Anda di halaman 1dari 111

IJCSI proceedings are currently indexed by:


In 2009, we witnessed a wide range of topics published in IJCSI.

Submissions of papers varied from short papers, to full research papers
without forgetting technical notes. The year 2009 has also been marked
with the publication of two special issues “Pervasive Computing Systems
and Technologies” and “Security Systems and Technologies”. Acceptance
rate for year 2009 varied from 25% to 37%.

Special thanks to all reviewers who have been taking hours out of their
precious time to review papers sent to them. One must not forget that a
published article does include the contribution of the reviewers as well in
the form of proof-reading, typos identification, grammar correction,
improving paper presentation, spotting weaknesses, and as whole help to
improve the quality of the final article before its publication.

We are pleased to present IJCSI Volume 7, Issue 1, split in three numbers

(IJCSI Vol. 7, Issue 1, No. 3). The call for papers started since early
November 2009 and attracted a total of 98 submissions out of which 35
papers have been accepted for publication, thus, resulting to an
acceptance rate of 35.7%.

We wish you a happy reading!

IJCSI Editorial Board

January 2010
IJCSI Editorial Board 2010

Dr Tristan Vanrullen
Chief Editor
LPL, Laboratoire Parole et Langage - CNRS - Aix en Provence, France
LABRI, Laboratoire Bordelais de Recherche en Informatique - INRIA - Bordeaux, France
LEEE, Laboratoire d'Esthétique et Expérimentations de l'Espace - Université d'Auvergne, France

Dr Constantino Malagôn
Associate Professor
Nebrija University

Dr Mokhtar Beldjehem
Sainte-Anne University
Halifax, NS, Canada

Dr Pascal Chatonnay
Assistant Professor
MaÎtre de Conférences
Laboratoire d'Informatique de l'Université de Franche-Comté
Université de Franche-Comté

Dr Yee-Ming Chen
Department of Industrial Engineering and Management
Yuan Ze University

Dr Vishal Goyal
Assistant Professor
Department of Computer Science
Punjabi University
Patiala, India
Dr Natarajan Meghanathan
Assistant Professor
REU Program Director
Department of Computer Science
Jackson State University
Jackson, USA

Dr Deepak Laxmi Narasimha

Department of Software Engineering,
Faculty of Computer Science and Information Technology,
University of Malaya,
Kuala Lumpur, Malaysia

Dr Navneet Agrawal
Assistant Professor
Department of ECE,
College of Technology & Engineering,
MPUAT, Udaipur 313001 Rajasthan, India

Prof N. Jaisankar
Assistant Professor
School of Computing Sciences,
VIT University
Vellore, Tamilnadu, India
IJCSI Reviewers Committee 2010

 Mr. Markus Schatten, University of Zagreb, Faculty of Organization and Informatics,

 Mr. Vassilis Papataxiarhis, Department of Informatics and Telecommunications, National
and Kapodistrian University of Athens, Panepistimiopolis, Ilissia, GR-15784, Athens,
Greece, Greece
 Dr Modestos Stavrakis, University of the Aegean, Greece
 Dr Fadi KHALIL, LAAS -- CNRS Laboratory, France
 Dr Dimitar Trajanov, Faculty of Electrical Engineering and Information technologies, ss.
Cyril and Methodius Univesity - Skopje, Macedonia
 Dr Jinping Yuan, College of Information System and Management,National Univ. of
Defense Tech., China
 Dr Alexis Lazanas, Ministry of Education, Greece
 Dr Stavroula Mougiakakou, University of Bern, ARTORG Center for Biomedical
Engineering Research, Switzerland
 Dr DE RUNZ, CReSTIC-SIC, IUT de Reims, University of Reims, France
 Mr. Pramodkumar P. Gupta, Dept of Bioinformatics, Dr D Y Patil University, India
 Dr Alireza Fereidunian, School of ECE, University of Tehran, Iran
 Mr. Fred Viezens, Otto-Von-Guericke-University Magdeburg, Germany
 Mr. J. Caleb Goodwin, University of Texas at Houston: Health Science Center, USA
 Dr. Richard G. Bush, Lawrence Technological University, United States
 Dr. Ola Osunkoya, Information Security Architect, USA
 Mr. Kotsokostas N.Antonios, TEI Piraeus, Hellas
 Prof Steven Totosy de Zepetnek, U of Halle-Wittenberg & Purdue U & National Sun
Yat-sen U, Germany, USA, Taiwan
 Mr. M Arif Siddiqui, Najran University, Saudi Arabia
 Ms. Ilknur Icke, The Graduate Center, City University of New York, USA
 Prof Miroslav Baca, Associated Professor/Faculty of Organization and
Informatics/University of Zagreb, Croatia
 Dr. Elvia Ruiz Beltrán, Instituto Tecnológico de Aguascalientes, Mexico
 Mr. Moustafa Banbouk, Engineer du Telecom, UAE
 Mr. Kevin P. Monaghan, Wayne State University, Detroit, Michigan, USA
 Ms. Moira Stephens, University of Sydney, Australia
 Ms. Maryam Feily, National Advanced IPv6 Centre of Excellence (NAV6) , Universiti
Sains Malaysia (USM), Malaysia
 Dr. Constantine YIALOURIS, Informatics Laboratory Agricultural University of Athens,
 Mrs. Angeles Abella, U. de Montreal, Canada
 Dr. Patrizio Arrigo, CNR ISMAC, italy
 Mr. Anirban Mukhopadhyay, B.P.Poddar Institute of Management & Technology, India
 Mr. Dinesh Kumar, DAV Institute of Engineering & Technology, India
 Mr. Jorge L. Hernandez-Ardieta, INDRA SISTEMAS / University Carlos III of Madrid,
 Mr. AliReza Shahrestani, University of Malaya (UM), National Advanced IPv6 Centre of
Excellence (NAv6), Malaysia
 Mr. Blagoj Ristevski, Faculty of Administration and Information Systems Management -
Bitola, Republic of Macedonia
 Mr. Mauricio Egidio Cantão, Department of Computer Science / University of São Paulo,
 Mr. Jules Ruis, Fractal Consultancy, The netherlands
 Mr. Mohammad Iftekhar Husain, University at Buffalo, USA
 Dr. Deepak Laxmi Narasimha, Department of Software Engineering, Faculty of
Computer Science and Information Technology, University of Malaya, Malaysia
 Dr. Paola Di Maio, DMEM University of Strathclyde, UK
 Dr. Bhanu Pratap Singh, Institute of Instrumentation Engineering, Kurukshetra
University Kurukshetra, India
 Mr. Sana Ullah, Inha University, South Korea
 Mr. Cornelis Pieter Pieters, Condast, The Netherlands
 Dr. Amogh Kavimandan, The MathWorks Inc., USA
 Dr. Zhinan Zhou, Samsung Telecommunications America, USA
 Mr. Alberto de Santos Sierra, Universidad Politécnica de Madrid, Spain
 Dr. Md. Atiqur Rahman Ahad, Department of Applied Physics, Electronics &
Communication Engineering (APECE), University of Dhaka, Bangladesh
 Dr. Charalampos Bratsas, Lab of Medical Informatics, Medical Faculty, Aristotle
University, Thessaloniki, Greece
 Ms. Alexia Dini Kounoudes, Cyprus University of Technology, Cyprus
 Mr. Anthony Gesase, University of Dar es salaam Computing Centre, Tanzania
 Dr. Jorge A. Ruiz-Vanoye, Universidad Juárez Autónoma de Tabasco, Mexico
 Dr. Alejandro Fuentes Penna, Universidad Popular Autónoma del Estado de Puebla,
 Dr. Ocotlán Díaz-Parra, Universidad Juárez Autónoma de Tabasco, México
 Mrs. Nantia Iakovidou, Aristotle University of Thessaloniki, Greece
 Mr. Vinay Chopra, DAV Institute of Engineering & Technology, Jalandhar
 Ms. Carmen Lastres, Universidad Politécnica de Madrid - Centre for Smart
Environments, Spain
 Dr. Sanja Lazarova-Molnar, United Arab Emirates University, UAE
 Mr. Srikrishna Nudurumati, Imaging & Printing Group R&D Hub, Hewlett-Packard,
 Dr. Olivier Nocent, CReSTIC/SIC, University of Reims, France
 Mr. Burak Cizmeci, Isik University, Turkey
 Dr. Carlos Jaime Barrios Hernandez, LIG (Laboratory Of Informatics of Grenoble),
 Mr. Md. Rabiul Islam, Rajshahi university of Engineering & Technology (RUET),
 Dr. LAKHOUA Mohamed Najeh, ISSAT - Laboratory of Analysis and Control of
Systems, Tunisia
 Dr. Alessandro Lavacchi, Department of Chemistry - University of Firenze, Italy
 Mr. Mungwe, University of Oldenburg, Germany
 Mr. Somnath Tagore, Dr D Y Patil University, India
 Ms. Xueqin Wang, ATCS, USA
 Dr. Fondjo Fotou Franklin, Langston University, USA
 Dr. Vishal Goyal, Department of Computer Science, Punjabi University, Patiala, India
 Mr. Thomas J. Clancy, ACM, United States
 Dr. Ahmed Nabih Zaki Rashed, Dr. in Electronic Engineering, Faculty of Electronic
Engineering, menouf 32951, Electronics and Electrical Communication Engineering
Department, Menoufia university, EGYPT, EGYPT
 Dr. Rushed Kanawati, LIPN, France
 Mr. Koteshwar Rao, K G Reddy College Of ENGG.&TECH,CHILKUR, RR DIST.,AP,
 Mr. M. Nagesh Kumar, Department of Electronics and Communication, J.S.S. research
foundation, Mysore University, Mysore-6, India
 Dr. Ibrahim Noha, Grenoble Informatics Laboratory, France
 Mr. Muhammad Yasir Qadri, University of Essex, UK
 Mr. Annadurai .P, KMCPGS, Lawspet, Pondicherry, India, (Aff. Pondicherry Univeristy,
 Mr. E Munivel , CEDTI (Govt. of India), India
 Dr. Chitra Ganesh Desai, University of Pune, India
 Mr. Syed, Analytical Services & Materials, Inc., USA
 Dr. Mashud Kabir, Department of Computer Science, University of Tuebingen, Germany
 Mrs. Payal N. Raj, Veer South Gujarat University, India
 Mrs. Priti Maheshwary, Maulana Azad National Institute of Technology, Bhopal, India
 Mr. Mahesh Goyani, S.P. University, India, India
 Mr. Vinay Verma, Defence Avionics Research Establishment, DRDO, India
 Dr. George A. Papakostas, Democritus University of Thrace, Greece
 Mr. Abhijit Sanjiv Kulkarni, DARE, DRDO, India
 Mr. Kavi Kumar Khedo, University of Mauritius, Mauritius
 Dr. B. Sivaselvan, Indian Institute of Information Technology, Design & Manufacturing,
Kancheepuram, IIT Madras Campus, India
 Dr. Partha Pratim Bhattacharya, Greater Kolkata College of Engineering and
Management, West Bengal University of Technology, India
 Mr. Manish Maheshwari, Makhanlal C University of Journalism & Communication,
 Dr. Siddhartha Kumar Khaitan, Iowa State University, USA
 Dr. Mandhapati Raju, General Motors Inc, USA
 Dr. M.Iqbal Saripan, Universiti Putra Malaysia, Malaysia
 Mr. Ahmad Shukri Mohd Noor, University Malaysia Terengganu, Malaysia
 Mr. Selvakuberan K, TATA Consultancy Services, India
 Dr. Smita Rajpal, Institute of Technology and Management, Gurgaon, India
 Mr. Rakesh Kachroo, Tata Consultancy Services, India
 Mr. Raman Kumar, National Institute of Technology, Jalandhar, Punjab., India
 Mr. Nitesh Sureja, S.P.University, India
 Dr. M. Emre Celebi, Louisiana State University, Shreveport, USA
 Dr. Aung Kyaw Oo, Defence Services Academy, Myanmar
 Mr. Sanjay P. Patel, Sankalchand Patel College of Engineering, Visnagar, Gujarat, India
 Dr. Pascal Fallavollita, Queens University, Canada
 Mr. Jitendra Agrawal, Rajiv Gandhi Technological University, Bhopal, MP, India
 Mr. Ismael Rafael Ponce Medellín, Cenidet (Centro Nacional de Investigación y
Desarrollo Tecnológico), Mexico
 Mr. Supheakmungkol SARIN, Waseda University, Japan
 Mr. Shoukat Ullah, Govt. Post Graduate College Bannu, Pakistan
 Dr. Vivian Augustine, Telecom Zimbabwe, Zimbabwe
 Mrs. Mutalli Vatila, Offshore Business Philipines, Philipines
 Dr. Emanuele Goldoni, University of Pavia, Dept. of Electronics, TLC & Networking
Lab, Italy
 Mr. Pankaj Kumar, SAMA, India
 Dr. Himanshu Aggarwal, Punjabi University,Patiala, India
 Dr. Vauvert Guillaume, Europages, France
 Prof Yee Ming Chen, Department of Industrial Engineering and Management, Yuan Ze
University, Taiwan
 Dr. Constantino Malagón, Nebrija University, Spain
 Prof Kanwalvir Singh Dhindsa, B.B.S.B.Engg.College, Fatehgarh Sahib (Punjab), India
 Mr. Angkoon Phinyomark, Prince of Singkla University, Thailand
 Ms. Nital H. Mistry, Veer Narmad South Gujarat University, Surat, India
 Dr. M.R.Sumalatha, Anna University, India
 Mr. Somesh Kumar Dewangan, Disha Institute of Management and Technology, India
 Mr. Raman Maini, Punjabi University, Patiala(Punjab)-147002, India
 Dr. Abdelkader Outtagarts, Alcatel-Lucent Bell-Labs, France
 Prof Dr. Abdul Wahid, AKG Engg. College, Ghaziabad, India
 Mr. Prabu Mohandas, Anna University/Adhiyamaan College of Engineering, india
 Dr. Manish Kumar Jindal, Panjab University Regional Centre, Muktsar, India
 Prof Mydhili K Nair, M S Ramaiah Institute of Technnology, Bangalore, India
 Dr. C. Suresh Gnana Dhas, VelTech MultiTech Dr.Rangarajan Dr.Sagunthala
Engineering College,Chennai,Tamilnadu, India
 Prof Akash Rajak, Krishna Institute of Engineering and Technology, Ghaziabad, India
 Mr. Ajay Kumar Shrivastava, Krishna Institute of Engineering & Technology,
Ghaziabad, India
 Mr. Deo Prakash, SMVD University, Kakryal(J&K), India
 Dr. Vu Thanh Nguyen, University of Information Technology HoChiMinh City,
 Prof Deo Prakash, SMVD University (A Technical University open on I.I.T. Pattern)
Kakryal (J&K), India
 Dr. Navneet Agrawal, Dept. of ECE, College of Technology & Engineering, MPUAT,
Udaipur 313001 Rajasthan, India
 Mr. Sufal Das, Sikkim Manipal Institute of Technology, India
 Mr. Anil Kumar, Sikkim Manipal Institute of Technology, India
 Dr. B. Prasanalakshmi, King Saud University, Saudi Arabia.
 Dr. K D Verma, S.V. (P.G.) College, Aligarh, India
 Mr. Mohd Nazri Ismail, System and Networking Department, University of Kuala
Lumpur (UniKL), Malaysia
 Dr. Nguyen Tuan Dang, University of Information Technology, Vietnam National
University Ho Chi Minh city, Vietnam
 Dr. Abdul Aziz, University of Central Punjab, Pakistan
 Dr. P. Vasudeva Reddy, Andhra University, India
 Mrs. Savvas A. Chatzichristofis, Democritus University of Thrace, Greece
 Mr. Marcio Dorn, Pontifical Catholic University of Rio Grande do Sul, Rio Grande do
Sul, Brazil
 Mr. Luca Mazzola, University of Lugano, Switzerland
 Mr. Nadeem Mahmood, Department of Computer Science, University of Karachi,
 Mr. Hafeez Ullah Amin, Kohat University of Science & Technology, Pakistan
 Dr. Professor Vikram Singh, Ch. Devi Lal University, Sirsa (Haryana), India
 Mr. M. Azath, Calicut/Mets School of Enginerring, India
 Dr. J. Hanumanthappa, DoS in CS, University of Mysore, India
 Dr. Shahanawaj Ahamad, Department of Computer Science, King Saud University, Saudi
 Dr. K. Duraiswamy, K. S. Rangasamy College of Technology, India
 Prof. Dr Mazlina Esa, Universiti Teknologi Malaysia, Malaysia
 Dr. P. Vasant, Power Control Optimization (Global), Malaysia
 Dr. Taner Tuncer, Firat University, Turkey
 Dr. Norrozila Sulaiman, University Malaysia Pahang, Malaysia
 Prof. S K Gupta, BCET, Guradspur, India
 Dr. Latha Parameswaran, Amrita Vishwa Vidyapeetham, India
 Mr. M. Azath, Anna University, India
 Dr. P. Suresh Varma, Adikavi Nannaya University, India
 Prof. V. N. Kamalesh, JSS Academy of Technical Education, India
 Dr. D Gunaseelan, Ibri College of Technology, Oman
 Mr. Sanjay Kumar Anand, CDAC, India
 Mr. Akshat Verma, CDAC, India
 Mrs. Fazeela Tunnisa, Najran University, Kingdom of Saudi Arabia
 Mr. Hasan Asil, Islamic Azad University Tabriz Branch (Azarshahr), Iran
 Prof. Dr Sajal Kabiraj, Fr. C Rodrigues Institute of Management Studies (Affiliated to
University of Mumbai, India), India
 Mr. Syed Fawad Mustafa, GAC Center, Shandong University, China
 Dr. Natarajan Meghanathan, Jackson State University, Jackson, MS, USA
 Prof. Selvakani Kandeeban, Francis Xavier Engineering College, India
 Mr. Tohid Sedghi, Urmia University, Iran
 Dr. S. Sasikumar, PSNA College of Engg and Tech, Dindigul, India
 Dr. Anupam Shukla, Indian Institute of Information Technology and Management
Gwalior, India
 Mr. Rahul Kala, Indian Institute of Inforamtion Technology and Management Gwalior,
 Dr. A V Nikolov, National University of Lesotho, Lesotho
 Mr. Kamal Sarkar, Department of Computer Science and Engineering, Jadavpur
University, India
 Dr. Mokhled S. AlTarawneh, Computer Engineering Dept., Faculty of Engineering,
Mutah University, Jordan, Jordan
 Prof. Sattar J Aboud, Iraqi Council of Representatives, Iraq-Baghdad
 Dr. Prasant Kumar Pattnaik, Department of CSE, KIST, India
 Dr. Mohammed Amoon, King Saud University, Saudi Arabia

1. RFID Applications: An Introductory and Exploratory Study [pg 1-7]

Kamran Ahsan, Faculty of Computing, Engineering and Technology & Centre for Ageing and
Mental Health, Staffordshire University, Stafford, United Kingdom
Hanifa Shah and Paul Kingston, Faculty of Computing, Engineering & Technology,
Staffordshire University, Stafford, ST18 0AD, United Kingdom

2. The Design of Circuit-Measuring Collaborative Learning System with Embedded

Broker [pg 8-17]
Fu-Chien Kao, Siang-Ru Wang and Ting-Hao Huang, Computer Science and Information
Engineering, Da-Yeh University, Taiwan

3. Implementation of an Innovative Bio Inspired GA and PSO Algorithm for Controller

design considering Steam GT Dynamics [pg 18-28]
R. Shivakumar, Department of Electrical and Electronics Engineering, Sona College of
Technology, Salem-636005, India
R. Lakshmipathi, Department of Electrical and Electronics Engineering, St.Peters Engineering
College, Chennai, India

4. Significant Interval and Frequent Pattern Discovery in Web Log Data [pg 29-36]
Kanak Saxena, Computer Application Department. R.G.P.V., S.A.T.I. Vidisha, M.P. , India
Rahul Shukla, Samrat Ashok Technological Institute, Vidisha (M.P.), India

5. Efficient Packet Forwarding Approach in Vehicular Ad Hoc Networks Using EBGR

Algorithm [pg 37-46]
K. Prasanth, Department of Information Technology, K.S.Rangasamy College of Technology,
Tiruchengode - 637 215, Tamilnadu, India
K. Duraiswamy, Department of Computer Science, K.S.Rangasamy College of Technology,
Tiruchengode - 637 215, Tamilnadu, India
K. Jayasudha, Department of Computer Applications, K.S.R College of Engineering,
Tiruchengode - 637 215, Tamilnadu, India
C. Chandrasekar, Periyar University, Salem, Tamilnadu, India

6. Framework for Visualizing Model-Driven Software Evolution and its Application

[pg 47-53]
Akepogu Anand Rao and Karanam Madhavi, Computer Science & Engineering Department,
Jawaharlal Nehru Technological University, Anantapur, Andhra Pradesh, India
7. Unidirectional Error Correcting Codes for Memory Systems: A Comparative Study
[pg 54-61]
Muzhir Al-Ani, Faculty of IT, Amman Arab University, Amman, Jordan
Qeethara Al-Shayea, MIS Department, Al-Zaytoonah University, Amman, Jordan

8. ICT in Universities of the Western Himalayan Region of India II: A Comparative SWOT
Analysis [pg 62-72]
Dhirendra Sharma, University Institute of Information Technology, Himachal Pradesh
University, Shimla, Himachal Pradesh 171 005, India
Vikram Singh, Department of Computer Science and Engg, Ch. Devi Lal University, Sirsa,
Haryna 125 055, India

9. Stochastic Model Based Proxy Servers Architecture for VoD to Achieve Reduced Client
Waiting Time [pg 73-80]
T. R. GopalaKrishnan nair, Director Research Industry and Incubation Centre DSI ,Bangalore,
M. Dakshayini, Dr. MGR University, Working with Dept. of ISE, BMSCE, Bangalore. Member,
Multimedia Research Group, Research Centre, DSI, Bangalore, India

10. Modified EESM Based Link Adaptation Algorithm for Multimedia Transmission in
Multicarrier Systems [pg 81-86]
R. Sandanalakshmi, Athilakshmi and K. Manivannan, Department of Electronics and
Communication Engineering, Pondicherry Engineering College, Pondicherry, India

11. Reliable Mining of Automatically Generated Test Cases from Software Requirements
Specification (SRS) [pg 87-91]
Lilly Raamesh, Research Scholar, Anna University, Chennai 25, India
G. V. Uma, CSE, Anna University, Chennai-25, India

12. Understanding Formulation of Social Capital in Online Social Network Sites (SNS)
[pg 92-96]
S. S. Phulari, Santosh Khamitkar, Nilesh Deshmukh, Parag Bhalchandra, Sakharam Lokhande
and A. R. Shinde, School of Computational Sciences, Swami Ramanand Teerth Marathwada
University, Nanded, MS, India, 431606
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 1
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

RFID Applications: An Introductory and Exploratory Study

Kamran AHSAN1, Hanifa SHAH2 and Paul KINGSTON3
Faculty of Computing, Engineering & Technology
Staffordshire University
Stafford, ST18 0AD, UK

1, 3
Centre for Ageing & Mental Health

Abstract realized as a performance differentiator for a variety of

RFID is not a new technology and has passed through many commercial applications, but its capability is yet to be
decades of use in military, airline, library, security, healthcare, fully utilised.
sports, animal farms and other areas. Industries use RFID for
various applications such as personal/vehicle access control,
departmental store security, equipment tracking, baggage, fast
food establishments, logistics, etc. The enhancement in RFID
2. RFID Evolution
technology has brought advantages that are related to resource
optimization, increased efficiency within business processes,
RFID technology has passed through many phases over
and enhanced customer care, overall improvements in business the last few decades (see figure 1). The technology has
operations and healthcare. Our research is part of a big project; been used in tracking delivery of goods, in courier
its aim is to produce a model for mobile technology services and in baggage handling. Other applications
implementation of hospital patients’ movement process. includes automatic toll payments, departmental access
However, the focus of this paper is to explore the main RFID control in large buildings, personal and vehicle control in
components, i.e. the tag, antenna and reader. The results of the a particular area, security of items which shouldn’t leave
investigations conducted on the three RFID components will be the area, equipment tracking in engineering firms,
used to develop our research model.
hospital filing systems, etc.[4, 5]. Figure 1 shows RFID
evolution over the past few decades.
Keywords: RFID technology, RFID discovery, RFID
components, RFID applications, RFID in healthcare.

1. Introduction
RFID stands for Radio Frequency Identification and is a
term that describes a system of identification [1]. RFID
is based on storing and remotely retrieving information
or data as it consists of RFID tag, RFID reader and back-
end Database [2]. RFID tags store unique identification
information of objects and communicate the tags so as to
allow remote retrieval of their ID. RFID technology
depends on the communication between the RFID tags
and RFID readers. The range of the reader is dependent
upon its operational frequency. Usually the readers have
their own software running on their ROM and also,
communicate with other software to manipulate these
unique identified tags [3]. Basically, the application
which manipulates tag deduction information for the end
user, communicates with the RFID reader to get the tag
information through antennas. Many researchers have
addressed issues that are related to RFID reliability and
capability [2]. RFID is continuing to become popular
because it increases efficiency and provides better
Fig. 1 RFID evolution: Over past the few decades adapted from [6]
service to stakeholders [1]. RFID technology has been
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 2

3. How RFID System Works The reader performs these operations one by one on each
tag. A typical RFID system work cycle can be seen in
Most RFID systems consist of tags that are attached to figure 2.
the objects to be identified. Each tag has its own “read-
only” or “rewrite” internal memory depending on the
type and application [7]. Typical configuration of this 4. Components of an RFID System
memory is to store product information, such as an
object’s unique ID manufactured date, etc. The RFID The RFID system consists of various components which
reader generates magnetic fields that enable the RFID are intergrated in a manner defined in the above section.
system to locate objects (via the tags) that are within its This allows the RFID system to deduct the objects (tag)
range [5]. The high-frequency electromagnetic energy and perform vaious operations on it. The intergration of
and query signal generated by the reader triggers the tags RFID components enables the implementation of an
to reply to the query; the query frequency could be up to RFID solution [8]. The RFID system consists of
50 times per second. [4]. As a result communication following five components (as shown in Figure 3):
between the main components of the system i.e. tags and • Tag (attached with an
reader is established [6]. As a result large quantities of object, unique identification).
data are generated. Supply chain industries control this
problem by using filters that are routed to the backend • Antenna (tag detector, creates magnetic field).
information systems. In other words, in order to control
this problem, software such as Savant is used. This • Reader (receiver of tag information,
software acts as a buffer between the Information manipulator).
Technology and RFID reader [6, 7]. • Communication infrastructure (enable
Several protocols manage the communication process reader/RFID to work through IT infrastructure).
between the reader and tag. These protocols (ISO 15693 • Application software (user database/application/
and ISO 18000-3 for HF or the ISO 18000-6, and EPC interface).
for UHF) begin the identification process when the
reader is switched on. These protocol works on selected
frequency bands (e.g. 860 – 915 MHz for UHF or 13.56
MHz for HF). If the reader is on and the tag arrives in the
reader fields, then it automatically wakes-up and decodes
the signal and replies to the reader by modulating the
reader’s field [7]. All the tags in the reader range may
reply at the same time, in this case the reader must detect
signal collision (indication of multiple tags) [6]. Signal
collision is resolved by applying anti-collision algorithm
which enables the reader to sort tags and select/handle
each tag based on the frequency range (between 50 tags
to 200 tags) and the protocol used. In this connection the
reader can perform certain operations on the tags such as
reading the tag’s identifier number and writing data into
a tag [7].
Fig. 3 Components of an RFID System

5. Tags
Tags contain microchips that store the unique
identification (ID) of each object. The ID is a serial
number stored in the RFID memory. The chip is made up
of intergrated circuit and embedded in a silicon chip [7].
RFID memory chip can be permanent or changeable
depending on the read/write characteristics. Read-only
and rewrite circuits are different as read-only tag contain
fixed data and can not be changed without re-program
electonically [5]. On the other hand, re-write tags can be
Fig. 2 A typical RFID System [7]
programmed through the reader at any time without any
limit. RFID tags can be different sizes and shapes
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 3

depending on the application and the environment at

which it will be used. A variety of materials are
intergrated on these tags. For example, in the case of the
credit cards, small plastic peaces are stuck on various
objects, and the labels. Labels are also emmbeded in a
variety of objects such as documents, cloths,
maufacturing mateirals etc [9]. Figure 4 demonstrates the
different sizes and shapes of the RFID tags.

Fig. 6 RFID active and passive tags comparison

5.1 Tag Frequencies

The range of the RFID tags depends on their frequency.
This frequency determines the resistance to interference
and other performance attributes [12]. The use/selection
of RFID tag depends on the application; different
frequencies are used on different RFID tags [10].
EPCglobal and International Standards Organization
Fig. 4 Varity of RFID tags (various shape & sizes) [9] (ISO) are the major organizations working to develop
international standards for RFID technologies in the
RFID tags can also be classified by their capabilities UHF band. These two organizations are still evolving
such as read and write data [10]. Figure 5 shows the five and are not fully compatible with each other [14]. In
classifications of the RFID tags [10]. order to avoid the use of different radio frequencies
standards, most of the international communities are
obligated to comply with the International
Telecommunication Union (ITU) standards. The
following are the commonly used frequencies:
• Microwave works on 2.45 GHz, it has good
reader rate even faster than UHF tags. Although
at this frequency the reading rate results are not
the same on wet surfaces and near metals, the
frequency produce better results in applications
such as vehicle tracking (in and out with
barriers), with approximately 1 meter of tags
read range [7].
• Ultra High Frequency works within a range of
860-930 MHz, it can identify large numbers of
tags at one time with quick multiple read rate at a
Fig. 5 RFID tags classifications
given time. So, it has a considerable good
reading speed. It has the same limitation as
There are three types of tags: the passive, semi-active Microwave when is applied on wet surface and
and active. Semi-active tags have a combination of active near metal. However, it is faster than high
and passive tags characteristics. So, mainly two types of frequency data transfer with a reading range of 3
tags (active and passive) are being used by industry and meters [7].
most of the RFID system [7]. The essential • High Frequency works on 13.56MHz and has
characteristics of RFID tags are their function to the less than one meter reading range but is
RFID system. This is based on their range, frequency, inexpensive and useful for access control, items
memory, security, type of data and other characteristics. identifications on sales points etc as it can
These characteristics are core for RFID performance and implanted inside thin things such as paper [6, 7].
differ in usefulness/support to the RFID system
operations [4, 11]. While considering these • Low Frequency works on 125 kHz, it has
characteristics, figure 6 compares the active and passive approximately half a meter reading range and
tags. mostly used for short reading range applications
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 4

such as shops, manufacturing factories, inventory

control through in and out counts, access control Near Fields: Near field uses method similar to
through showing a card to the reader. These low transformer, and employs inductive coupling of the tag to
frequency tags are mostly not affected when the magnetic field circulating around the reader antenna
applied on wet and near metal surfaces [7, 9]. (see figure 8).

6. Antennas
RFID antennas collect data and are used as a medium for
tag reading [7]. It consists of the following:

Fig. 7 RFID antennas types [11]

Fig. 8 RFID near field methodology [7]
(1) Patch antennas, (2) Gate antennas, (3) Linear
polarized, (4) Circular polarized, (5) Di-pole or multi- Far Field: Far field uses method similar to radar,
pole antennas, (6) Stick antennas, (7) Beam-forming or backscatter reflection by coupling with the electric field.
phased-array element antennas, (8) Adaptive antennas,
and (9) Omni directional antennas.

7. RFID Reader
RFID reader works as a central place for the RFID
system. It reads tags data through the RFID antennas at a
certain frequency [7, 9]. Basically, the reader is an
electronic apparatus which produce and accept a radio
signals [15]. The antennas contains an attached reader,
the reader translates the tags radio signals through Fig. 9 RFID near field methodology [7]
antenna, depending on the tags capacity [16]. The
readers consist of a build-in anti-collision schemes and a The distinction between the RFID systems with far fields
single reader can operate on multiple frequencies. As a to the near fields is that the near fields use LF (lower
result, these readers are expected to collect or write data frequency) and HF (higher frequency) bands [17, 18].
onto tag (in case) and pass to computer systems. For this While RFID systems with far fields usually use longer
purpose readers can be connected using RS-232, RS-485, read range UHF and microwave [17].
USB cable as a wired options (called serial readers) and
connect to the computer system. Also can use WiFi as
wireless options which also known as network readers 8. Advantages & Disadvantages of RFID
[8, 12]. Readers are electronic devices which can be System
used as standalone or be integrated with other devices
and the following components/hardware into it [12]. Table 1: Comparison of RFID System
Advantage Disadvantage
Power for running reader, (2) Communication interface, High speed Interference
(3) Microprocessor, (4) Channels, (5) Controller, (6) Multipurpose and many High cost
Receiver, (7) Transmitter, (8) Memory. format
Reduce man-power Some materials may create
signal problem
7.1 Tag Standards High accuracy Overloaded reading (fail to
Readers use near and far fields of methodology to Complex duplication
communicate to the tag through its antennas [7]. If a tag Multiple reading (tags)
wants to respond to the reader then the tag will need to
receive energy and communicate with a reader. For
example, passive tags use either one of the two following
methods [7, 11].
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 5

9. Study Model explored. In order to understand the benefits of RFID

application in a hospital case, this paper explores general
Our research aims to use “context based knowledge RFID applications shown in figure 10.
management” [4] to produce a model for mobile
technology implementation within patients’ movement
processes. In order to do this we need to investigate
RFID feasibility and integration with hospital
information systems for improving healthcare [16]. This
paper explores the fundamentals of RFID technology.
That is the advantages, types, limitations and
applications that will further help to develop a model for
hospital application. The implementation of RFID in
patient’ flow is intended to improve healthcare especially
in hospital settings. RFID technology can provide new
capabilities as well as an efficient methods for several
applications such as health care, access control, manage,
store and analyze information inventory, business
processes, and security controls through access to
information [17]. The following section will describe the Figure 10. General RFID applications according to its capabilities
RFID applications [18] which will further help us to
Object identification can be given through various ways
develop a model for mobile technology implementation
such as barcode, biometric and RFID. RFID has two
within patients’ movement processes.
basic categories (short & long range). The short range
applications need tags to be near reader, it is useful in
various condition such as when a patient is required to
10. Application Investigation come near the door/reader and only one person can get
access (access control). The long range applications may
The hospital case study conducted by this research shows
not need tags that are closer to reader. Similar scenarios
that there are objects which need to be considered when
are successfully been used on various items in the
developing a model to represent patients’ flow [4]. These
warehouse (logistics) as shown in figure 9. The most
objects are associated with location. The finite set of
common application is defined in figure 9 and is
locations within the hospital will be captured through
discussed in Section 10.2 together with other
mobile technology in a live environment [16]. The
following components have been observed in an overall
picture of patients’ movement [4]. However, the in depth
investigation of each component is yet to be explored. 10.2 Healthcare Applications

• The number of paramedical staff involved in RFID applications in healthcare could save important
patients’ movement processes. resources that can further contribute to better patient
care. RFID applications could reduce the number of
• The number of actions performed in patients’ errors by tagging medical objects in the healthcare
movement processes. setting such as patients’ files and medical equipment
• The resources involved in an patients’ movement tracking in a timely manner. RFID further improves the
processes. situation for patients’ care by integrating medical objects
involved throughout the patients’ care. RFID based
• The finite number of locations used for patients’ timely information about the location of objects would
movement processes. increase the efficiency and effectiveness of paramedical
staff leading to improved patients’ experience [4, 16].
• The process of integrating patients’ movement
information with an existing IT infrastructure.
10.3 Security & Control Applications
The system should enable the integration and
optimization of resources while improving accuracy and RFID tags can be attached to the equipment/user
minimizing patients’ transition time leading to personal/official belongings such as organization ID
improvements in patients’ services. cards and vehicles. By applying RFID application in
secure zones, not only permission can be granted to and
10.1 Type of RFID Applications revoke for the users/persons in particular zone but also
record individual access and the length of their stay. It is
In previous section components are identified in hospital also good for audit trial. These types of application
case. However, the detailed investigation is yet to be
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 6

consider time and flow carefully and as an aspect that is advantages and study model. The last part explores RFID
very important [19]. technology applications. The paper considers RFID
technology as a means to provide new capabilities and
10.4 Patrolling Log Applications efficient methods for several applications. For example,
in heathcare, access control, analyzing investory
RFID is also used for auditing and controlling security information, and business processes. RFID technology
persons themselves. Application provides checkpoints needs to develop its capability to be used with computing
for patrolling the security guards. Checkpoints are devices. This will allow businesses to get real potential
basically a RFID tag which security guard needs to scan benefits of RFID technology. This study facilitates
during their sequential patrol through the reader. The adoption of location deduction technology (RFID) in a
reader maintains the record of the time and point at healthcare environment and shows the importance of the
which the security guard swapped his card. This will not technology in a real scenario and application in
only help security firms administration to check the connection with resource optimization and improving
performance of its security guards but also used as a effectiveness. However, there is no doubt in the future
reference to track events. This application can also help that many companies and organisations will benefit from
to improve the patrolling process, e.g. through RFID technology.
identifying the need to increase patrols or check points in
a patrolled area.
Future Work
10.5 Baggage Applications
Our work continues to develop an enterprise architectural
Airline industries, package and delivery service lose a lot framework for managing contextual knowledge by
of money on lost or late delivery of baggage/packages. exploiting object location deduction technologies such as
Handling large amount of packages from many places to RFID in healthcare processes that involve the movement
various destinations on different routes can be very of patients. Such a framework is intended to facilitate
complex. In this scenario RFID application provide best healthcare managers in adopting RFID for patient care
resource management, effective operation and efficient resulting in improvements in clinical process
transfer of packages. RFID helps to identify the management and healthcare services.
packages, and provide records that can advice the
industry on possible areas that may require some
improvements. It also keeps customers informed about Acknowledgments
their packages.
This research/project is funded by SaTH NHS Trust, UK
10.6 Toll Road Applications and this RFID applications study is one of many
objectives set in the larger project on “Context based
RFID applications make the toll collection/charging knowledge management in healthcare”.
better with improved traffic flow, as cars/vehicles cannot
pass through toll stations without stopping for payment.
RFID is used to automatically identify the account holder References
and make faster transactions. This application helps to [1] J. Bohn, “Prototypical implementation of location-aware
services based on a middleware architecture for super-
keep good traffic flow and to identify traffic patterns distributed RFID tag infrastructures”, Pers Ubiquit
using data mining techniques that can inform the omputing, (2008) Journal 12:155-166.
administration or decision support systems. For example, [2] J. Schwieren1, G. Vossen, “A Design and Development
the information can be used to report the traffic Methodology for Mobile RFID Applications based on the
conditions or to extend and develop future policies [19]. ID-Services Middleware Architecture”, IEEE Computer
Society, (2009), Tenth International Conference on Mobile
Data Management: Systems, Service and Middleware.
[3] B. Glover, & H. Bhatt, RFID Essentials, O’Reilly Media,
11. Conclusions Inc, Sebastopol, (2006), ISBN 0-596-00944-5.
[4] K. Ahsan, H. Shah, P. Kingston, “Context Based
This study has identified and explained the nature of Knowledge Management in Healthcare: An EA
RFID technology evolution with respect to RFID Approach”, AMCIS 2009, Available at AIS library.
applications. RFID technology will open new doors to [5] S. Garfinkel, B. Rosenberg, “RFID Application, Security,
make organisations, companies more secure, reliable, and Privacy”, USA, (2005), ISBN: 0-321-29096-8.
and accurate. The first part of this paper has explained [6] L. Srivastava, RFID: Technology, Applications and Policy
Implications, Presentation, International
and described the RFID technology and its components, Telecommunication Union, Kenya, (2005).
and the second part has discussed the main [7] Application Notes, “Introduction to RFID Technology”
considerations of RFID technology in terms of CAENRFID: The Art of Identification (2008).
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 7

[8] L. Sandip, “RFID Sourcebook”, IBM Press, USA, (2005) Paul Kingston is Professor of Health Sciences and Director of
ISBN: 0-13-185137-3. the Centre for Ageing and Mental Health at Staffordshire
[9] T. Frank, H. Brad, M. Anand, B. Hersh, C. Anita, K. John, University. Professor Kingston has 34 years of experience as a
“RFID Security”, (2006) ISBN: 1-59749-047-4. Healthcare researcher and also practitioner. He has presented
101 conference papers in a number of different countries. He
[10] A. Narayanan, S. Singh & M. Somasekharan, has co-authored or authored 7 books, and contributed 11
“Implementing RFID in Library: Methodologies, chapters. He has seen 33 journal articles published, produced 10
Advantages and Disadvantages”, (2005). reports/ monographs, and 3 working papers. Additionally Paul
[11] Intermec, “ABCs of RFID: Understanding and using radio has been involved in the development of training material in the
frequency identification”, White Paper, (2009). area of adult protection and was a key founder of the charity
[12] E. Zeisel & R. Sabella, “RFID+”, Exam Cram, (2006), Action on Elder Abuse. Furthermore he is a consulting editor for
ISBN: 0-7897-3504-0. the Journal of Elder Abuse and Neglect. Over the last ten years
[13] US. Department of Homeland Security, “Additional Paul has developed a strong and respected track record for
Guidence and Security Controls are needed over Systems developing research activity. During ten years he was either in
using RFID and DHS”, Department of Homeland Security receipt of, co-managing, or leading awards totaling £1,785,633.
(Office of Inspector General), (2006), OIG-06-53.
[14] US. Department of Homeland Security, “Enhanced
Security Controls needed for US-Visit’s System using
RFID Technology”, Department of Homeland Security
(Office of Inspector General), (2006), OIG-06-39.
[15] US. Government Accountability Office, “Information
Security: Radio Frequency Identification Technology in
the Federal Government”, (2005), Report to Congressional
Requesters, GAO-05-551.
[16] K. Ahsan, H. Shah, P. Kingston, “Role of Enterprise
Architecture in healthcare IT”, Proceeding ITNG2009,
(2009), IEEE.
[17] Y. Meiller & S. Bureau, “Logistics Projects: How to
Assess the Right System? The Case of RFID Solutions in
Healthcare”, Americas Conference on Information
Systems (AMCIS) 2009 Proceedings, Association for
Information Systems Year 2009.
[18] R. Parks, W. Yao & C. H. Chu, “RFID Privacy Concerns:
A Conceptual Analysis in the Healthcare Sector”,
Americas Conference on Information Systems (AMCIS)
2009 Proceedings, Association for Information Systems
Year 2009.
[19] S. Shepard, (2005), “RFID Radio Frequency
Identification”, (2005), USA, ISBN:0-07-144299-5.

Kamran AHSAN has an MSc in Mobile Computer Systems from

Staffordshire University in Computer Science from University of
Karachi. Kamran is a PhD researcher and lecturer in FCET
(Faculty of Computing, Engineering and Technology) and, web
researcher in Centre for Ageing and Mental Health, Staffordshire
University, UK since 2005. He has published 10 research
papers so far. He has involved in several research funding (UK)
includes KTP, NHS Trust, Index Vouchers etc. He is Visiting
Faculty at University of Karachi. He is a consultant to businesses
in IT applications, software development and web tools. His
research interests are mobile technology applications in
healthcare including knowledge management.

Hanifa Shah is Professor of Information Systems and Director of

the Centre for Information, Intelligence & Security Systems at
Staffordshire University. She is also a Visiting Professor at
Manchester University. Professor Shah has 30 years of
experience as an IS/IT researcher and also practitioner. She is a
Fellow of the British Computer Society and has led both UK
Research Council (EPSRC) funded and industrially funded
research projects. Her research has resulted in publications in
international journals and conferences. She has research
interests in a number of areas including the development of IS/IT
systems, enterprise architecture, knowledge management,
mobile technology based information systems, university-
industry collaborations, learning through action research and the
professional development of IS/IT practitioners in industry.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 8
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

The Design of Circuit-Measuring Collaborative Learning System

with Embedded Broker
Fu-Chien Kao, Siang-Ru Wang and Ting-Hao Huang

Computer Science and Information Engineering,

Da-Yeh University, Taiwan

ABSTRACT from the communication and argumentation. The

Recently, the academic community has been giving much traditional learning becomes an accumulation of
attention to Cooperative Learning System, a group learning knowledge. The teachers are responsible for inculcating
method combined with pedagogy and social psychology. It the results of the evolvement to the students. This learning
allows group members to gain knowledge through collaborations activity does not only drab but it is not meaningful.
and interactions. Nowadays, most Internet cooperative learning
systems are designed to provide students mainly with a
convenient online environment to study theoretical courses but In general, there are there learning modes, including
rarely with an online environment to operate practical Competitive Learning, Individualistic Learning and
instruments. Hence, this paper designed a 3D online cooperative Cooperative Learning (Huang, 1996). In the Competitive
learning system for operating virtual instruments with Learning, the students compete for the achievement with
circuit-measuring function. By integrating with Virtual Reality, each other. They pursue the targets that can be achieved by
Remote Control Parameter Transmission and embedded system a few students. Thus, the students may make every effort
techniques, this system gives learners not only a cooperative to win, or they may give up totally and escape from the
learning environment via networking to jointly operate the 3D reality. In the Individualistic Learning, the students are
virtual instruments (for example, multi-meters, power supplies
and oscilloscopes) but also the functions of instant messages and
arranged to make effort independently to achieve
3D puzzles to interact with one another. Therefore, learners can individual targets. The students pursue for individual
effectively improve learning interests and results. benefits, and they attach importance to their own efforts
Keywords: Cooperative learning system, embedded system, and results. They do not mind the learning and
virtual instruments achievements of other colleagues. In the Cooperative
Learning, they must help each other. They discuss, ask and
answer questions mutually, and feedback in groups in
1.Introduction order to achieve the individual and group targets. This
does not only benefit the individual students but also
In the traditional learning, the learning process of the benefits to the group members (Chen, 1997; Huang, 1996).
students is similar to photocopy. They catch the In the process of Cooperative Learning, all members are
knowledge transferred by teachers. The teachers teach in benefited from the groups, and the learning usually
one-way. Although the questions and suggestions from the happens when student are active and collaborate in solving
students are important, the influence is in one-way. The a problem in a social environment (Laurillard, 2002). In
teachers have been acting the models all the time. Starting fact, recent pedagogical research shows that learning is not
from the putting forward of the questions, analysis, simply knowledge assimilated with the help of a more
reasoning to answer, all of them are dominated and knowledgeable person or mediated by a computer system,
demonstrated by the teachers and the students are passive. but also jointly constructed through solving problems with
They observe and simulate the contents taught by the peers by a process of building shared understanding
teachers and the learning results are not good. Under the (Scardamalia & Bereiter, 2003). It is a kind of life
traditional learning, the students are short of chance for community and is a learning trend that is suitable for team
discussion, communication and argumentation with cooperation and innovation in the existing high-tech
colleagues. Therefore, they have no idea on how the others society.
consider on solving the same problem, how they can
clarify the own views through the interaction with others There is a series of innovative teaching methods applied in
and how the others organize the knowledge and concept educational domain recently. The innovative methods
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 9
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

develop the education to reach a new era. The results from 2. Embedded System and Applications
many researches (Zhang, 2003; Lin, 2005; Neo, 2006) also
show that the teaching contents should interact with The "Embedded system" integrates the application of
students regardless of the multimedia network teaching or information software and hardware. The application of
the recent virtual reality auxiliary teaching system in order embedded system can be found everywhere, including the
to reach the best learning performance. The computer is life facilities such as mobile phone, electric toy and
not the machine for data analysis only. From the designed video-audio instrument, and transportation system and
digital teaching materials, it does not only promote the automation in factories. The trend of the functions of
interests of the learners but also improves the learning common embedded systems is simplification. The
efficiency. Furthermore, the learners all over the world can software and hardware only include the modules required
use the teaching materials easily through the network. for specific functions. The peripheral products from
Many domestic and foreign researchers on network different manufacturers are combined to become an intact
Cooperative Learning wish to establish a more effective embedded system according to the provided Intelligence
learning environment through the Internet in order to Property (IP). Comparing with the common computers, the
improve the learning ability and performance of the building cost can be reduced significantly. Besides, due to
students (Shi, 2002; Chen, 2004; Dong, 2004). The the domain of the application, most of them include the
existing =be classified into three types according to the characteristics of miniaturization and low-power
tools and technologies of design. The first type is the consumption. Since the system only includes the specific
network Cooperative Learning system that provides functions, the design of the system will be optimized to
documents mainly (Johnson, 1994;Maier, 1994). This ensure the stability of the system (Kao, 2007).
type of Cooperative Learning system discusses the design
of the shared document database and the system
infrastructure of the network Cooperative Learning. The 2.1 Development Approach of Embedded System
second type is the Cooperative Learning system that
provides video conferencing function. As per the name, it Since the embedded system is developed for specific
is suggested that the Cooperative Learning system mainly functions, the developing tools include In-Circuit
provides the mechanism for the members of the group to Emulator (ICE), Development Board, Integrated
discuss face-to-face on network. The third type of Development Environment (IDE) and compiler
Cooperative Learning system provides virtual reality (Microtime Computer, 2004). The Development Board is
environment (Maier, 1994). This type of Cooperative the design sample provided by the hardware manufacturers.
Learning system mainly provides a learning environment The debugging mechanism in the central processing unit
of simulated virtual reality to the members of the group. of the In-Circuit together with the pre-set messages of the
However, most of the existing network Cooperative compiler can simulate the program running on the
Learning system a convenient network environment for hardware. The product developers may refer to their
the students to conduct online Cooperative Learning of circuit design and integrate to the corresponding
theoretical courses. It is rarely to discuss how the developing environment. Therefore, most of the
Cooperative Learning environment of operation training of Development Boards have been built with many peripheral
online-simulated instruments and the circuit current function modules such as network card, seven-segment
measurement can be established. display, parallel port, LED and dip switch, for the
convenient of product development and speedup the
This paper has proposed a Web-based Cooperative product development cycle. The application integrating
Learning system for remote operation 3D virtual electronic the development environment and complier refer to the
instruments with circuit-measuring function. The system development software on the personal computer of the
does not only provide the cooperative operation practice of developer. Therefore, the system developers may design
simulated instruments on network for the learners, it also the working platforms and the corresponding applications
improves the interests of the learners in practical courses of the embedded system on the personal computer, and
and the learning performance through the system functions complete the compiling. Through the parallel port or serial
including real time text chatting and discussion, and actual port, the corresponding software can be embedded to the
circuit measurement. The chapters of this article also target board and the cross-platform development process is
include embedded system and applications, virtual reality, finished. Since the common low-cost multimeter cannot
and system infrastructure implementation, the Cooperative communicate with computer, this study has made use of
Learning of virtual instruments operation and circuit low-cost 8-bit 89c51 chip to develop Embedded RS-2332
measurement, and conclusions. Module to capture the data shown on the panel of the
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 10
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

multimeter and transfer it to the computer (as shown in

Figure 1).

Figure 2 89C51 block diagram

2.3Embedded Broker
Figure 1 Embedded RS-232 module
As the network technology and applications are developed
2.2 Embedded RS-232 Module rapidly, large amount of servers for different network
services are built. They include file server, web server and
Since most of the low-cost multimeters used by students Email server. These servers are built on high-end servers.
do not include the serial communication function with the According to the functions built by the users, these
computer, this study makes use of 8-bit 89C51 single chip powerful servers become the specific function servers. The
to implement an Embedded RS-232 Module. It is then hardware is very complicated and expensive, and it
integrated with the low-cost multimeter to replace the increases the building cost of the system directly. To build
high-cost multimeter. The 89C51 single chip includes the the network server by the embedded system, the relevant
following advantages: (1) the program is easy to learn (2) functions can be enhanced according to the requirement of
low-cost (3) simple circuit (4) small size. The the service. For example, the files operating performance
infrastructure of the single chip is shown in Figure 2. The and the space of the hard disks of the file server should be
corresponding specification is as follows: increased but the relatively low performance of CPU can
(1) CPU: the automatic controlled high performance 8-bit be used. This can reduce the cost effectively and increase
CPU is used to implement the whole operation of the the competitiveness. This paper has adopted ARM7
computer. development board (as shown in Figure 3) to develop
(2) Program memory: the ROM or EPROM is used to embedded Broker to replace the expensive server.
store the program and the constant numbers. Different
have different codes.
(3) Data memory: the RAM is used to store the data that is
dynamic during the running of the program. The data can
accessed in the memory with the CPU. The data will
disappear if the power is off.
(4) Oscillator: An oscillator is included in the MCS-51
series single chip. It will produce the clock for the whole
system if
it is connected with a quartz crystal.
Figure 3 ARM7TDMI development board
(5) I/O pins: there are 32 input / output pins to use.
(6) Timer / Counter: the commands are used to set the
16-bit Timer or 16-byte Counter.
3. Application of Virtual Reality Technology
The Virtual Reality (VR) is the term for science and
technology domain. The virtual reality includes many
components such as vision, hearing, and even sense of
touch and sense of smell. The users can determine the path
for browsing the information freely in the virtual reality.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 11
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Therefore, the virtual reality can be regarded as an intact real and smooth. The system has no strict limitation to the
multimedia system. The major characteristics of the virtual hardware and it supports most of the peripheral devices of
reality system are the interaction and the real time virtual reality. In addition, the SDK (Superscape
response. When it is applied on application, the users can Developers Kit) is applicable to compiling to the Dynamic
operate the computers freely. They can observe the Linking Libraries (.dll file) so that it can plug-in to the
products in any angle and position. The virtual reality virtual scene and develop the own drivers. The virtual
technology makes use of the computer drawing or image reality tool can develop online virtual reality and multiple
synthetic technology together with the virtual reality users’ connections on network. It can also process
constructed by sound processing, sense of touch and sense dynamic data link with other applications for information
of taste. In this virtual world, the 3D simulated object can sharing. It provides an intact virtual scene to the users and
be our famous things, or something that we cannot see, or they can browse the virtual scene from network after they
a simulated imagined space. Virtual reality is an integrated have installed ActiveX on the browser. This software
technology in order to provide a higher level of includes different editing tools, such as model editor,
man-machine interface (Kao, 2007). Dong (2004) has image editor, 3D scene and 3D virtual object editor. In
proposed three important topics for virtual reality. They order to increase the control of the 3D virtual object and
are immersion, interaction and imagination (3I). These flexibility of the interactive design of the 3D virtual object,
three topics are explained respectively as follows: the 3D Webmaster provides Superscape Control Language
(1) Immersion: (SCL), which is similar to programming language C.
The users can feel the scene of virtual reality directly Therefore, the users can write with the control language in
produced by computer. The display system provides a the virtual scene designed by themselves (Superscape,
simulated and First-person virtual world and scene. 1996) as shown in Figure 4.
Besides, it can be controlled by the users directly. The
users can really "immerse" in the world of virtual reality.
(2) Interaction
The users interact with the objects in the virtual reality
world. For example, the users can walk in the virtual
world, or wear the glove to catch the objects in the virtual
world, or even interact with other users in the same virtual
(3) Imagination:
The virtual reality provides imagining space to the
users. If the virtual reality technology is applied in the
interacting environment of the virtual group, then multiply
users can remote logon simultaneously, and share and
interact in the same space, and meet the characteristics of
virtual reality. This must bring better efficiency on
interaction and communication approach to the virtual
group. The users can talk and interact as if they are in the
real world. The application and development of virtual Figure 4 SCL program editor of 3D Webmaster
reality technology on different professional domains are
increasing gradually. One of the reasons is that the
function of the software of virtual reality is enhanced. It is 4. System Infrastructure and Implementation
more important that the cost for virtual reality on personal
computer (PC) is relatively low and it is widely accepted The network technology is developed rapidly. For
by all trades and professions. The 3D Webmaster and example, the web language starts from HTML, SMIL,
Virtools Dev 3.0 have been adopted in this study to VRML to XML. The media appeared starts from text,
develop the relevant virtual instruments. picture, and image to streaming media. The transmission
channel starts from modem, ISDN, T1 lease line to ADSL.
The 3D Webmaster does not only provide the editing The change causes the revolutionary development on
function for the 3D virtual object but it can also transform teaching method, teaching style and interactive mechanism.
the DXF or VRML2.0 format to other 3D format. It can In order to recognize the characteristics of Cooperative
also export to VRML2.0 format and the audio file is Learning, the system of this study provides the functions
mainly in wav format. The Direct 3D is supported and Z including group real time discussion, real time interaction,
buffer is used for image processing so that the images look operation of 3D virtual electronic instruments that include
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 12
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

transmission technology of remote control parameters and 4.2 Cooperative Learning of 3D virtual instruments
circuit measurement. with circuit- measuring function
The proposed cooperative practice environment with
4.1 System infrastructure circuit measurement function sends the measured value
from the multimeter Embedded RS-232 Module to the 3D
The functions of the system are divided into four major virtual multimeters of the learners in the same group. Any
parts as shown in Figure 5. learner can operate the own multimeter to progress
(1) Cooperative Learning system: the learners can cooperative practice. As shown in Figure 6, when the user
download the installation file from the link of the portal sends the control command (or get the data from circuit
website. After the installation, the Cooperative Learning measurement) to the 3D virtual instrument through the
system can be entered. Operating Interface, it will be determined and processed
(2) Cooperative Digital Learning: in the Cooperative by the embedded control program and a control parameter
Learning system, many items of Cooperative Digital will be obtained. Then with the remote control parameter
Learning (system brief, 3D virtual lab, 3D puzzle that transmission technology, the embedded Broker transmits
trains up initiative and interaction, 3D virtual multimeter the control parameter to other learners of the same group
and 3D virtual power supply) can be found. The learners so that the 3D virtual instruments of the learning partners
can click and start Cooperative Digital Learning. can be updated simultaneously. The detailed transmission
(3) Teaching material: after the learners have started procedures of the remote control parameter are as follows:
the Cooperative Learning system, they can start learning (1) The user sends the control command (or receives the
from the digital teaching content. The users must study the measured data of the circuit) on the Operating Interface.
operation of the electronic instruments first. Then they can The obtained control parameter will be sent to 3D virtual
make use of the real time discussion to practice operating instrument first and responded properly in accordance with
the instruments and start Cooperative Learning with the the control parameter.
group members. (2) The control parameter is sent to the embedded
(4) Circuit measurement: the learners of the same TCP/IP control function.
group can progress Cooperative Learning of multimeter (3) The TCP/IP control function sends the control
with embedded RS-232 module for circuit measurement. parameter to the Embedded Broker.
The members of the same group can see the data of the (4) Embedded Broker transmits the control parameter to
circuit obtained by the measuring team from the 3D virtual the TCP/IP control functions of other users.
instrument simultaneously. (5) Other users will send the received control parameter
to the embedded control program.
(6) The embedded control program from other users
then sends the control parameters to the 3D virtual
instrument, and it responds properly in accordance to the
control parameter.

Figure 5 System infrastructure diagram

Figure 6 System operation procedures

IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 13
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

In this way, the users can interact with the 3D virtual

instruments and the display on the instruments can be
synchronized among the users in the same group. The
Embedded Broker is the backend of the whole system.
When Embedded Broker is started, the main duty is to
listen to the status of each user constantly. Once the status
of a user is changed, Embedded Broker will receive the
status changing information of the users. It will determine
the groups of the learner immediately, and copies and
sends the received information to the members of the same
group in order to progress real time update of the status.
Thus, the simulated operation of the instruments and the
measured data of the circuit among the members in the
same group will show the same result as shown in Figure 7
and Figure 8-10.

Figure 8 Voltage value shown from the 3D virtual instruments of the

members in the same group

Figure 7 Circuit measurements by the multimeter

IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 14
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Figure 10 Diode measurement shown from the 3D virtual instruments of

Figure 9 Resistance value shown from the 3D virtual instruments of the the members in the same group
members in the same group

4.3 Proposed 3D virtual instrument for Cooperative

learning of simulated operation training
The designed 3D virtual instrument interface for
Cooperative Learning of simulated operation training is
shown in Figure 11. The functions of the system include
login, group study, 3D virtual object operating interface
and real time discussion. The corresponding function
blocks are described below:
(1) The interactive 3D Virtual Instrument
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 15
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

(2) In the real time discussion, the users can transmit utilized: Ethereal (Network Protocol Analyzer, version
messages to the members of the group for real time 0.10.10).
discussion. The corresponding learning records can be In Figure 12 and 13, the x-axis stands for the number of
saved for the teacher assessing the learning performance. learners while the average response time for 3D virtual
(3) User inputs the IP of the Embedded Broker and the cooperative learning system (unit: ms, millisecond) is
assigned Port in order to connect to the Embedded Broker. indicated by the y-axis. While 40 learners are using the
(4) After grouping, team study via networking available system simultaneously, the response time for each learner
in addition to login learning. will be different because the received status data
(5) The users can make use of the virtual machine calculated by averaging the sum of each learner’s response
operating interface to operate the 3D virtual instrument. time are orderly updated via Embedded Broker into each
learner’s PC. Under PC and Embedded Broker
Architectures, the experimental results for the system’s
➂ response time against increasing the number of learners
are listed below:
(1) Under the system with PC Broker, while learning is
➃ done by grouping, the average response times for Group 1,
2, 3, and 4 are respectively “132ms,” “144ms,” “148ms,”
and “152~155ms,” as illustrated in Figure 12.
➀ ➄ (2) Under the system with Embedded Broker, while
learning is done by grouping, the average response times
for Group 1, 2, 3, and 4 are respectively “130ms,”
“141ms,” “147ms,” and “152~155ms,” as illustrated in
Figure 13.
From the above results, the proposed embedded Broker
➁ would also provide service capabilities similar to that of a
high-end server system, and significantly reduce the
system costs.

Figure 11 3D instruments in cooperative learning system

5. Performance Evaluation and Statistic

In the experiment, we adopts as the setting 40 learners
with 3 users per group with 4 users in the 13th group, and
connect into Embedded Broker to do cooperative learning
by means of the cooperative learning materials offered by
the system.

5.1 System Performance Testing

The hardware specification used by the 40 learners is Intel Figure 12 Average response time via PC Broker’s group learning
Pentium 3 Processor as CPU, main memory 512 MB, and
Microsoft Window 2000 Professional Operating System.
Learners can be allowed to make use of the Internet
browser and go into the portal to download and install the
system. The testing is focused on the performance of
getting curriculum materials against increasing the number
of user connections (one group per unit) to further analyze
the transmission efficiency reached by the system. In order
to retrieve the packets communicated between Embedded
Broker and each learner’s PC for analysis of data
transmission, a professional network packet analyzer is
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 16
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Table 1 Statistic analysis

Questionnaires Average Total
Score Average
1. The operations of virtual 3.5
instruments really Helps
the understanding of real
2. The operation of the 3D
cooperative type 3.26 3.80
Instruments can really
promote the positive
Interdependence attitude
each other among the
Numbers of group.
Figure 13 Average response time via embedded Broker’s group learning 3. The operation of the 3D
cooperative type 3.52
5.2 Questionnaire design and statistic analysis Instruments can promote
the learning interest And
The design of the questionnaire is using Likert-type scale. learning efficiency.
Each question have five options from not very agree, not 4. The appearance design of 3.63
agree, normal, agree, to strongly agree. The best result is 5 the 3D cooperative Type
points, and the lowest is 1 point. The learners can choose instruments is very
the situation that they feel from question that they describe. suitable for learning.
The purpose of this survey is to analyze the actual use with 5. It is very simple that the 4.22
co-operative practice system, and then make an cooperative operation of
assessment of learners progress the curriculum. We use instruments with
online Assessment methods, and let 40 learners fill in the embedded RS-232
survey online after they use the cooperative learning module.
system that is proposed by this research. From collected 6. The operating interface is 4.08
questionnaire data, the experts can analyze each question suitable and easy use for
to related statistical results such as Table 1. From Table learners.
1, the average grade is between 3.26 to 4.36. The statistic 7. The cooperative circuit 4.36
results show that the design of cooperative circuit measuring of real VOM
measuring really improve the learning efficiency, and instrument can really
almost all students think that the hardware design of improve the learning
instrument with embedded RS-232 module is easy to efficiency.

6. Conclusions
Most of the existing network Cooperative Learning system
provides a convenient network environment for the
students to conduct online Cooperative Learning of
theoretical courses. It is rarely to discuss how the
Cooperative Learning environment of operation training of
online-simulated instruments and the circuit measurement
can be established. This causes the existing network
Cooperative Learning system cannot implement the
characteristics of Cooperative Learning in the learning
environment on the Internet.

This study has broken through the traditional "lab

operation” practice mode. The Cooperative Learning
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 17
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

environment of the instrument practice is implemented in [7] Kao, F. C. (2007): Design of 3D Virtual Instrument for
the Internet. The proposed Cooperative Learning system Cooperative Learning System. TWELF2007, 602-606.
of 3D virtual instruments with circuit-measuring function [8] Kao, F.C., Chiang, K.Y. & Kuo, C.L. (2006). The Design of
integrates with the Virtual reality technology, embedded Load-Balancing Computer- Assisted Instruction System with
Embedded 3D Virtual Instrument s. ICIMB, 30.
system technology and the transmission technology of the [9] Kao, F.C., Tseng, C.W. & Ji, J.H. (2007). The Design of
remote control parameter. The Cooperative Learning Embedded LCMS Broker with Load- Balancing Function.
environment for operation of instruments on Internet is ICMLC, 3770-3776.
established completely. It provides the Cooperative [10] Lin, D. S. (2005). Studying on Applying Conceptual
Learning practice of circuit measurement for the learners Graph to Junior High Students Cooperative Learning.
to operate online anytime. Besides, the simulating training Nanhua General Education Research, Vol. (2), 39-67.
provided by the 3D virtual instruments, and the online [11] Liu, X. M. (1998). Teaching Strategy of Cooperative
grouped discussion do not only increase the interests of the Learning. Bulletin of Civic and Moral Education, Vol. (5),
learners in practice courses and train up their initiative and 285-294.
[12] Laurillard. D.(2002). Rethinking University Teaching, 2nd
interaction on Cooperative Learning. This also decreases ed: Routledge Falmer.
the damage rate of the traditional instruments and reduces [13] Maier, M. H. & Keenan, D. (1994). Teaching Tools
the purchasing cost of school facilities. Besides, the Cooperative Learning in Economics. Economic Inquiry,
proposed embedded Broker would also provide service Vol.32, 358-361.
capabilities similar to that of a high-end server system, and [14] Microtime Computer Inc., "Implementation of Embedded
significantly reduce the system costs. uClinux on PreSOCes", Chuan Hwa Book Co., Ltd, ISBN
[15] Neo, S. J. & Yang, G. X. (2006). The History and Existing
Acknowledgment Development Analysis of Network Education.
This paper has been supported in part by the grant [16] Parker, D.B. (1985). Learning Logic. Technical Report
NSC95-2520-S-212-001-MY3 from the National Science TR-47, Center for Computational Research in Economics
and Management Science, Massachusetts Institute of
Council of Taiwan.
Technology, Cambridge, MA, USA.
[17] Rosenberg, M.J. (2001). E-Learning: Strategies for
References Delivering Knowledge in the Digital Age. New York,
[1] Chen, S.J. (1997). Discussion on Applying Coaching McGraw-Hill.
Cooperative Learning Strategy on Reading Teaching in [18] Scardamalia, M and Bereiter, C. (2003). Computer
Primary School. Journal of National Taichung Teachers Support for knowledge-Building Communities. The Journal
College, Vol. (11), 65-111. of the Learning Science, vol.3, pp. 265-283.
[2] Chen, X. Z., Liu, J. R., & Ke, J. J. (2004). Performance [19] Shi, D. Q. (2002). Discussion on Improving the Reading
Assessment for Cooperative Learning on Information Ethics. and Understanding Ability from the Cooperative Learning
Journal of Kao Yuan University, Vol. (10), 161-168. Group Discussion. Practice Teachers Quarterly, Vol. (24),
[3] Dong, W. F. (2004). Applying Cooperative Learning on 70-76.
Teaching the English Class. [20] Slavin. R.E. (1985). Learning to cooperate, cooperating to
[4] Huang, Z. J. & Lin, P. X. (1996). Cooperative Learning. learn. 147- 172.
Wu Nan Book Inc. [21] Wei, J. L. & Chen, Y. M. (1994). The Method Breaking
[5] Johnson, D.W. & Johnson. R.T. (1994). The new circles of Through the Teaching Predicament of Normal Class:
learning cooperation in the classroom and school, American: Cooperative Learning. Educational Research & Information,
ASCD. Vol. (35), 59-65.
[6] Kao, F. C. (2005). Design of SCORM Learning Management [22] Zhang, J. H. (2003). Enlightenment of Cooperative
System with Embedded 3D Virtual Instrument. OOTA2005, Learning Theory on Teaching the English Class. Journal of
286-290. Liaocheng University, Philosophy and Social Science Edition,
Vol. (5), 120-122.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 18
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Implementation of an Innovative Bio Inspired GA and PSO

Algorithm for Controller design considering Steam GT Dynamics
R.Shivakumar and Dr. R.Lakshmipathi
Department of Electrical and Electronics Engineering
Sona College of Technology, Salem-636005, INDIA

Department of Electrical and Electronics Engineering
St.Peters Engineering College, Chennai, INDIA.

Abstract Systems [2-3].Designing and applying the PSS is a

The Application of Bio Inspired Algorithms to complicated complex phenomenon. In recent years, several approaches
Power System Stability Problems has recently attracted the based on Modern control theory have been applied to PSS
researchers in the field of Artificial Intelligence. Low frequency design problem. These include optimal control, adaptive
oscillations after a disturbance in a Power system, if not control, variable structure control and intelligent control
sufficiently damped, can drive the system unstable. This paper
provides a systematic procedure to damp the low frequency
oscillations based on Bio Inspired Genetic (GA) and Particle
Swarm Optimization (PSO) algorithms. The proposed controller Despite the potential of modern control techniques, Power
design is based on formulating a System Damping ratio System utilities still prefer the conventional lead lag PSS
enhancement based Optimization criterion to compute the structure [6].The parameters of CPSS are based on a
optimal controller parameters for better stability. The Novel and linearized model of the Power System. Since modern
contrasting feature of this work is the mathematical modeling power systems are dynamic and non linear in nature,
and simulation of the Synchronous generator model including the CPSS performance is degraded whenever the operating
Steam Governor Turbine (GT) dynamics. To show the robustness point changes from one point to another because of the
of the proposed controller, Non linear Time domain simulations
fixed parameters of the stabilizer. Unfortunately, the
have been carried out under various system operating conditions.
Also, a detailed Comparative study has been done to show the conventional techniques are time consuming, as they are
superiority of the Bio inspired algorithm based controllers over iterative and require heavy computation burden and slow
the Conventional Lead lag controller. convergence.
Keywords: BioInspired Algorithms, Power System Optimization,
Genetic Algorithm, Low frequency Oscillations, Particle Swarm Recently, Bio Inspired optimization techniques like
Optimization. Genetic Algorithm, Evolutionary Programming, Simulated
Annealing, Bacteria foraging, Particle Swarm optimization
etc have been applied for PSS parameter optimization [7-
1. Introduction 9].In this work, Genetic algorithm and Particle Swarm
Optimization algorithms has been implemented for
Modern Bio Inspired Algorithms include a wide variety of computing the parameters of the optimal controller
Population based algorithms which can be applied to including Steam Governor Turbine dynamics for Power
various Power System Optimization problems. The System Stability.
Phenomenon of stability of modern interconnected power
systems has received a great deal of attention in recent The Main Objectives of this work are summarized as
years. One problem that faces power systems nowadays is follows:
the Low frequency oscillations arising from (1).To develop a linearized State Space model of the
interconnected Power Systems [1]. Sometimes, these SMIB model (including Steam Governor Turbine
oscillations sustain for minutes and grow to cause system dynamics) with and without the damping controller.
separation, if adequate damping is not provided. A cost
efficient and satisfactory solution to the problem of low (2).To formulate a Damping ratio enhancement based
frequency oscillations is to provide damping by optimization criterion to compute the Optimal PSS
implementing Power System Stabilizers (PSS), which are parameters.
supplementary controllers in the Generator Excitation
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 19

(3).To compare the computed damping ratios of the

Conventional PSS (CPSS), Genetic based PSS (GAPSS)
and Particle Swarm based PSS(PSOPSS) for damping the
poorly damped Electromechanical modes of oscillation.

(4). To carry out Parameter Sensitivity analysis by

performing Non linear Time domain based simulation to
validate the robustness of the proposed controllers under
wide variations of system operating conditions and also
under variations in system parameter involved in the

2. Modeling of Power System including GT

2.1 System Model under Study Fig.1. Heffrons-Phillips SMIB Model with PSS

Fig.1.shows the Heffron Phillips block diagram [10] of

Single Machine Infinite bus model (SMIB) equipped with
PSS. EXC(s) in the Heffron model represents the IEEE
Type 1 Excitation system involving Amplifier, Exciter and
Rate Derivative feedback compensation as in Fig(2).

In all the classical model analysis for SMIB system, the

mechanical power input remains constant during the
period of the Transient (i.e) Governor Turbine Dynamics
is not included in the modeling and analysis. But in this
work, the mechanical power input is included in the
Modeling and Simulation. PSS in the model represents the
Power System Stabilizer with ∆ω as the input and output Fig.2. IEEE type 1 Excitation System Model
is ∆UE given to the Generator Excitation System summing

Fig3.represents the Steam Governor Turbine Model with

Time Constants TRH, TCH, and TGV.
Where TRH = Reheater Time Delay.
TCH = Inlet and Steam Chest Delay.
TGV = Main gate Servo Motor Time Constant
RP = Steady State Speed Droop.
FHP = High Pressure (HP) flow fraction
Fig.3. Steam Governor Turbine Model
In this work, Non Reheat type Steam Turbine is used in
the modeling and simulation.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 20
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

For a Non Reheat Steam Turbine, TRH =0, where

TRH = Reheater Delay (Typically 6 secs). The Transfer function of the PSS Model is given by
Hence the model in Fig (3) is simplified to the Model as in ⎡∆U E ⎤ ⎡ (1 + s T 1) ⎤ ⎡ (1 + s T 3)⎤ ⎡ (s T w ) ⎤ (4)
⎥ = Ks ⎢ ⎥⎢ ⎥⎢ ⎥
⎣⎢ (1 + s T 2 )⎦⎥ ⎣⎢ (1 + s T 4 )⎦⎥ ⎣⎢ (1 + s T w )⎦⎥

Fig (4). ⎣ ∆ω ⎦

In Fig (4), TGS = Steam Governor Time Constant Where Ks = PSS gain
TTS = Steam Turbine Time Constant. Tw = Washout Time Constant
T1,T2,T3,T4 = PSS Time Constants
Fig(4) represents the model for Steam Governor and
Time Constants T1=T3, T2=T4
Turbine(Non reheat type) to be equipped with the Heffron
(Identical Compensator Block).
Phillips Generator model.(i.e) Output(∆Tm) of Steam GT
Model is given as input to the Heffron-Phillips generator
Hence Ks, T1, T2 are the PSS parameters which should be
computed using Conventional Lead Lag stabilizer and
All the abbreviations for the Constants and Variables
optimally tuned using GAPSS and PSOPSS. The washout
involved in the model are given in Appendix-II.
time constant (Tw) is in the range of 1 to 20 seconds and
in this work, Tw is taken as 10 seconds.

3. Proposed Optimization Criterion for


3.1 Criteria for Damping

The Rate of Decay of amplitude of oscillations is best
expressed in terms of the Damping ratio (ξ).For an
Fig.4. Steam Governor-Non Reheat type Turbine Model
Oscillatory mode represented by an Complex Eigen value
The Dynamic Model in state space is given by (σ ± j ω), the Damping ratio is given by
. [ξ ] = −2 σ 2
x = Ax + Bu (1) σ +ω
The damping ratio of all system modes of oscillation
where [x] = Vector of State Variables should exceed a specified value. In Power Systems,
A,B = State Matrix and Input Matrix respectively. Electromechanical oscillations with damping ratios more
In this work, for Open loop 8 state variables and for than 0.05 are considered satisfactory.
Closed loop (with PSS) 11 state variables are used in the
modeling. 3.2 Proposed Optimization Criterion
The Main objective of this formulation is to compute the
[x]open = [∆ω, ∆δ, ∆Eq’, ∆EFD, ∆VR, ∆VE,
optimal value of PSS parameters for system oscillations
∆PG, ∆Tm]T (2)
damping. A good measure of system damping is the
damping ratio or damping factor. Hence a Damping Ratio
[x]Closed = [∆ω, ∆δ, ∆Eq’, ∆EFD, ∆VR, ∆VE,
enhancement based objective function is selected for PSS
∆PG, ∆Tm, ∆P1, ∆P2, ∆UE ]T (3)
parameter optimization.
All the Test System parameters used for simulation
[11] are given in Appendix-I.
[ J ] = min ξ i ( ) ,ξ εξ i EM

where ξi = Damping Ratio of ith Electromechanical

2.2. Structure of PSS Mode of Oscillation.

The PSS model consists of the Gain Block, Cascaded ξEM = Damping Ratios of all the
identical Phase Compensation block and the washout Electromechanical modes of
block. The input to the controller is the Rotor speed Oscillation.
Deviation (∆ω) and output is the Supplementary Control
signal (∆UE) given to Generator excitation system.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 21

The Objective here is to Maximize J, in order to enhance m in

≤ ≤
m ax
the Damping Ratio of the poorly damped modes of T 2 T 2 T 2

oscillation for better Stability.

The poorly damped electromechanical modes of Typical ranges selected for Ks, T1 and T2 are as follows:
oscillation will have its Eigen values located in right half For Ks [5 to 60], for T1 [0.1 to 1] and for T2 [0.1 to 1].
of complex s plane, thus making the system Unstable. This Maximization criterion has been implemented in this
The above criterion formulation is represented in simpler work to compute the Optimum Value of the PSS
form as: Parameters Ks, T1 and T2.
Maximize [J] such that

(ξ )⎦⎥⎤ ⎯⎯⎯→ ⎣⎢⎡ξ ⎦⎥⎤ ,ξ ≥ ξ

⎡ min
4. Bio Inspired Algorithms
⎣⎢ i sm sm T

4.1 Genetic Algorithm- An Overview.

Where ξT = Threshold Level of Damping ratio
for System Stability. Genetic Algorithms are numerical optimization algorithms
min (ξi) = Minimum Damping ratio value inspired by Natural selection and Natural Genetics [12-
among the Electromechanical 13].GA techniques differ from more traditional search
Modes of Oscillation. algorithms in that they work with a number of candidate
ξsm = Required Damping ratio for solutions rather than one candidate solution. Each
Stability with m = 1, 2, 3. candidate solution of a problem is represented by a data
structure known as an individual. A group of individuals
Here m represents the method used for optimization. (i.e) collectively comprise what is known as a population. GAs
m=1 indicate CPSS, m=2 indicate GAPSS, m=3 indicate are initialized with a population of random guesses. GA
PSOPSS. includes operators such as Reproduction, Crossover,
Mutation and Inversion.

Reproduction is a process in which a new generation of

population is formed by selecting the fittest individuals in
the current population. Crossover is responsible for
producing new offsprings by selecting two strings and
exchanging portions of their structures. The new
offsprings may replace the weaker individuals in the
Mutation is a local operator which is applied with a very
low probability of occurrence. Its function is to alter the
value of a random position in a string.
Finally, Inversion is a process which inverts the order of
Fig.5. Pictorial Representation of the proposed
the elements between two randomly chosen points on the
Objective function. string.
The Algorithmic Steps involved in Genetic Algorithm are
In Fig.5, Damping ratio corresponding to min(ξi) location as follows:
is to be shifted to locations beyond ξT for system stability.
Let it be ξS1, ξS2, ξS3 for the three techniques implemented Step 1. Specify the various parameters for GA
in this work (CPSS,GAPSS, PSOPSS). Optimization.

The Design problem including the constraints imposed on Step 2. Create an Initial Population of
the various PSS parameters is given as follows: individuals randomly.

Optimize J Step 3. Evaluate the Fitness of each individual

Subject to (i.e) Evaluating the optimization criterion J.
m in m ax
     ≤ ≤                                         (8)
Step 4. If value of J obtained is minimum, then
m in m ax
≤T ≤T (9)
T 1 1 1 Optimum value of PSS parameters is equal
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 22
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

to those obtained in current generation, The velocity of each agent can be modified by the
Otherwise Goto step 5. following equation
Step 5. Based on the fitness, select the best k +1
= W .V i + C 1.rand 1 * ⎛⎜ Pbest i − S i ⎞⎟ +
k k

Individuals and perform recombination Vi ⎝ ⎠

through a crossover process. * ⎛⎜ g − S i ⎞⎟
C . rand
⎝ best

Step 6. Mutate the new generation with a given Where Vik = Velocity of agent i at iteration k.
Probability. W = Weighting Function.
Cj = Weighting factor.
Step 7. If termination condition (Maximum no rand = random number between 0 and 1.
of Generations) is not reached, go back Sik = Current position of agent i at
to step (3). iteration k.
Pbest = Pbest of agent i.
gbest = gbest of the group.
4.2 Particle Swarm Optimization- An Overview.
The following Weighting Function is usually utilized in
PSO is an Evolutionary Computation Technique equation (11).
developed by Eberhart and Kennedy [14-15] in 1995, ⎡ ⎤
which was inspired by the Social behavior of Bird W =[ ]− W max −W min * iter
W max ⎢

flocking and fish schooling. PSO has its roots in artificial ⎣⎢ iter max ⎦⎥
life and social psychology as well as in Engineering and
Computer science [16-17].It is not largely affected by the where Wmax = Initial Weight
size and Non linearity of the problem and can converge to Wmin = Final Weight.
the optimal solution in many problems where most itermax = Maximum Iteration number
analytical methods fail to converge. iter = Current iteration number.

Particle Swarm Optimization has more advantages over The Current position can be modified by the
Genetic Algorithm as follows: following equation

k +1 k +1
(a). PSO is easier to implement and there are fewer k
= S i +V i (13)
parameters to adjust. Si
(b). In PSO, every particle remembers its own
previous best value as well as the
neighbourhood best ; therefore, it has a more The Algorithmic Steps involved in Particle Swarm
effective memory capability than GA. Optimization Algorithm are as follows:
(c).PSO is more efficient in maintaining the
diversity of the swarm, since all the particles use Step 1: Select the various parameters of PSO.
the information related to the most successful
particle in order to improve themselves, whereas Step 2: Initialize a Population of particles with
in Genetic algorithm, the worse solutions are random Positions and Velocities in the
discarded and only the new ones are saved; (i.e) problem space.
in GA the population evolves around a subset of
the best individuals. Step 3 : Evaluate the Desired Optimization Fitness
Function for each particle.
PSO utilizes a population of particles that fly through the
problem space with given velocities. Each particle has a Step 4: For each Individual particle, Compare the
memory and it is capable of remembering the best position Particles fitness value with its Pbest. If the
in the search space ever visited by it. The Positions Current value is better than the pbest value,
corresponding to the Best fitness is called Pbest and the then set this value as the Pbest for agent i.
overall best out of all the particles in the population is
called gbest. Step 5: Identify the particle that has the Best Fitness
At each iteration, the velocities of the individual particles Value. The value of its fitness function is
are updated according to the best position for the particle identified as gbest.
itself and the neighborhood best position.
Step 6: Compute the new Velocities and
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 23

Positions of the particles according to Here e(t) refers to the error involving Rotor Speed
equation (11) & (13). deviation (∆ω) and Power Angle deviation (∆δ).Ts
represent the Time of Simulation.
Step 7: Repeat steps 3-6 until the stopping The State Space modeling of the SMIB model including
Criterion of Maximum Generations is met. steam Governor Turbine dynamics was performed and the
system open loop eigen values and damping ratios was
computed as listed in Table 1 and Table 2.The
Electromechanical modes of oscillation indicate that the
5. Simulation Results test system is unstable having positive real part eigen
values located in right half of the complex s plane.
For all the Computation, Simulation and Analysis of the Also, the time domain analysis involving Load change
results in this work, MATLAB 7.0 / SIMULINK platform disturbance (∆PL) in Fig (6) and Fig (7) reveal that the
was used. open loop system without PSS are having oscillatory
The Two main analysis involved in the simulation in this responses with huge overshoots and large settling times,
work are thus making the system unstable.
(1). Small Signal Stability Analysis.
(2). Non Linear Time Domain Analysis.

(1).Small Signal Stability Analysis.

Small Signal Stability is the ability of the Power

System to maintain synchronism when subjected to small
disturbances.In this work, the stability analysis is based on
computation of eigen values and damping ratios of the
system for open loop, with CPSS, GAPSS and PSOPSS
and its comparison.
The disturbances involved are variation in system
operating point with Load change disturbance, variation in
system parameters namely variation in line reactance,
variation in Amplifier Gain with respect to the normal
operating point.
Fig.6.Open Loop Speed Deviation Response at (0.4, 0.008),
The Eigen values located in the left half of complex s- ∆PL =0.1p.u Operating
plane will determine the stability of the system, whereas Condition.
the Eigen values located in right half of complex s-plane
will make the system unstable.The Damping ratios more
than the threshold value of damping will provide better
system damping.

(2).Non linear Time Domain Analysis.

This analysis is to show the effectiveness of the
proposed controllers in damping the low frequency
oscillations under wide variations in operating conditions
and system parameters.
The objective is to minimize the integral squared
error (ISE) involved in the system. The error refers to the
speed deviation (∆ω) and the power angle deviation (∆δ).

The Integral squared error is given by,

Fig.7.Open Loop Power Angle Deviation Response at (0.4, 0.008),
Ts ∆PL=0.1p.u Operating Condition
[ ISE ] = ∫ e
(t )dt (14)
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 24
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Implementation of CPSS, GAPSS and PSOPSS provide

the Optimal PSS parameters as listed in Table 2.
Table 1. Computed Eigen Values for Open Loop without PSS, GAPSS and PSOPSS

Eigen Values
S. Operating Open Loop CPSS GAPSS PSOPSS
No Conditions
without PSS
1 P = 0.4 -13.1440 -16.7162 ± j 7.1748 -15.3069 ± j 5.6824 -15.8761 ± j 6.8962
Q = 0.008 0.1218 ± j 5.452 -0.3956 ± j 8.6327 -0.7011 ± j 7.2045 -0.8895 ± j 8.4064
∆PL = 0.1 -6.5964 -0.8315 ± j 3.4119 -1.3476 ± j 3.9158 -5.8273 ± j 0.7754
p.u -3.2215 ± j 4.7092 -4.950 ± j 0.7994 -5.4757 ± j 1.0014 -2.1101 ± j 2.4145
-2.0101 -0.0500 -0.0501 -0.0504
-1.3302 -2.2377 -2.2702 -1.6176
-1.3260 -1.3303 -1.8079
2 P = 0.4 -13.1452 -16.6293 ± j 7.0840 -14.8144 ± j 5.1911 -15.7799 ± j 4.7389
Q = 0.06 0.1231 ± j 5.405 -0.4355 ± j 8.5255 -0.5502 ± j 6.7845 -0.6361 ± j 7.0777
∆PL = 0.2 -6.6024 -0.8501 ± j 3.4180 -1.6839 ± j 3.9568 -2.9594 ± j 2.6991
p.u -3.2247 ± j 4.7140 -4.9452 ± j 0.8198 -5.7962 ± j 0.9729 -0.0504
-2.0000 -0.0501 -0.0501 -1.6220 ± j 0.2209
-1.3292 -2.2343 -2.2542 -2.2438
-1.3233 -1.3364 -1.3253
3 P = 0.4 -13.0586 -17.1748 ± j 7.6603 -0.3440± j 5.8868 -16.7640 ± j 7.1869
Q = 0.06 0.1272 ± j 5.6292 -0.2107 ± j 9.1457 -12.1442 ± j 1.6734 -0.5372 ± j 8.7165
∆PL = 0.3 -6.5683 -0.7241 ± j 3.3516 -3.2217 ± j 4.0608 -0.7460 ± j 3.5708
p.u -3.2617 ± j 4.6158 -4.7499 ± j 0.6580 -9.0214 -4.8116 ± j 0.7338
+ 10 % increase -2.0552 -0.0500 -6.2420 -0.0500
in Line -1.3371 -2.2364 -0.0501 -2.2373
reactance Xe. -1.3329 -1.3637 -1.3334

Table 2. Computed Damping Ratios for Open loop without PSS, GAPSS and PSOPSS

Optimal Damping Damping Ratios of Poorly Damped Electromechanical

Controller Parameters Modes of Oscillation

Threshold Level of Damping Ratio (ξT)= 0.06

No Conditions [Ks,T1,T2] [Ks,T1,T2] [Ks,T1,T2] Without PSS
1 P = 0.4 6.1692 6.2634 52.1596
Q = 0.008 0.6707 0.4557 0.2353 -0.02233 0.04578 0.09693 0.105225
∆PL = 0.1p.u 0.1000 0.5823 0.5176

2 P = 0.4 6.2986 8.1981 43.2273

Q = 0.06 0.6487 0.3527 0.1605 -0.022769 0.051016 0.08083 0.089513
∆PL = 0.2 p.u 0.1000 0.8915 0.3724

3 P = 0.4, 5.1944 8.7912 14.9908

Q = 0.06, 0.8100 0.1359 0.4651 -0.022591 0.023032 0.05834 0.061513
∆PL = 0.3 p.u 0.1000 0.2346 0.1521
+ 10 % increase
in Line reactance
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 25

Table.3. Parameters selected for GA Implementation. better damping to the oscillatory modes. Though CPSS
and GAPSS provide good damping , the PSO based
GA Parameters controller (PSOPSS) provide better damping to the
Population Size 20 oscillatory modes, with damping ratios more than the
No of Generations 10 threshold level of damping (ξT=0.06) for all the conditions
Selection Operator Roulette Wheel Selection involved (last Column of Table 2).
Generation gap 0.9
Crossover Probability 0.95
Non linear Time domain simulations involving wide
Mutation Probability 0.10
variations in operating points and system model
Termination Method Maximum Generations
parameters have been carried out to show the robustness
Table.4. Parameters selected for PSO Implementation.
of the proposed controllers in damping the low frequency
PSO Parameters
Swarm Size 20
wmax ,wmin 1 , 0.5
C1 , C2 1.0 , 1.0
No of Generations 10
No of Variables 03
Termination Method Maximum Generations

Table 3, Table 4 represents the various parameters

selected for Bio Inspired algorithms implementation.
Parameter Sensitivity Analysis:

Parameter Sensitivity analysis refers to analyzing the Fig.8. Speed Deviation response for (0.4, 0.008), ∆PL=0.1p.u condition
System Stability performance, whenever the System
parameters involved in the System model and System
operating conditions are varied over a wide range with
respect to the normal operating point.

In this work, the system is subjected to wide variations as


(a).Wide variation in operating condition(P,Q) with

Load change disturbances (∆PL) introduced in
the system model.
(b). 10 % variation (increase) in Line Reactance with
respect to the normal operating point.
(c). 10 % variation (increase) in Amplifier gain KA
with respect to the normal operating point.

In Power Systems, Electromechanical oscillations with

damping ratios greater than 0.05 are considered
satisfactory [18].Based on this criterion, the desired
Fig.9. Power Angle Deviation response for (0.4, 0.008),∆PL=0.1p.u
threshold level of damping is taken as ξT = 0.06 in this condition with CPSS, GAPSS and PSOPSS
work. Fig (8) and Fig (9) shows the effectiveness of the Bio
inspired controllers in damping the low frequency
The Damping Ratios are computed for the poorly damped oscillations better than the CPSS for the operating
Electromechanical modes of oscillation using the optimal condition (P=0.4, Q=0.008, Load Change disturbance of
PSS parameters and eigen values, listed in Table 2.The 0.1 p.u).
Damping ratios calculated for poorly damped modes of
oscillation reveal that the proposed controllers provide
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 26
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Fig (10) and Fig (11) indicate the Speed deviation and compared to the CPSS and the Genetic based PSS
Power Angle response of the system with operating (GAPSS).
condition (P=0.4, Q= 0.06, ∆PL=0.2 p.u). In order to enhance the system stability, the CPSS,
GAPSS and PSOPSS reduces the oscillations overshoot
These responses reveal the dominance of the Bio Inspired and also make the oscillations to settle at a quicker settling
optimal damping controllers in damping out the low time. For instance, in Fig (13), the maximum overshoot is
frequency oscillations, in particular, the PSO based 0.06 p.u for CPSS, for GA based PSS it is 0.054p.u,
controller damp the oscillations with reduced overshoot whereas for PSOPSS, the maximum oscillation overshoot
and quick settling time compared to the CPSS and the is only 0.038 p.u. This shows the optimal tuning and
Genetic based PSS (GAPSS). effective damping exerted by the PSOPSS.

Fig.10. Speed Deviation response for (0.4, 0.06), ∆PL=0.2 p.u condition
Fig.12. Speed Deviation response for (0.4, 0.06), ∆PL=0.3 p.u and 10%
increase in Line Reactance(Xe) condition with CPSS, GAPSS

Fig.11. Power Angle Deviation response for (0.4, 0.06),∆PL=0.2 p.u

Condition with CPSS, GAPSS and PSOPSS.
Fig.13. Power Angle Deviation response for (0.4, 0.06), ∆PL=0.3 p.u and
10% increase in Line Reactance (Xe) condition with CPSS,
Fig (10) and Fig (11) indicate the Speed deviation and GAPSS
Power Angle response of the system with operating and PSOPSS.
condition (P=0.4, Q= 0.06, ∆PL=0.2 p.u).These responses These responses clearly indicate the damping phenomenon
reveal the dominance of the Bio Inspired optimal damping exerted by the various controllers. In all the cases, the Bio
controllers in damping out the low frequency oscillations, Inspired algorithms(GA and PSO) based controller
in particular, the PSO based controller damp the provide better damping to the Low frequency oscillations
oscillations with reduced overshoot and quick settling time compared to the conventional lead lag stabilizer,thus
enhancing Power system stability .
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 27

6. Conclusions ∆ω = Incremental Change in Rotor Speed

∆δ = Incremental Change in Rotor Power
This work provides a better and efficient solution to the Angle.
complicated Engineering optimization problem of ∆Eq’ = Incremental Change in Generator Internal
damping the low frequency oscillations, by implementing Voltage.
the GA and PSO algorithms, thus enhancing Power ∆EFD = Incremental Change in Generator Field
System Stability. Voltage.
∆VR = Incremental Change in Amplified
The following are the important features implemented Voltage.
satisfactorily in this work: ∆VE = Incremental Change in Rate Feedback
(a).State Space Modeling of the SMIB System including Compensation output voltage.
the Governor Turbine(Non Reheat Type) dynamics for SE(Efd) = Excitation Saturation Function.
analysis and Simulation. KF = Gain of Rate Feedback compensation
TF = Time Constant of Rate Feedback
(b).Implementation of Damping Ratio based optimization Compensation
criterion to compute the optimal PSS parameters based on KE = Gain of Exciter
Conventional, Genetic and PSO based algorithms. TE = Time Constant of Exciter
KA = Gain of Amplifier
(c).Carried out a detailed comparative study on the TA = Time Constant of Amplifier.
damping ratios for CPSS, GAPSS and PSOPSS to damp Tdo’ = Field Open circuit Time constant
out the poorly damped electromechanical modes of ∆P1 = Output State Variable of PSS Washout
oscillation. Block.
∆P2 = Output State variable of first PSS Phase
(d).Implementation of Parameter Sensitivity analysis to Compensation block
validate the robustness of the proposed controllers by ∆PG = Incremental Change in Generation (Output
performing Non linear Time domain based simulations State Variable of Steam Governor)
under wide system loading conditions and also under ∆UE = Supplementary Excitation signal from
various system parameter variations. PSS.
K1-K6 = K Constants/ Coefficients involved in
Heffrons Phillips Model.
Appendix – I ξ = System Damping ratio
ξT = Threshold level of Damping ratio.
Test System Parameters σ = Real part of Eigen value
ω = Imaginary part of Eigen Value
Generator : xd = 0.973, xd’=0.190, xq = 0.550, Xe = Transmission Line Reactance
D=0, M= 9.26, Tdo’ = 7.76 secs ∆PL = Load Change Disturbance
Excitation : IEEE ST1A type Excitation M = Inertia Constant
(For Speed input Stabilizer) D = System Damping
KA = 190, TA =0, KF = 0,TF =1Sec Tw = Washout Time Constant
Ks = Gain of Power System Stabilizer
Line and Load: R = 0.034, Xe = 0.997, G = 0.249, ISE = Integral Squared Error.
B= 0.262, Vto = 1.05, = 0.4, Q = 0.008. CPSS = Conventional Lead Lag Stabilizer.
Governor and GAPSS = Genetic Algorithm based PSS.
Turbine : Steam Type PSOPSS= Particle Swarm based PSS.
TRH=0 (Non Reheat Type), RT = 0.4, SMIB = Single Machine Infinite Bus System.
RP = 0.05, TGS = 0.2, TTS = 0.3, VT = Generator Terminal Voltage
TCH = 0.3 Secs. ∆Tref = Incremental Change in Reference Torque
All Parameters are in p.u unless specified otherwise. References
[1] K.Sebaa, M.Boudour,. “Optimal Locations and Tuning of
Robust Power System Stabilizer using Genetic Algorithm”,
Appendix –II Electric Power Systems Research, Vol.79, 2009,pp.406-419.
[2] H.Shayeghi, A. Safari and H.A.Shayanfar, “Multimachine
Power System Stabilizers Design using PSO Algorithm”,
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 28
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

International Journal of Electrical Power and Energy Assistant Professor in Electrical and Electronics Engineering
Systems Engineering”1:4,Fall 2008,pp.226-233. Department. He is now working towards his PhD in Computational
[3] C.Y.Chung,K.W.Wang,C.T.Tse,X.Y.Bian and A.K.David, Intelligence at Anna University, Chennai. His areas of research
include Artificial Intelligence, Optimization Techniques, FACTS
“Probabilistic Eigen Value Sensitivity Analysis and PSS and Power System Stability Analysis. He is an annual member of
design in Multimachine systems”, IEEE Transactions on IEEE and Life member of ISTE.
Power Systems, Vol.18,No.4,2003,pp.1439-1445.
[4] D.K.Chaturvedi, O.P.Malik and K.Kalra, “Experimental Dr.R.Lakshimpathi received the B.E degree in 1971 and M.E
studies with a Generalized Neuron based Power System degree in 1973 from College of Engineering, Guindy, and
Stabilizer”, IEEE Transactions on Power Systems, Vol.19, Chennai.He received his PhD degree in High Voltage Engineering
from Indian Institute of Technology, Chennai, India. He has 36
No.3, Aug 2004, pp.1445-1453. years of teaching experience in various Government Engineering
[5] R.Segal,M.L.Kothari and S.Madani, “Radial Basic function Colleges in Tamilnadu and he retired as Principal and Regional
(RBF) network adaptive Power System Stabilizer”, IEEE Research Director at Alagappa Chettiar College of Engineering
Transactions on Power Systems,Vol.15,May 2000,pp.722- and Technology, Karaikudi.He is now working as Professor in
727. Electrical and Electronics Engineering department at St.Peters
[6] M.J.Gibbard, “Robust Design of Fixed parameter Power Engineering College, Chennai.His areas of research include HVDC
Transmission, Power System Stability and Electrical Power
System Stabilizer”, IEEE Transactions on Power Systems, Semiconductor Drives.
Vol.6, 1991, pp.794-800.
[7] M.A.Abido, “Robust design of Multimachine Power
System Stabilizers using Simulated Annealing”, IEEE
Transaction on Energy Conversion, vol.15, No.3, 2003,
[8] M.A.Abido,Y.L.Abdel Magid, “Optimal Design of Power
System Stabilizers using Evolutionary Programming”,IEEE
Transactions on Energy Conversion,vol.17,No.4,2002,
[9] S.Mishra, M.Tripathy, J.Nanda, “Multimachine Stabilizer
design by rule based Bacteria Foraging”, Electric Power
Systems Research, Vol.77, 2007, pp.1595-1607.
[10] Bikash Pal, Balarko Chaudhuri, “Robust Control in Power
Systems”, Springer Series, 2005.
[11] Yao Nan Yu, “Electric Power System Dynamics”,
Academic Press, Newyork, 1983.
[12] A.Andreoiu, K.Bhattacharya, “Robust Tuning of PSS
using a Lyapunov method based Genetic Algorithm”,IEE
proceedings on Generation, Transmission and Distribution
,Vol.149,No.5,Sep 2002,pp.585-592.
[13] K.Sebaa, M.Boudour, “Optimal locations and Tuning of
Robust Power system Stabilizer using Genetic Algorithms”,
Electric Power Systems Research, vol.79, 2009,pp.406-419.
[14] J.Kennedy and R.Eberhart, “Particle Swarm Optimization”,
in proceedings, IEEE International Conference on Neural
Networks (ICNN), vol.4, Nov 1995, pp.1942-1948.
[15] R.Eberhart and J.Kennedy, “A New optimizer using
Particle Swarm theory”, in proceedings,6th international
symposium on Micro Machine and Human
Science(MHS),Oct 1995,pp.39-43.
[16] Jong Bae Park,Ki Song Lee, Joong Rin Shin and Kwang
Y.Lee, “A Particle Swarm Optimization for Economic
Dispatch with Non Smooth Cost functions”, IEEE
Transactions on Power Systems,vol.20,No.1,Feb 2005,pp.34-
[17] M.A.Abido, “Optimal Design of Power System Stabilizers
using Particle Swarm Optimization”, IEEE Transactions on
Energy Conversion, vol.17, No.3, Sep 2002, pp.406-413.
[18] Graham Rogers, “Power System Oscillations”, Kluwer
Academic Publishers, USA, 2000.

R.Shivakumar received the M.E degree in Power System

Engineering from Annamalai University, India in 1999. Currently he
is working in Sona College of Technology, Salem, India as
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 29
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Significant Interval and Frequent Pattern Discovery in

Web Log Data
Dr. Kanak Saxena1 and Mr. Rahul Shukla2
Professor in Computer Application Department. R.G.P.V., S.A.T.I.
Vidisha, M.P. , India

M-Tech (Software System) Student, R.G.P.V., S.A.T.I.
Vidisha, M.P. , India

Used which better fit the properties of Web data.

Abstract Furthermore, not only data mining algorithms, but also
There is a considerable body of work on sequence mining of artificial intelligence information retrieval and natural
Web Log Data We are using One Pass frequent Episode language processing techniques can be used efficiently.
discovery (or FED) algorithm, takes a different approach than the Thus, Web mining has been developed into an
traditional apriori class of pattern detection algorithms. autonomous research area.
In this approach significant intervals for each Website are
Data Mining refers to the extracting or mining knowledge
computed first (independently) and these interval used for
detecting frequent patterns/Episode and then the Analysis is From large amounts of data' [3]. The growing interest in
performed on Significant Intervals and frequent patterns That can the field of data mining over the past few decades has
be used to forecast the user’s behavior using previous trends and resulted in a number of algorithms for processing various
this can be also used for advertising purpose. This type of kinds of data including time-series data. In general, time-
applications predicts the Website interest. In this approach, time- series data can be defined as an ordered sequence of
series data are folded over a periodicity (day, week, etc.) Which values of a variable at specific (mostly periodic) time
are used to form the Interval? Significant intervals are discovered intervals. Time series data analysis is used in a variety of
from these time points that satisfy the criteria of minimum data-centric applications such as economic forecasting,
confidence and maximum interval length specified by the user
sales forecasting, budgetary analysis, stock market
Keywords: Web log data, minimum confidence; periodicity;
analysis, inventory studies, census analysis and so forth. A
significant interval discovery; Frequent Episode/Pattern, Web
access, access count. considerable amount of work [6] has been done to process
and mine through the large collections of time-series data
sets using sequential mining. Most of the existing
1. Introduction sequential mining techniques use individual time points,
that is, events are considered to occur at specific time
The expansion of the World Wide Web (Web for short) points. On the other hand, in many real-world scenarios,
has resulted in a large amount of data that is now in events are likely to occur with a high degree of certainty,
general freely available for user access. The different not at specific time points, but within intervals. For
types of data have to be managed and organized in such a example, it is useful to extract intervals of high activity
way that they can be accessed by different users from telephone logs to understand network usage.
efficiently. Therefore, the application of data mining In the Discovery of Frequent Pattern/Episode, it is
techniques on the Web is now focus of an increasing important to predict website access in tight and accurate
number of researchers. Several data mining methods are intervals to effectively predict the user behavior. Although
used to discover the hidden information in the Web. there is a considerable body of work on sequential mining
However, Web mining does not only mean applying data of transactional data, most of them deal with time-point
mining techniques to the data stored in the Web. The data and make several passes over the entire data set in
algorithms have to be modified Such that they better suit order to discover frequently occurring patterns We use an
the demands of the Web. New approaches should be Approach in which significant intervals representing
intrinsic nature of data are discovered in a single pass. In
this approach, time-series data are folded over a
periodicity (day, week, etc.) Which are used to form the
Interval? Significant intervals are discovered from these
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 30
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

time points that satisfy the criteria of minimum confidence 3. Terms and Definition and Significant
and maximum interval length specified by the user. These Interval
Significant Interval are used to discover the frequent
patterns in Web log data We are Apply the One Pass SI algorithm[4],One Pass
In this Paper, we have used One PassSI Algorithm [4] and AllSI[4] and One Pass FED[5] on the web log data so we
One Pass AllSI Algorithm [4] for generating the need the different website name such as
Significant Interval. These algorithm are single-pass,, etc The time-
algorithm, these algorithm use the main-memory approach series can be represented with an Website timestamp
to discover significant intervals in the Web Log Data. This model. A website w (for example is access,
approach not only makes use of a reduced data set by is not access, etc.) is associated with a
compressing point-based data to an interval-based data, sequence of timestamps {T1, T2, • ••, Tn} that describes
its access over a period of time. The notion of periodicity
but also makes only a single pass over the entire data set. (such as daily, weekly, monthly, etc.) is used to group the
We are also performing the analysis on One Pass SI website accesses. For each website, the number of
algorithm [4], One Pass AllSI Algorithm [4] and One Pass accesses at each time point can be obtained by grouping
FED algorithm [5]. on the timestamp (or periodicity attribute). We term the
The remainder of the paper is organized as follows. number of accesses of each website as access count (ac).
Section - 2 discussed the Related work , section-3 Thus the time series data can be represented as < w {Tl,
Defined the Related terms and Definition and Significant al}, {T2, a2}, {T3, a3}, {Tn, an}>, where Ti represents
Interval.section-4 describe the process of significant the timestamp associated with the website w and ai
Interval Discovery with the Example section-5 discuss the represents number of accesses. ai can be referred as the
frequent pattern with the Example of Frequent Pattern access count of the website w at Timestamp Ti . This is
referred to as folding of the time-series data using
Discovery and section-6 shows the Experimental Analysis
periodicity and time granularity (e.g., daily on seconds,
on Significant Interval and Frequent pattern finally daily on minutes, weekly on minutes).
section-7 discussed the conclusion and future work
3.1 Significant Interval
2. Related Work Intervals associated with a website are characterized by
confidence and density. The confidence of a time point is
The significant intervals used for the lock and unlock the ratio (expressed as percentage) of the access count at
operations. These significant intervals can then be used to that point to the periodicity of data collection (number of
find an association or a relationship between the Websites. days or weeks (N)). When the interval consists of several
We have worked on the Dynamic Data (i.e. Web Log time points, total access count of the interval (the sum of
Data) that is dependent on the interest of user. These the access count of the points that form the interval) is
Intervals are used to discover the frequent pattern/Episode. used. The density of an interval is the ratio of the total
Though the concept of intervals has been used for finding access count of the interval to the interval length. We
association rules ([11] Miller & Yang, 1997; [12] Srikant define a Window as a time-interval wd [Ts, Tel where (Ts
& Agrawal, 1996b), not many data mining algorithms <= Te), Ts is the start time and Te is the end time of the
discuss the formation of intervals on time-series data interval. An interval associated with a website is
based on the interaction of events. WinEpi (Mannila et al., represented as [Ts, Te, ac, 1, d, cl where ac represents total
1995) makes multiple passes over the data for counting the access count of the interval, 1 denotes the length of an
support of the candidates in each pass. Algorithm [4] is interval (Te-Ts+1), d indicates the density, and c
closer to MinEpi (Mannila et al., 1995) as we obtain the represents confidence of the interval (ac/N * 100). Given a
event count in a single pass over the data and use it to time sequence T, minimum confidence min-conf and
obtain support for the intervals. Regarding timing interval length max-Len, we define the interval wd [Ts,
constraints, WinEpi and MinEpi find all sequences that Tel as a Significant Interval in T if:
satisfy the time constraint maximum span, which is 1) Confidence(c) of wd >= min-conf
defined as the maximum allowed time difference between 2 length (l) of wd <= max-Len
latest and earliest occurrences of events in the entire 3) There is no window wd'= [Ts ', Te'] in T such that Ts
sequence and minimum support, counted with the one '>=Ts and Te'=<Te that satisfies i) and ii).
occurrence per span window method. They have applied
this approach on the static data but we are applying this
approach on the Dynamic Data. 3.2 Valid Significant Interval
Significant intervals can be of unit size, overlapping or
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 31
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

disjoint. All valid intervals should be discovered by a 4. Process of Significant Interval Discovery
significant interval discovery (SID) algorithm. Figure 3.1
shows all possible valid significant intervals. That B The Raw data is collected and then apply the
combines with C since B started after A. The Figure3.1 preprocessing steps The Preprocessing steps are
taken from [4]. 1) Data Cleaning: In this step we are separate the data into
1) Unit Significant Interval: It is a significant interval with the different website record and use the different views to
the same start and end time; that is, Ts = Te. show the record of each website. This process is called
2) Disjoint Significant Intervals: They are defined as two Data cleaning
significant intervals (that is, wd [Ts, Te] and wd` [Ts`, 2) Data Folding: In the preprocessing phase, the input data
Te`]), which do not overlap (that is, Ts` >= Te and Te` is set is combined to periodicity of interest (example, daily
not in the interval [Ts, Te] or Te`> Ts and Ts` is not in the or weekly). This process is called Data Folding.
interval [Ts, Te]). Now One Pass SI or One Pass AllSI [4] is applied for
3) Overlapping Significant Intervals: They are defined as generate the significant Interval. Those Time points
two significant Intervals wd [Ts, Te] and wd’ [Ts’, Te’]. satisfied the Minimum Constraint (i.e. min-conf and max
(if Ts<=Te’<Te and Ts’<Ts) or (Ts<Ts’<=Te and Te’> Len for One Pass SI[4] or min-Conf for One Pass
Te). AllSI[4]) generate the Significant Interval are Discovered.

4.1 Example
Before applying One Pass SI or One Pass AllSI Algorithm
[4] The Preprocessing steps must be taken.
Consider a 7 day data set. This dataset contains the
Website name, its Access status (Access/Not Access), and
timestamp (this combination is considered as a unique
Website). There are two different website in the Table
4.1(raw data) so After Data Cleaning the Table 4.1 Break
Figure3.1 Valid Significant Interval up in the two different Tables. Table 4.2 Data Cleaning
( and Table 4.3 Date Cleaning (

3.3 Invalid Significant Interval Table 4.1 Raw data

An invalid significant interval is identified by the third Website Access Status Timestamp
condition of the definition for a significant interval; that is, Access 4/15/2009 2:05 pm
a significant interval cannot have an embedded valid Access 4/15/2009 2:10 pm
significant interval (including a unit significant interval). Access 4/16/2009 2:10 pm Access 4/17/2009 2:40 pm Access 4/18/2009 2:40 pm Access 4/19/2009 2:10 pm Access 4/19/2009 2:05 pm Access 4/20/2009 2:20 pm Access 4/20/2009 2:10 pm Access 4/21/2009 2:05 pm Access 4/21/2009 2:05 pm Access 4/21/2009 2:10 pm Access 4/22/2009 2:05 pm Access 4/22/2009 2:10 pm
Figure3.2 Invalid Significant Interval Access 4/22/2009 2:20 pm

Figure 3.2 gives an example of an invalid significant

Table 4.2 Data Cleaning (
interval. A is not a significant interval as it contains
Website Access Status Timestamp
another significant interval B. Hence, only B is a Access 2/15/2009 2:05 pm
significant interval. In other words, B is the tightest Access 2/16/2009 2:10 pm
significant interval. The Figure3.2 taken from [4]. Access 2/17/2009 2:40 pm Access 2/18/2009 2:40 pm Access 2/19/2009 2:05 pm
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 32
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814 Access 2/21/2009 2:05 pm the constraint of min-conf, it is not a significant interval Access 2/21/2009 2:10 pm because it does not satisfy the max-Len constraint. Access 2/22/2009 2:10 pm 3) Next, the algorithm starts significant interval discovery
at time point 2:40 which cannot be further combined with
any more time points. Since all time points in the data set
Table 4.3 Data Cleaning (
have been considered there are no more time points and
Website Access Status Timestamp Access 2/15/2009 2:10 pm
the algorithm stops. The significant Interval is given in Access 2/19/2009 2:10 pm Table 4.6 Access 2/20/2009 2:20 pm
Table 4.6 Significant Interval Using One PassSI algorithm Access 2/20/2009 2:10 pm Access 2/21/2009 2:05 pm Website Start Time End Time Interval
Confidence Access 2/22/2009 2:05 pm 2:05 2:10 85.71% Access 2/22/2009 2:20 pm 2:05 2:10 71.42% 2:10 2:20 71.42%
Now the Data Folding step is taken on Table 4.2 (for
Now for 60% minimum confidence (min-conf) in Daily and table 4.3 (for result of
Periodicity and max_Len is not required in the One Pass
Data Folding step is given in Table 4.4 and Table 4.5.The
AllSI [4], the One Pass-AllSI algorithm [4] performs the
Table 4.4 and Table 4.5 contain the time point at which
following steps for Table4.4:
the website was ‘Access’ and number of times (over the
1) The algorithm starts significant interval discovery at
entire dataset) it was ‘Access’ at that time point. For
time point 2:05 and combines it with time point 2:10 and
example, has been `Access' at 2:05 three
computes a confidence of 85.71%. Since the required 60%
times (on three different days) in the data set (of seven
confidence has achieved by combining these two time
Points. Discovery of a significant interval (say A) with
start time of 2:05 and end time of 2:10 and confidence of
Table 4.4 Data Folding (
Website Access Times of access Access
Status Count 2) Next, the algorithm starts significant interval discovery Access 2:05 3 at time point 2:10 and combines it with time point 2:40 to Access 2:10 3 give a confidence of 71.42% with interval length of 30 Access 2:40 2 minutes. This interval satisfies the constraint of min-conf,
it is a significant interval because there are No max_Len
Table 4.5 Data Folding ( Constraint. As we see in previous, when we used One
Website Access Times of Access Access Pass-SI algorithm for significant interval discovery, even
Status Count though this interval satisfied the constraint of min-conf, it Access 2:05 2 was not discovered as a significant interval because it did Access 2:10 3 not satisfy the constraint of max-Len. But here we are Access 2:20 2 using the One Pass-AllSI [4] algorithm for significant
interval discovery, this interval (say B) is classified as a
Consider a 7 day data set which gives a folded table as significant interval with start time of 2:10, end time of
given in Figure 4.4 and 4.5. For 60% minimum confidence 2:40 and confidence of 71.42%.
(min-conf), maximum interval length (max-Len) of 20 3) Next, the algorithm starts significant interval discovery
minutes and Daily periodicity, the One PassSI algorithm at time point 2:40 which cannot be further combined with
[4] performs the following steps for Table4.4: any more time points. Since all time points in the data set
1) The algorithm starts significant interval discovery at have been considered there are no more time points and
time point 2:05 and combines it with time point 2:10 and the algorithm stops. The significant Interval is given in
computes a confidence of 85.71% ((3+3)*100/7). Since Table 4.7
the required 60% confidence has achieved by combining
these two time points, and also satisfied the max-Len Table 4.7 Significant Interval using One Pass AllSI Algorithm
because the length is 5 minutes. Discovery of a significant Website Start Time End Time Interval
interval (say A) with start time of 2:05 and end time of Confidence
2:10 and confidence of 85.71%. 2:05 2:10 85.71%
2) Next, the algorithm starts significant interval discovery 2:10 2:40 71.42%
at time point 2:10 and combines it with time point 2:40 to 2:05 2:10 71.42%
give a confidence of 71.42 %(( 3+2)*100/7) with interval 2:10 2:20 71.42%
length of 30 minutes. Even though this interval satisfies
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 33
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

5. Frequent Pattern/Episode 1:00 1:15 1:10 1:20
As we saw in the significant Interval Discovery 2:00 2:10
confidence of an interval has been defined for a single 2:00 2:10
website .However, One Pass FED algorithm [5] detects 2:05 2:15
frequent episodes/pattern. Hence, we need to define the
confidence of a pattern using the confidence associated For this example, we shall consider Semantics-s and
with the websites forming the pattern. This is termed as Sequential Window-30 minutes. The One Pass FED
Pattern Confidence (PC). The Pattern Confidence of an Algorithm [5] goes through the following steps:
episode within an interval is defined as the minimum of The number of distinct Website (n) is found as 3, that is n
the confidence of its constituent websites. When several = 3.The One Pass FED algorithm [5] starts with first
websites access with different confidence in an interval, significant interval [, 1:00, 1: 15, 70], that is,
we can only infer that all websites access with minimum start is pointing at this significant interval.
confidence in that interval. Hence, the confidence of a The next pointer moves to significant interval [,
pattern interval represents the minimum number of 1:10, 1:20, 80]. The start of this significant interval lies
occurrence of an episode within an interval. With within Sequential Window units (that is, 30 minutes).
frequently occurring patterns, Pattern Confidence Hence, these two significant intervals form a 2nd level
underestimates the actual probability of the websites frequent pattern/episode [, 1:00,
access together but retains its significance or order relative 1:20, 70]. The Pattern Confidence of the frequent episode
to the other patterns discovered. Pattern Confidence is is taken as the minimum of the two confidences which is
used instead of access count in the One Pass FED 70 in this case. Now, the next pointer moves to significant
algorithm [5]. Frequent pattern/episodes are discovered by interval [, 2:00, 2: 10, 75]. The Sequential
combining/merging Website. The parameters that define Window constraint is violated. Hence, the next pointer
how this combination/ merge take place are Sequential will not move any further. None of the significant intervals
Window and Semantics. Sequential Window defines the after this significant interval will not satisfy the Sequential
maximum allowed time within which Website in a Window constraint. Hence, no more frequent pattern can
frequent pattern/episode may Start/End. The way in which be further formed with [, 1:00, 1:15, 70] as
this maximum allowed time is interpreted is defined by the start (base) significant interval. Now, the start pointer
Semantics. There are two types of Semantics which a user moves to the significant interval [, 1:10, 1:20, 80].
may define: Semantics-s and Semantics-e. Semantics-s The next pointer moves to significant interval
generates all possible combinations of websites, which [, 2:00, 2:10, 75]. Again, the Sequential
Start within Sequential Window units of the first websites. Window constraint is violated and no more frequent
Semantics-e, on the other hand, generates combinations of pattern can be discovered with the current base significant
websites that Start and End within the Sequential Window interval.
units of the first websites. We have Developed the Now, the start pointer moves to the significant interval
Frequent Pattern using Semantic-s and now we are [, 2:00, 2:10, 75]. The next pointer moves
working to generate the Frequent pattern Using Semantic- to significant interval [, 2:00, 2: 10, 80]. Both
e. these significant intervals combine to form a 2nd level
frequent pattern/episode [,,
2:00, 2:10, 75]. The next pointer now moves to the
significant interval [, 2:05, 2:15, 70]. This
combines with the current frequent pattern to form a 3rd
5.1 Example level frequent pattern [,,, 2:00, 2:15, and 70]. Now, as the value of n is 3,
In this subsection, we shall explain the above One Pass we can conclude that no more higher level frequent pattern
FED Algorithm [5] with the help of an example. This will will be formed using the base significant interval
give a clear view of the algorithm. One Pass FED [, 2:00, 2:10, 75] Now, the start pointer
algorithm [5] works on the significant intervals generated moves to the significant interval [,2:00,
by the significant interval discovery algorithms. For this 2:10,80] and the process will continue
example, let us assume that the significant intervals have
been generated using one of the significant interval Table 5.1 Second Level Frequent Pattern
discovery algorithms and are given in Table5.0
Second Level Frequent Pattern
Table 5.0 Significant Interval Discovered by SID Website1 Website2 Start End Confide
Website Start Time End Time Time Time nce (%)
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 34
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814 1:00 1:20 70 2:00 2:10 75 2:00 2:15 70

Number of Significant
Vary Max Length
Table 5.2 Third Level Frequent Pattern 570

Third Level Frequent Pattern 565
Website1 Website2 Website3 Start End Confi 555
Time Time dence 550
(%) 10 20 30 40 50 60
Newswo Citeseer. 2:00 2:15 70
Max Length com
Finally, the frequent episodes have detected and given in Confidence-40
the Table 5.1 and 5.2.
Figure 6.2

4. Experimental Analysis 6.1.2 Effect of min-conf

Now we want to test the effect of the changing the min-
A number of experiments were carried out to compare the conf user parameter for the same max-Len and periodicity.
performance of One Pass SI by changing the confidence The chart seen in Figure 6.3 shows the effect of varying
and max_Len and compare the performance of One Pass the min-conf for a set of experiments keeping the max-Len
SI[4] and one pass AllSI[4] and also Check the No. of = 20 and periodicity as Weekly. As seen from the figure
Frequent Pattern/Episode have been Discovered when we 5.3 as the min-conf increases the No. of Significant
are changing the Sequential window length. The Website Interval intervals are decreases.
Interest is calculating in different months. The Algorithm
One Pass SI [4], One Pass AllSI [4] and One Pass Fed
Algorithm [5] were implemented in VB.Net as a Front Vary Confidence
End and SQL Server 2000 As the Back End. The data has 1000
Number of

been taken from the cyber café which is providing the 800

services to their customer in the 24 hours. 400
Data set used for experiments Figure 6.1
10 30 50 70 90
No of Days No of Tuples No of Websites
90 1700 5 Confidence

M ax Length-20

Figure 6.3

This is because as the min-conf increases the numbers of

time points which are checked for forming significant
6.1 Effect of Varying Parameter on One Pass SI intervals are decreases
6.2 Comparison of One Pass SI and One Pass AllSI
6.1.1 Effect of max_Len Algorithm
In these experiment we want to test the effect of the
changing the max-Len user parameter for the same min- 6.2.1 Comparison between the numbers of Significant
conf and periodicity. The chart seen in Figure 6.2 shows Intervals
the effect of varying the max-Len for a set of experiments
keeping the min-conf = 40% and periodicity as Weekly.
As seen from the figure 5.2 as the max-Len increases the An experiment was carried out to compare the number of
No. of Significant Interval intervals are also increases. significant intervals discovered by both the One Pass-SI
This is because, as the max-Len increases the number of and One Pass-AllSI algorithms [4]. The One Pass SI [4]
time Points which are checked for forming significant has Discover more significant Interval than One Pass
intervals Also increases. AllSI [4] up to confidence 60. This is because the One
Pass-AllSI algorithm does not have a limitation on the
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 35
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

max-Len but one Pass SI has this limitation When the Figure 6.5
confidence increases more than 60 One Pass SI [4]
generates the more significant Interval as compare to One
Pass AllSI [4] Actually this is dependent on the Data The 6.3 Effect of Varying Sequential Window Length on
parameter for Experiment was max_Len=20 minutes for One Pass FED Algorithm
One PassSI [4] comparison between the One Pass SI [4]
and One Pass AllSI [4] is given in Figure 6.4. In the First phase of paper we have developed the
significant Interval using One Pass SI algorithm [4] then
we have applied the One Pass FED algorithm [5] for
generate the Frequent Pattern/Episode. In this experiment
we want to test the effect of the changing the sequential
Comparison Between OnePassSI and window length user parameter for the weekly periodicity.
The Figure 6.6 shows the effect of varying the sequential
No of Significant Interval

1000 window length. It is clear from the figure 6.6 when we

800 increase the sequential window length the no. of FED also
600 increase.
100 Vary Sequential Window Length
10 20 30 40 50 60 70 80 90 100 570
No. of FED 565
OnePass One Pass ALLSI

Figure 6.4 555

10 20 30 40 50
Sequential Window Length

6.2.2 Time Comparison

As the One Pass AllSI algorithm [4] produces a superset Figure 6.6
of significant intervals as compared to its One Pass SI
algorithm [4] counterpart, we wanted to check the extra
time taken for finding all significant intervals at same min- 6.4 Contribution of Website in Frequent Pattern
conf. An experiment was carried out to analyze the Discovery
difference in the time taken by each of these algorithms.
As seen from Figure 6.5, the One Pass-AllSI algorithm [4] In the final phase we are showing the interest of user in
takes more time than the One Pass-SI algorithm [4] for the the particular website in the particular period of time. We
different min-conf values. have taken the data of month April, May and June. In the
figure 6.7 we are showing the contribution of website in
the Frequent pattern/Episode Discovery. The table-a gives
the full form of website which are use in the Figure6.7.

Comparison of Time Execution Table-a


10 SC
8 RN
4 EC
10 30 50 70 90 In the starting of April the,
Confidence, is heavily accessed because this
is the time of Election and IPL Tournament and M-Tech
One Pass One Pass ALLSI
research student are also doing their research. In the month
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 36
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

of May the people are also interested in the [6] John F. Roddick and Myra Spiliopoulou. A survey of
because this is the time of admission in the different temporal knowledge discovery paradigms and methods.
courses and this time result of different courses are also IEEE Transactions on Knowledge and Data Engineering,
declared. In the month of June interest in little bit 14(4):750-767, Jul/Aug. 2002
increase as compare to previous month. [7] Cook, D. J., Youngblood, G. M., III, E. O. H., Gopalratnam,
K., Rao, S., Litvin, A., et al. (2003).
[8] Mannila, H., Toivonen, H., & Verkamo, A. I. (1995).
Contribution of Website in Frequent Pattern Discovering frequent episodes in sequences. In KDD (pp.
Discovery 210-215).
[9]Srikant, R., & Agrawal, R. (1996a). Mining sequential
contribution of Website

50 CC patterns: Generalizations and performance improvements. In

EDBT (pp. 3-17).
in FED(%)

30 [10] Srinivasan, A., Bhatia, D., & Chakravarthy, S. (2006).

SC Discovery of interesting episodes in sequence data. In Proc.
10 RN of the 21st ACM SAC (pp. 599-602). France:
EC [11] Miller, R. J., & Yang, Y. (1997). Association rules over
-10 15 30 45 60 75 90 interval data. In SIGMOD conference (pp. 452-461).
[12] Srikant, R., & Agrawal, R. (1996b). Mining quantitative
No. of Days
association rules in large relational tables. In SIGMOD
conference (pp. 1-12).

Figure 6.7 Dr. Kanak Saxena is the Professor in the Computer Application
Department of Samrat Ashok Technological Institute, Vidisha
(M.P.), India. This Institute is Affiliated Rajiv Gandhi
Proudyogiki Vishwavidyalaya, Bhopal (M.P.), India
7. Conclusion and Future Work
Mr.Rahul Shukla is the Student of M-Tech (Software System)
from Samrat Ashok Technological Institute, Vidisha (M.P.),
In this paper, we have presented the discovery of
significant intervals in Web Log Data Using One Pass SI
and One Pass AllSI algorithm [4] and these significant
Intervals are used to generate the Frequent Pattern. The
frequent Pattern is generating by the One Pass FED
algorithm [5]. These patterns are used to forecast the
trend. We are applied this [4] for web Log Data Which is
the Dynamic nature Data. And analysis performs on Web
Log Data. And Experimental result shows the result of this
analysis. Currently, we are working on Frequent Pattern
Generation using Semantic-e on time-series web Log Data
This process applied for other domain, such as traffic
analysis, and others.

[1] Renate Iváncsy, István Vajk, Frequent Pattern Mining in Web
Log Data Acta Polytechnica Hungarica Vol. 3, No. 1, 2006
[2] Michael H. Bohlen, Renato Busatto, and Christian S. Jensen.
Point-versus interval-based temporal data models. In ICDE,
pages 192-200, 1998.
[3] Jiawei Han and Micheline Kamber. Data Mining: Concepts
and Techniques, chapter 1.2, page 5. Morgan Kaufmann
Publishers, 2005
[4] Sagar Savla, Sharma Chakravarthy, A Single Pass Algorithm
for Discovering Significant Intervals In Time series data
,International Journal of Data Warehousing and Mining,
Volume 3, Issue 3 , IGI Global
[5] Sagar Savla and Sharma chakravarthy, An Efficient single
pass approach to frequent episode discovery in sequence
data, IET 4th International Conference 2008
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 37
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Efficient Packet Forwarding Approach in Vehicular Ad Hoc

Networks Using EBGR Algorithm
K.Prasanth1, Dr. K. Duraiswamy2, K. Jayasudha3 and Dr. C. Chandrasekar4
Research Scholar, Department of Information Technology
K.S.Rangasamy College of Technology, Tiruchengode – 637 215, Tamilnadu, India

Dean Academic, Department of Computer Science
K.S.Rangasamy College of Technology, Tiruchengode – 637 215, Tamilnadu, India

Research Scholar, Department of Computer Applications
K.S.R College of Engineering, Tiruchengode – 637 215, Tamilnadu, India

Reader, Periyar University, Salem, Tamilnadu, India

Abstract technologies, employing short-range to medium-range

communication systems. The radio range of VANETs is
VANETs (Vehicular Ad hoc Networks) are highly mobile several hundred meters, typically between 250 and 300
wireless ad hoc networks and will play an important role in meters. In US, the Federal Communications Commission
public safety communications and commercial applications. (FCC) has allocated 75 MHz in 5.9 GHz band for licensed
Routing of data in VANETs is a challenging task due to rapidly Dedicated Short Range Communication (DSRC) for
changing topology and high speed mobility of vehicles.
Conventional routing protocols in MANETs (Mobile Ad hoc
vehicle-to-vehicle and vehicle to infrastructure
Networks) are unable to fully address the unique characteristics communications. In Europe, the Car-to- Car
in vehicular networks. In this paper, we propose EBGR (Edge Communication Consortium (C2C-CC) has been initiated
Node Based Greedy Routing), a reliable greedy position based by car manufacturers and automotive OEMs (original
routing approach to forward packets to the node present in the equipment manufacturers), with the main objective of
edge of the transmission range of source/forwarding node as increasing road traffic safety and efficiency by means of
most suitable next hop, with consideration of nodes moving in inter-vehicle communication.
the direction of the destination. We propose Revival Mobility
model (RMM) to evaluate the performance of our routing
technique. This paper presents a detailed description of our 2. Related Work
approach and simulation results show that packet delivery ratio is
improved considerably compared to other routing techniques of In this section, we briefly summarize the characteristics of
VANET. VANETs related to routing and also we will survey the
existing routing schemes in both MANETs and VANETs
Keywords: Vehicular Ad hoc Networks, Greedy Position Based in vehicular environments.
Routing, EBGR, Revival Mobility Model, Packet Delivery Ratio.
2.1. VANETs Characteristics
1. Introduction
In the following, we summarize the unique characteristics
Inter-vehicle communication (IVC) is attracting of VANETs compared with MANETs.
considerable attention from the research community and
the automotive industry. It is beneficial in providing Unlimited transmission power: Mobile device power
intelligent transportation system (ITS) as well as drivers issues are not a significant constraint in vehicular
and passenger’s assistant services. VANETs is a form of Networks. Since the vehicle itself can provide continuous
mobile ad hoc network providing communications among power to computing and communication devices.
nearby vehicles as well as between vehicles and nearby High computational capability: Operating vehicles can
fixed equipment, usually described as roadside equipment. afford significant computing, communication and sensing
VANETs have similar or different radio interface capabilities.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 38

Highly dynamic topology: Vehicular network scenarios representative of the reactive approaches we have chosen
are very different from classic ad hoc networks. In DSR, since it has been shown to be superior to many other
VANETs, vehicles can move fast. It can join and leave the existing reactive ad-hoc routing protocols in [8].
network much more frequently than MANETs. Since the Position-based routing algorithms require information
radio range is small compared with the high speed of about the physical position of the participating nodes. This
vehicles (typically, the radio range is only 250 meters position is made available to the direct neighbors in the
while the speed for vehicles in freeway will be 30m/s). form periodically transmitted beacons. A sender can
This indicates the topology in VANETs changes much request the position of a receiver by means of a location
more frequently. service. The routing decision at each node is then based on
Predicable Mobility: Unlike classic mobile ad hoc the destination’s position contained in the packet and the
networks, where it is hard to predict the nodes’ mobility, position of the forwarding node’s neighbors. Position-
vehicles tend to have very predictable movements that are based routing does not require the establishment or
(usually) limited to roadways. The movement of nodes in maintenance of routes. Examples for position-based
VANETs is constrained by the layout of roads. Roadway routing algorithms are face-2 [9], GPSR [10], DREAM
information is often available from positioning systems [11] and terminodes routing [12]. As a representative of
and map based technologies such as GPS. Each pair of the position based algorithms we have selected GPSR,
nodes can communicate directly when they are within the (which is algorithmically identical to face-2), since it
radio range. seems to be scalable and well suited for very dynamic
Potentially large scale: Unlike most ad hoc networks networks.
studied in the literature that usually assume a limited
network size, vehicular networks can is extended over the 2.3. Routing protocols in VANET
entire road network and include many participants.
Partitioned network: Vehicular networks will be Following are a summary of representative VANETs
frequently partitioned. The dynamic nature of traffic may routing algorithms.
result in large inter-vehicle gaps in sparsely populated
2.3.1 GSR (Geographic Source Routing)
scenarios and hence in several isolated clusters of nodes.
Network connectivity: The degree to which the network is Lochert et al. in [13] proposed GSR, a position-based
connected is highly dependent on two factors: the range of routing with topological information. This approach
wireless links and the fraction of participant vehicles, employs greedy forwarding along a pre-selected shortest
where only a fraction of vehicles on the road could be path. The simulation results show that GSR outperforms
equipped with wireless interfaces. topology based approaches (AODV and DSR) with respect
to packet delivery ratio and latency by using realistic
2.2 Routing protocols in MANET vehicular traffic. But this approach neglects the case that
The routing protocols in MANETs can be classified by there are not enough nodes for forwarding packets when
their properties. On one hand, they can be classified into the traffic density is low. Low traffic density will make it
two categories, proactive and reactive. difficult to find an end-to-end connection along the pre-
selected path.
Proactive routing algorithms employ classical routing
strategies such as distance-vector routing (e.g., DSDV [1]) 2.3.2 GPCR (Greedy Perimeter Coordinator Routing)
or link-state routing (e.g., OLSR [2] and TBRPF [3]). To deal with the challenges of city scenarios, Lochert et al.
They maintain routing information about the available designed GPCR in [14]. This protocol employs a restricted
paths in the network even if these paths are not currently greedy forwarding procedure along a preselected path.
used. The main drawback of these approaches is that the When choosing the next hop, a coordinator (the node on a
maintenance of unused paths may occupy a significant part junction) is preferred to a non coordinator node, even if it
of the available bandwidth if the topology of the network is not the geographical closest node to destination. Similar
changes frequently [4]. Since a network between cars is to GSR, GPCR neglects the case of low traffic density as
extremely dynamic we did not further investigate well.
proactive approaches.
Reactive routing protocols such as DSR [5], TORA [6], 2.3.3 A-STAR (Anchor-based Street and Traffic Aware
and AODV [7] maintain only the routes that are currently Routing)
in use, thereby reducing the burden on the network when
only a small subset of all available routes is in use at any To guarantee an end-to-end connection even in a vehicular
time. It can be expected that communication between cars network with low traffic density, Seet et al. proposed A-
will only use a very limited number of routes, therefore STAR [15]. A-STAR uses information on city bus routes
reactive routing seems to fit this application scenario. As a to identify an anchor path with high connectivity for
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 39

packet delivery. By using an anchor path, A-STAR node, which is considered as most suitable next hop, due
guarantees to find an end-to-end connection even in the to high dynamics of vehicles. This will lead to low packet
case of low traffic density. This position-based scheme delivery ratio, high end to end delay and increased packet
also employs a route recovery strategy when the packets drops.
are routed to a local optimum by computing a new anchor The various routing protocols of MANETs and VANETs
path from local maximum to which the packet is routed. are analyzed and drawbacks of those routing protocols are
The simulation results show A-STAR achieves obvious described in the Table 1.
network performance improvement compared with GSR
Table 1
and GPSR. But the routing path may not be optimal Drawbacks of routing protocols in MANET and VANET
because it is along the anchor path. It results in large
delay. Routing Protocols Drawbacks
Frequent network disconnection.
2.3.4 MDDV (Mobility-Centric Data Dissemination Routing loops.
Algorithm for Vehicular Networks) Too many hops.
Routing in wrong direction.
To achieve reliable and efficient routing, Wu et al. End to end connection is difficult in
proposed MDDV [16] that combines opportunistic low traffic density.
forwarding, geographical forwarding, and trajectory-based GPCR
End to end connection is difficult in
forwarding. MDDV takes into account the traffic density. low traffic density.
A forwarding trajectory is specified extending from the Routing paths are not optimal and
source to the destination (trajectory-based forwarding), A-STAR results in large delay of packet
along which a message will be moved geographically
closer to the destination (geographical forwarding). The Large delay if the traffic density varies
selection of forwarding trajectory uses the geographical by time.
knowledge and traffic density. MDDV assumes the traffic Large delay due to varying topology
density is static. Messages are forwarded along the and varying traffic density.
forwarding trajectory through intermediate nodes which Too many hops.
Large delay if the traffic density is
store and forward messages opportunistically. This PDGR
approach is focusing on reliable routing. The trajectory- Low packet delivery ratio.
based forwarding will lead to large delay if the traffic Frequent network disconnection.
density varies by time.

2.3.5 VADD (Vehicle-Assisted Data Delivery)

3. Proposed Routing Algorithm

To guarantee an end-to-end connection in a sparse network 3.1. Edge Node Based Greedy Routing Algorithm
with tolerable delay, Zhao and Cao proposed VADD [17] (EBGR)
based on the idea of carry and forward by using predicable EBGR is a reliable greedy position base routing algorithm
mobility specific to the sparse networks. Instead of routing designed for sending messages from any node to any other
along a pre-select path, VADD chooses next hop based on node (unicast) or from one node to all other nodes
the highest pre-defined direction priority by selecting the (broadcast/multicast) in a vehicular ad hoc network. The
closest one to the destination. The simulation results show general design goals of the EBGR algorithm are to
VADD outperforms GPSR in terms of packet delivery optimize the packet behavior for ad hoc networks with
ratio, data packet delay, and traffic overhead. This high mobility and to deliver messages with high reliability.
approach predicts the directions of vehicles movement.
But it doesn’t predict the environment change in the The EBGR algorithm has six basic functional units. First is
future. Neighbor Node Identification (NNI), second is Distance
2.3.6 PDGR (Predictive Directional Greedy Routing) Calculation (DC), third is Direction of Motion
Identification (DMI), fourth is Reckoning Link Stability
Jiayu Gong proposed PDGR [18], in which the weighted (RLS), fifth is Potential score calculation (PS) and sixth is
score is calculated for current neighbors and possible Edge Node Selection (ENS). The NNI is responsible for
future neighbors of packet carrier. With Predictive DGR collection of information of all neighbor nodes present
the weighted scores of immediate nodes 2-hops away are within the transmission range of source/forwarder node at
also calculated beforehand. Here next hop selection is any time. The DC is responsible for calculating the
done on prediction and it is not reliable at all situations. It closeness of next hop using distance information from the
doesn’t guarantee the delivery of packet to the node GPS. DMI is responsible to identify the direction of
present in the edge of the transmission range of forwarding motion of neighbor nodes which is moving towards the
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 40

direction of destination. The RLS is responsible for using periodic beacon messages. The neighbor node which
identifying link stability between the source/forwarder is closer to the destination node is calculated. The
node and its neighbor nodes. The PS is responsible to closeness of next hop is identified by the mathematical
calculate potential score and identifies the neighbor node model [18] and it is shown in Fig.1.
having higher potential for further forwarding of a
particular packet to destination. The ENS is responsible to 𝑫𝑫𝒊𝒊
𝑫𝑫𝑫𝑫 = � 𝟏𝟏 − �
select an edge node having higher potential score in 𝑫𝑫𝒄𝒄
different levels of transmission range. In the following
section, the general assumptions of EBGR algorithm are 𝑯𝑯𝑯𝑯𝑯𝑯𝑯𝑯,
briefly discussed and then functional units of EBGR 𝑫𝑫𝒊𝒊 ∶ 𝑆𝑆ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖 𝑡𝑡𝑡𝑡
𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝐷𝐷.
algorithm are discussed in detail.
𝑫𝑫𝒄𝒄 ∶ 𝑆𝑆ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
3.2. Assumptions 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑐𝑐 𝑡𝑡𝑡𝑡 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝐷𝐷.
∶ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑜𝑜𝑜𝑜 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑜𝑜𝑜𝑜.
The algorithm design is based on the following 𝑫𝑫𝒄𝒄
assumptions: All nodes are equipped with GPS receivers,
digital maps, optional sensors and On Board Units (OBU). Fig. 1 Distance Calculation in EBGR
Location information of all vehicles/nodes can be
identified with the help of GPS receivers. The only 3.5. Direction of Motion Identification (DMI)
communications paths available are via the ad-hoc
network and there is no other communication The appropriate neighbor node which is moving towards
infrastructure. Node power is not the limiting factor for the the direction of destination node is identified using the
design. Communications are message oriented. The mathematical model [18] and it is shown in Fig.2.
Maximum Transmission Range (MTR) of each node in the
environment is 250m. �⃗𝒊𝒊 , 𝒍𝒍⃗𝒊𝒊,𝒅𝒅 �
𝑫𝑫𝑫𝑫𝑫𝑫 = 𝐜𝐜𝐜𝐜 𝐬𝐬� 𝝊𝝊

3.3. Neighbor Node Identification (NNI) 𝑯𝑯𝑯𝑯𝑯𝑯𝑯𝑯,

�⃗𝒊𝒊 ∶ 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 𝑓𝑓𝑓𝑓𝑓𝑓 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖.
Neighbor node identification is the process whereby a 𝒍𝒍⃗𝒊𝒊,𝒅𝒅 ∶ 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 𝑓𝑓𝑓𝑓𝑓𝑓 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖 𝑡𝑡𝑡𝑡 𝑡𝑡ℎ𝑒𝑒
vehicle/node identifies its current neighbors within its 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝐷𝐷.
transmission range. For a particular vehicle, any other �⃗𝒊𝒊 , 𝒍𝒍⃗𝒊𝒊,𝒅𝒅 � ∶ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚
𝐜𝐜𝐜𝐜 𝐬𝐬� 𝝊𝝊
vehicle that is within its radio transmission range is called 𝑏𝑏𝑏𝑏 𝑡𝑡ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣
a neighbor. All vehicles consist of neighbor set which
holds details of its neighbor vehicles. Since all nodes
Fig. 2 Direction of Movement Identification in EBGR
might be moving, the neighbors for a particular mobile
node are always changing. The neighbor set is dynamic
The cosine value of vector for velocity of edge node i and
and needs to be updated frequently. Generally, neighbor
vector for location of edge node i to the location of
node identification is realized by using periodic beacon
destination node D is measured. A large cosine value
messages. The beacon message consists of node ID, node
implies a vehicle/node can still approach the destination
location and timestamp. Each node informs other nodes of
closer and closer along its current direction.
its existence by sending out beacon message periodically.
All nodes within the transmission range of source/packet 3.6 Reckoning Link Stability (RLS)
forwarding node will intimate its presence by sending a
beacon message every µ second. After the reception of a Each vehicle estimates the Link Stability (LS) for each
beacon, each node will update its neighbor set table. If a neighboring vehicle before selecting the next hop for the
node position is changed, then it will update its position to data forwarding/sending. The LS is a relation between the
all neighbors by sending beacon signal. If a known link communication lifetime and a constant value (say: σ)
neighbor, times out after α * µ seconds without having which represents in general cases the routing route validity
received a beacon ( α is the number of beacons that a time, and it depends on the used routing protocol. Fig.3
shows how link lifetimes are estimated [19] based on
node is allowed to miss) and it will be removed from the
neighbors' movement information.
neighbor set table.
3.4. Distance calculation (DC) The lifetime of the link (i, j) 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿[𝑖𝑖, 𝑗𝑗] corresponds to
The location and distance information of all the estimated time 𝛥𝛥𝛥𝛥 = 𝑡𝑡1 − 𝑡𝑡0 with 𝑡𝑡1 is the time when
vehicles/nodes can be identified with the help of GPS 𝐷𝐷1 becomes equal or bigger than the communication range
receivers. It can be communicated to neighbor vehicles R (i.e. the time when j goes out of the communication
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 41

range of i). 𝐷𝐷1 and 𝛥𝛥𝛥𝛥 are estimated using the initial
positions of i and j ( (𝑋𝑋𝑖𝑖0 , 𝑌𝑌𝑖𝑖0 ) and �𝑋𝑋𝑗𝑗 0 , 𝑌𝑌𝑗𝑗 0 �, and their 𝑫𝑫𝒊𝒊
𝑷𝑷𝑷𝑷𝒊𝒊 = 𝛒𝛒 × � 𝟏𝟏 − �⃗𝒊𝒊 , 𝒍𝒍⃗𝒊𝒊,𝒅𝒅 � + 𝛌𝛌 × 𝑳𝑳𝑳𝑳𝒄𝒄,𝒊𝒊
� + 𝝎𝝎 × 𝐜𝐜𝐜𝐜 𝐬𝐬� 𝝊𝝊
�⃗𝑗𝑗 respectively).
�⃗𝑖𝑖 and 𝑉𝑉 𝑫𝑫𝒄𝒄
initial speeds 𝑉𝑉
𝑷𝑷𝑷𝑷𝒊𝒊 ∶ 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑜𝑜𝑜𝑜 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖
𝐷𝐷1 2 = ((𝑋𝑋𝑖𝑖0 + 𝑉𝑉𝑉𝑉𝑖𝑖 𝛥𝛥𝛥𝛥) − �𝑋𝑋𝑗𝑗 0 + 𝑉𝑉𝑉𝑉𝑗𝑗 𝛥𝛥𝛥𝛥� 𝛒𝛒, 𝝎𝝎, 𝛌𝛌 ∶ 𝑃𝑃𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
+ (𝑌𝑌𝑖𝑖0 + 𝑉𝑉𝑉𝑉𝑖𝑖 𝛥𝛥𝛥𝛥) − (𝑌𝑌𝑗𝑗 0 + 𝑉𝑉𝑉𝑉𝑗𝑗 𝛥𝛥𝛥𝛥))2 𝐿𝐿𝐿𝐿𝐿𝐿 𝛒𝛒 + 𝝎𝝎 + 𝛌𝛌 = 𝟏𝟏 ; 𝛌𝛌 > 𝛒𝛒 𝑎𝑎𝑎𝑎𝑎𝑎 𝛌𝛌 > 𝝎𝝎
2 2
𝐷𝐷1 = 𝐴𝐴𝐴𝐴𝐴𝐴 + 𝐵𝐵𝐵𝐵𝐵𝐵 + 𝐶𝐶 𝑫𝑫𝒊𝒊 ∶ 𝑆𝑆ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖 𝑡𝑡𝑡𝑡
2 2 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝐷𝐷.
𝐴𝐴 = �𝑉𝑉𝑉𝑉𝑖𝑖 − 𝑉𝑉𝑉𝑉𝑗𝑗 � + �𝑉𝑉𝑉𝑉𝑖𝑖 − 𝑉𝑉𝑉𝑉𝑗𝑗 �
𝑫𝑫𝒄𝒄 ∶ 𝑆𝑆ℎ𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
𝐵𝐵 = 2[�𝑋𝑋𝑖𝑖0 − 𝑋𝑋𝑗𝑗 0 ��𝑉𝑉𝑉𝑉𝑖𝑖 − 𝑉𝑉𝑉𝑉𝑗𝑗 � 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑐𝑐 𝑡𝑡𝑡𝑡 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝐷𝐷.
+�𝑋𝑋𝑖𝑖0 − 𝑋𝑋𝑗𝑗 0 ��𝑉𝑉𝑉𝑉𝑖𝑖 − 𝑉𝑉𝑉𝑉𝑗𝑗 �] 𝑫𝑫𝒊𝒊
2 2 ∶ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑜𝑜𝑜𝑜 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑜𝑜𝑜𝑜.
𝐶𝐶 = �𝑋𝑋𝑖𝑖0 − 𝑋𝑋𝑗𝑗 0 � + �𝑌𝑌𝑖𝑖0 − 𝑌𝑌𝑗𝑗 0 � 𝑫𝑫𝒄𝒄
�⃗𝒊𝒊 ∶ 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 𝑓𝑓𝑓𝑓𝑓𝑓 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖.
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑡𝑡ℎ𝑒𝑒 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒: 𝒍𝒍⃗𝒊𝒊,𝒅𝒅 ∶ 𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 𝑓𝑓𝑓𝑓𝑓𝑓 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛
𝐴𝐴𝐴𝐴𝐴𝐴 2 + 𝐵𝐵𝐵𝐵𝐵𝐵 + 𝐶𝐶 − 𝑅𝑅 2 = 0 𝑖𝑖 𝑡𝑡𝑡𝑡 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝐷𝐷
𝑤𝑤𝑤𝑤 𝑐𝑐𝑐𝑐𝑐𝑐 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓 𝛥𝛥𝛥𝛥. 𝐜𝐜𝐜𝐜 𝐬𝐬� 𝝊𝝊 �⃗𝒊𝒊 , 𝒍𝒍⃗𝒊𝒊,𝒅𝒅 � ∶ 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚
𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿[𝑖𝑖, 𝑗𝑗] = 𝛥𝛥𝛥𝛥 𝑏𝑏𝑏𝑏 𝑡𝑡ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣
𝑳𝑳𝑳𝑳𝒄𝒄,𝒊𝒊 ∶ 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓𝑓
𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑐𝑐 𝑡𝑡𝑡𝑡 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖.
𝑳𝑳𝑳𝑳[𝒊𝒊, 𝒋𝒋] =
Fig. 4 Potential Score Calculation in EBGR
𝐿𝐿𝐿𝐿[𝑖𝑖, 𝑗𝑗] = 1 𝑤𝑤ℎ𝑒𝑒𝑒𝑒 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿[𝑖𝑖, 𝑗𝑗] ≥ σ
3.8. Edge Node Selection (ENS)
𝑳𝑳𝑳𝑳𝒊𝒊,𝒋𝒋 ∶ 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑡𝑡𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎 𝑗𝑗.
In the Edge Node Selection, edge nodes are selected for
packet forwarding event. An edge node is a node which
has shortest distance to the destination D compared to all
Fig. 3 Reckoning Link Stability in EBGR other nodes within the different levels of transmission
range of source/packet forwarding node.
Once LS is calculated for each neighboring vehicle, EBGR
selects the node corresponding to the highest LS
(corresponding to the most stable neighboring link) as next
hop for data forwarding. This approach should help as well
in minimizing the risk of broken links and in reducing
packet loss.

3.7 Potential Score Calculation (PS)

The potential score (PS) of all nodes present within the

different levels of transmission range of source/packet
forwarding node is calculated. The potential score (PS) is
calculated to identify the closeness of next hop to
destination, direction of motion of nodes and reliability of
neighbor nodes. The appropriate edge node with largest
potential score will be considered as having higher
potential to reach the destination node and that particular
node can be chosen as next hop to forward the packet to
the destination node. Potential score is calculated by
addition of DC, DMI and LS and that mathematical model
represented in Fig.4.

Fig 5: Different Levels of Transmission Range in EBGR

𝑷𝑷𝑷𝑷𝒊𝒊 = 𝛒𝛒 × 𝐃𝐃𝐃𝐃 + 𝝎𝝎 × 𝐃𝐃𝐃𝐃𝐃𝐃 + 𝛌𝛌 × 𝐋𝐋𝐋𝐋
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 42

The different levels of transmission range are considered 26. 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖
to avoid packet loss due to high speed mobility of 27. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇
vehicles. An edge node has the responsibility of saving 28. 𝒆𝒆𝒆𝒆𝒆𝒆𝒆𝒆 𝒊𝒊𝒊𝒊 (𝐷𝐷𝑐𝑐𝑐𝑐 < 𝐿𝐿2𝑇𝑇𝑇𝑇 && 𝐷𝐷𝑐𝑐𝑐𝑐 > 𝐿𝐿3𝑇𝑇𝑇𝑇)
received data packets in forwarding table and transfers it 29. 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 = 𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 – 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖
later when those nodes meet new neighbors. The overall 30. 𝑃𝑃𝑃𝑃𝑖𝑖 = ρ × � 1 − 𝑖𝑖 � + 𝜔𝜔 × co s� 𝜐𝜐⃗𝑖𝑖 , 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 � + λ × 𝐿𝐿𝐿𝐿𝑐𝑐,𝑖𝑖
objective of the algorithm is to forward the packet as soon 31. 𝒇𝒇𝒇𝒇𝒇𝒇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑃𝑃𝑃𝑃𝑖𝑖 𝒅𝒅𝒅𝒅
as possible to increase packet delivery ratio, minimize the 32. 𝑃𝑃𝑃𝑃 = 𝑃𝑃𝑃𝑃𝑖𝑖
end to end delay and avoid packet loss. The MTR of a 33. 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖
vehicle/node is 250m.The other levels of transmission 34. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇
range is considerably less than MTR. The different levels 35. 𝒆𝒆𝒆𝒆𝒆𝒆𝒆𝒆 𝒊𝒊𝒊𝒊 (𝐷𝐷𝑐𝑐𝑐𝑐 < 𝐿𝐿3𝑇𝑇𝑇𝑇 && 𝐷𝐷𝑐𝑐𝑐𝑐 > 𝐿𝐿4𝑇𝑇𝑇𝑇)
of transmission range is shown in Fig.5 which includes, 36. 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 = 𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 – 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖
Maximum Transmission Range (i.e. MTR=250m) 37. 𝑃𝑃𝑃𝑃𝑖𝑖 = ρ × � 1 − 𝑖𝑖 � + 𝜔𝜔 × co s� 𝜐𝜐⃗𝑖𝑖 , 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 � + λ × 𝐿𝐿𝐿𝐿𝑐𝑐,𝑖𝑖
Level1 transmission range (i.e.L1TR=200m)
38. 𝒇𝒇𝒇𝒇𝒇𝒇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑃𝑃𝑃𝑃𝑖𝑖 𝒅𝒅𝒅𝒅
Level2 transmission range (i.e.L2TR=150m)
39. 𝑃𝑃𝑃𝑃 = 𝑃𝑃𝑃𝑃𝑖𝑖
Level3 transmission range (i.e.L3TR=100m) 40. 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖
Level4 transmission range (i.e.L4TR=50m). 41. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇
42. 𝒆𝒆𝒆𝒆𝒆𝒆𝒆𝒆 𝒊𝒊𝒊𝒊 (𝐷𝐷𝑐𝑐𝑐𝑐 < 𝐿𝐿4𝑇𝑇𝑇𝑇)
𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳: 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿2 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 = 150𝑚𝑚 44. 𝑃𝑃𝑃𝑃𝑖𝑖 = ρ × � 1 − � + 𝜔𝜔 × co s� 𝜐𝜐⃗𝑖𝑖 , 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 � + λ × 𝐿𝐿𝐿𝐿𝑐𝑐,𝑖𝑖
𝑳𝑳𝑳𝑳𝑳𝑳𝑳𝑳: 𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿𝐿3 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 = 100𝑚𝑚 45. 𝒇𝒇𝒇𝒇𝒇𝒇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑃𝑃𝑃𝑃𝑖𝑖 𝒅𝒅𝒅𝒅
𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄𝒄: 𝑡𝑡ℎ𝑒𝑒 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 47. 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖
𝒍𝒍𝒍𝒍𝒍𝒍𝒄𝒄 : 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 48. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇
𝒗𝒗𝒄𝒄 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑓𝑓𝑓𝑓𝑓𝑓 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑛𝑛𝑛𝑛 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 49. 𝒆𝒆𝒆𝒆𝒆𝒆𝒆𝒆
𝒅𝒅𝒅𝒅𝒅𝒅𝒅𝒅: 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 𝑜𝑜𝑜𝑜 𝑡𝑡ℎ𝑒𝑒 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 50. 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑡𝑡ℎ𝑒𝑒 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐
𝒍𝒍𝒍𝒍𝒍𝒍𝒅𝒅 : 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑓𝑓𝑓𝑓𝑓𝑓 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 51. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒊𝒊𝒊𝒊
𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏: 𝑡𝑡ℎ𝑒𝑒 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑎𝑎𝑎𝑎 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 ℎ𝑜𝑜𝑜𝑜 52. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇
𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒏𝒊𝒊 : 𝑡𝑡ℎ𝑒𝑒 𝒊𝒊𝑡𝑡ℎ 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑏𝑏𝑏𝑏𝑏𝑏 53. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇
𝒍𝒍𝒍𝒍𝒍𝒍𝒊𝒊 : 𝑡𝑡ℎ𝑒𝑒 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙 𝑜𝑜𝑜𝑜 𝑡𝑡ℎ𝑒𝑒 𝒊𝒊𝑡𝑡ℎ 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑏𝑏𝑏𝑏𝑏𝑏
𝝊𝝊𝒊𝒊 : 𝑡𝑡ℎ𝑒𝑒 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣𝑣 𝑜𝑜𝑜𝑜 𝑡𝑡ℎ𝑒𝑒 𝒊𝒊𝑡𝑡ℎ 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑏𝑏𝑏𝑏𝑏𝑏
Fig. 6 Pseudo code of EBGR Algorithm
1. 𝑙𝑙𝑙𝑙𝑙𝑙𝑐𝑐 ← 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔(𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐)
2. 𝑣𝑣𝑐𝑐 ← 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔(𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐)
���⃗ Step1: Neighbor nodes having distance between 250m and
3. 𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 ← 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔(𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑) 200m from the current node falls between MTR and
4. 𝐷𝐷𝑐𝑐 = 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑(locc , locd) L1TR. The potential score of all nodes present between the
5. ������⃗
𝑙𝑙𝑐𝑐,𝑑𝑑 = 𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 – 𝑙𝑙𝑙𝑙𝑙𝑙𝑐𝑐 transmission range of MTR and L1TR are calculated. The
6. 𝑃𝑃𝑃𝑃 = 𝜔𝜔 × cos⁡(����⃗, 𝑣𝑣𝑐𝑐 ������⃗
𝑙𝑙𝑐𝑐,𝑑𝑑 ) node which is having higher potential score is considered
7. 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 as edge node of the MTR. So the packet from the current
8. 𝒇𝒇𝒇𝒇𝒇𝒇 𝒂𝒂𝒂𝒂𝒂𝒂 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝒅𝒅𝒅𝒅 node is forwarded to that particular edge node. If no node
9. 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖 ← 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔( 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 ) present between MTR and L1TR, then L1TR and L2TR
10. 𝜐𝜐⃗𝑖𝑖 ← 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔(𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 ) are considered.
11. 𝐷𝐷𝑖𝑖 = 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑(𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 , 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖 )
12. 𝐷𝐷𝑐𝑐𝑐𝑐 = 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑(𝑙𝑙𝑙𝑙𝑙𝑙𝑐𝑐 , 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖 ) Step2: Neighbor nodes having distance between 200m and
13. 𝒇𝒇𝒇𝒇𝒇𝒇 𝒂𝒂𝒂𝒂𝒂𝒂 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑏𝑏𝑏𝑏𝑏𝑏𝑏𝑏 𝑜𝑜𝑜𝑜 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝐷𝐷𝑐𝑐𝑐𝑐 𝒅𝒅𝒅𝒅
150m from the current node falls between L1TR and
14. 𝒊𝒊𝒊𝒊 (𝐷𝐷𝑐𝑐𝑐𝑐 < 𝑀𝑀𝑀𝑀𝑀𝑀 && 𝐷𝐷𝑐𝑐𝑐𝑐 > 𝐿𝐿1𝑇𝑇𝑇𝑇)
L2TR. The potential score of all nodes present between the
15. 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 = 𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 – 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖
transmission range of L1TR and L2TR are calculated. The
16. 𝑃𝑃𝑃𝑃𝑖𝑖 = ρ × � 1 − 𝑖𝑖 � + 𝜔𝜔 × co s� 𝜐𝜐⃗𝑖𝑖 , 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 � + λ × 𝐿𝐿𝐿𝐿𝑐𝑐,𝑖𝑖 node which is having higher potential score is considered
17. 𝒇𝒇𝒇𝒇𝒇𝒇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑃𝑃𝑃𝑃𝑖𝑖 𝒅𝒅𝒅𝒅 as edge node of the L1TR.So the packet from the current
18. 𝑃𝑃𝑃𝑃 = 𝑃𝑃𝑃𝑃𝑖𝑖 node is forwarded to that particular edge node. If no node
19. 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 present between L1TR and L2TR, then L2TR and L3TR
20. 𝒆𝒆𝒆𝒆𝒆𝒆 𝒇𝒇𝒇𝒇𝒇𝒇 are considered.
21. 𝒆𝒆𝒆𝒆𝒆𝒆𝒆𝒆 𝒊𝒊𝒊𝒊 (𝐷𝐷𝑐𝑐𝑐𝑐 < 𝐿𝐿1𝑇𝑇𝑇𝑇 && 𝐷𝐷𝑐𝑐𝑐𝑐 > 𝐿𝐿2𝑇𝑇𝑇𝑇)
22. 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 = 𝑙𝑙𝑙𝑙𝑙𝑙𝑑𝑑 – 𝑙𝑙𝑙𝑙𝑙𝑙𝑖𝑖 Step3: Neighbor nodes having distance between 150m and
23. 𝑃𝑃𝑃𝑃𝑖𝑖 = ρ × � 1 − 𝑖𝑖 � + 𝜔𝜔 × co s� 𝜐𝜐⃗𝑖𝑖 , 𝑙𝑙⃗𝑖𝑖,𝑑𝑑 � + λ × 𝐿𝐿𝐿𝐿𝑐𝑐,𝑖𝑖 100m from the current node falls between L2TR and
24. 𝒇𝒇𝒇𝒇𝒇𝒇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛ℎ𝑖𝑖 𝑤𝑤𝑤𝑤𝑤𝑤ℎ 𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔𝑔 𝑃𝑃𝑃𝑃𝑖𝑖 𝒅𝒅𝒅𝒅 L3TR. The potential score of all nodes present between the
25. 𝑃𝑃𝑃𝑃 = 𝑃𝑃𝑃𝑃𝑖𝑖 transmission range of L2TR and L3TR are calculated. The
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 43

node which is having higher potential score is considered

as edge node of the L2TR.So the packet from the current
node is forwarded to that particular edge node. If no node
present between L2TR and L3TR, L3TR and L4TR are

Step4: Neighbor nodes having distance between 100m and

50m from the current node falls between L3TR and L4TR.
The potential score of all nodes present between the
transmission range of L3TR and L4TR are calculated. The
node which is having higher potential score is considered
as edge node of the L3TR.So the packet from the current
node is forwarded to that particular edge node. If no node
present between L3TR and L4TR, then L4TR are

Step5: Neighbor nodes having distance within 50m from

the current node falls to L4TR. The potential score of all
nodes present L4TR are calculated. The node which is
having higher potential score is considered as edge node of
Fig.7 Revival Mobility Model
the L4TR.So the packet from the current node is
forwarded to that particular edge node. If no node present
In this mobility, deterministic and instantaneous
in any of the above mentioned range, then the current node
transmission mechanism in which a message is available
store and carry the packet until it find some other node
for receiving within a certain radius r=250m from the
comes within its transmission range. The pseudo code of
sender with certainty, but unavailable further away.
ENS algorithm is illustrated in Fig.6.
Vehicles can unicast, multicast and broadcast packets to
the neighbor vehicle which is present within its
4. Simulation Results and Analysis transmission range.
In this section, we evaluate the performance of routing Table 2
protocols of vehicular networks in an open environment. Simulation Parameters
So among the routing protocols we aforementioned, we Parameter Value
choose GPSR, PDGR and EBGR for comparison. Simulation Area 1000m * 1000m
Number of Vehicles 20 - 100
4.1 Revival Mobility model (RMM) Average speed of vehicles 0 – 25 metre/second
Number of packet Senders 40
We use Revival Mobility model (RMM) to simulate the Transmission Range 250m
movement pattern of moving vehicles on streets or roads Constant Bit Rate 2 (Packets/Second)
defined by maps from the GPS equipped in the vehicles. Packet Size 512 Bytes
In Revival Mobility model (RMM), the road comprises of Vehicle beacon interval 0.5 (Seconds )
two or more lanes. Vehicles or nodes are randomly MAC Protocol 802.11 DCF
distributed with linear node density. Each vehicle can
move in different speed. This mobility model allows the The Simulations were carried out using Network
movement of vehicles in two directions. i.e. north/south Simulator (NS-2) ([20]). We are simulating the vehicular
for the vertical roads and east/west for the horizontal ad hoc routing protocols using this simulator by varying
roads. In cross roads, vehicles choose desired direction the number of nodes. The IEEE 802.11 Distributed
based on the shortest path. A security distance should be Coordination Function (DCF) is used as the Medium
maintained between two subsequent vehicles in a lane. Access Control Protocol. The packet size was fixed to 512
Overtaking mechanism is applicable and one vehicle can bytes. The Traffic sources are UDP. Initially the nodes
able to overtake the preceding vehicle. Packet transmission were placed at certain specific locations, and then the
is possible and can be done by vehicles moving in both nodes move with varying speeds towards new locations.
directions, which means front hopping and back hopping The nodes move with speeds up to 25 meter/sec. For
of data packet is possible as shown in the Fig.7. fairness, identical mobility and traffic scenarios were used
across the different simulations. The simulation
parameters are specified in Table 2
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 44

4.2. Performance Metrics to evaluate simulation

In order to evaluate the performance of vehicular ad hoc
network routing protocols, the following metric is 0.8
considered. 0.7

Packet Delivery Ratio

Packet delivery ratio (PDR): The ratio of the packets that 0.6
successfully reach destination.
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 0.4
𝑃𝑃𝑃𝑃𝑃𝑃 = × 100
𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑜𝑜𝑜𝑜 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 0.3
4.3 Packet Delivery Ratio vs. Number of Nodes
In this part, we compare the packet delivery ratio with 0
number of nodes and it is shown in Fig.8. Initially the 0 50 100 150 200 250
packet delivery ratio is less for GPSR, PDGR and EBGR.
When the number of node increases, then packet delivery Transmission Range (metres)
for all routing algorithms increases. More nodes in
network will provide more opportunities to find some GPSR PDGR EBGR
suitable node for forwarding of packet. When no node
available, GPSR switches to perimeter mode and it
increases delay of packet transmission, which results in Fig. 9: Packet Delivery Ratio vs. Transmission range
lower packet delivery ratio. PDGR have comparatively
high packet delivery ratio compared with GPSR. PDGR
considers its 2 hop neighbors for packet forwarding based 0.9
on prediction. But prediction is not reliable at all 0.8
situations. Due to high mobility, packet forwarded to edge 0.7
of the transmission range will be lost. In EBGR, the next
Packet Delivery Ratio

hop selection is done by considering the potential score.

By using EBGR, the packet loss is minimized 0.5
considerably and the packet delivery ratio is improved for 0.4
about 11.6% in comparison with PDGR with the increase
in number of vehicles.

1 0
0.9 0 5 10 15 20 25
Mobility (metre/second)
Packet Delivery Ratio


Fig. 10: Packet Delivery Ratio vs. Mobility.
4.4 Packet Delivery Ratio vs. Transmission Range
0.1 In this part, we compare the packet delivery ratio with
0 different levels of transmission range and it is shown in
20 40 60 80 100
Fig.9.The GPSR and PDGR always select the immediate
one hop and two hop neighbors respectively to forward the
Number of Nodes
packet. When many neighbor nodes are present, then
GPSR PDGR EBGR numbers of hops are increased in GPSR and PDGR. This
will decrease PDR. In EBGR, the vehicle always selects
the neighbor node based on different levels of transmission
Fig. 8: Packet Delivery Ratio vs. Number of Nodes range (i.e. L1TR, L2TR, and L3TR & L4TR) using
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 45

distance information from GPS. By using EBGR, the Tomasz Imielinske and Hank Korth, Eds., vol. 353. Kluwer
numbers of hops are minimized considerably and the Academic Publishers, 1996.
packet delivery ratio is improved for about 15.5% in [6]. Vincent D. Park and M. Scott Corson, “A highly adaptive
comparison with PDGR with the different levels of distributed routing algorithm for mobile wireless networks,”
in Proceedings of IEEE INFOCOMM, 1997, pp. 1405–
transmission range. 1413.
[7]. Charles E. Perkins and Elizabeth M. Royer, “Adhoc on-
4.5 Packet Delivery Ratio vs. Mobility demand distance vector routing,” in Proceedings of the 2nd
IEEE Workshop on Mobile Computing Systems and
In this part, we compare the packet delivery ratio with Applications, February 1999, pp. 1405–1413.
varying speed of vehicles and it is shown in Fig.10. When [8]. Josh Broch , David A. Maltz , David B. Johnson , Yih-Chun
the speed of vehicle increases, the packet delivery ratio of Hu , and Jorjeta Jetcheva , “A performance comparison of
GPSR and PDGR decreases much faster than others. The multi-hop wireless ad hoc network routing protocols,” in
high speed of vehicles leads to packet loss in edge of Proceedings of the Fourth Annual ACM/IEEE International
MTR. By using EBGR, the packet loss at the edge of MTR Conference on Mobile Computing and Networking
(MobiCom ’98), Dallas, Texas, U.S.A., October 1998, pp.
is minimized considerably and the packet delivery ratio is 85 – 97.
improved for about 12.5% in comparison with PDGR with [9]. P. Bose, P. Morin, I. Stojmenovic, and J. Urrutia, “Routing
the increase in speed of vehicles. with guaranteed delivery in ad hoc wireless networks,” in
Proc. of 3rd ACM Intl. Workshop on Discrete Algorithms
5. Conclusion and Methods for Mobile Computing and Communications
DIAL M99, 1999, pp. 48–55.
In this paper we have investigated routing aspects of [10]. Brad Karp and H. T. Kung, “GPSR: Greedy perimeter
VANETs. We have identified the properties of VANETs stateless routing for wireless networks,” in Proceedings of
and previous studies on routing in MANETs and the 6th Annual ACM/IEEE International Conference on
Mobile Computing and Networking (MobiCom 2000),
VANETs. We have commented on their contributions, and
Boston, MA, U.S.A., August 2000, pp. 243–254.
limitations. By using the uniqueness of VANETs, we have [11]. Stefano Basagni, Imrich Chlamtac, Violet R. Syrotiuk, and
proposed Revival Mobility Model and a new position Barry A.Woodward, “A distance routing effect algorithm
based greedy routing approach EBGR. Our simulation for mobility (dream),” in ACM MOBICOM ’98. ACM,
results have shows EBGR outperform GPSR and PDGR 1998, pp. 76 – 84.
significantly in the terms of improving the packet delivery [12]. Ljubica Blazevic , Silvia Giordano , and Jean- Yves Le
ratio. In the future, our approach requires modifications by Boudec , “Self-organizing wide-area routing,” in
taking into account the city environment characteristics Proceedings of SCI 2000/ISAS 2000,Orlando, July 2000.
and different mobility models with obstacles. Comparison [13]. C. Lochert, H. Hartenstein, J. Tian, D. Herrmann, H. Fubler,
M. Mauve: “A Routing Strategy for Vehicular Ad Hoc
of proposed EBGR approach with other existing approach
Networks in City Environments”, IEEE Intelligent Vehicles
shows that our routing algorithm is considerably better Symposium (IV2003).
than other routing algorithms in improving the packet [14]. C. Lochert, M. Mauve, H. Fler, H. Hartenstein. “Geographic
delivery ratio. Routing in City Scenarios” (poster), MobiCom. 2004, ACM
SIGMOBILE Mobile Computing and Communications
References Review (MC2R) 9 (1), pp. 69–72, 2005.
[15]. B.-C. Seet, G. Liu, B.-S. Lee, C. H. Foh, K. J. Wong, K.-K.
Lee. “A-STAR: A Mobile Ad Hoc Routing Strategy for
[1]. Charles E. Perkins and Pravin Bhagwat, “Highly dynamic Metropolis Vehicular Communications”, NETWORKING
destination-sequenced distance-vector routing (DSDV),” in 2004.
Proceedings of ACM SIGCOMM’94 Conference on [16]. H. Wu, R. Fujimoto, R. Guensler and M. Hunter. “MDDV:
Communications Architectures, Protocols and Applications, A Mobility-Centric Data Dissemination Algorithm for
1994. Vehicular Networks”, ACM VANET 2004.
[2]. T. H. Clausen and P. Jacquet. “Optimized Link State [17]. J. Zhao and G. Cao. “VADD: Vehicle-Assisted Data
Routing (OLSR)”, RFC 3626, 2003. Delivery in Vehicular Ad Hoc Networks”, InfoCom 2006.
[3]. Richard G. Ogier , Fred L. Templin , Bhargav Bellur , and [18]. Jiayu Gong, Cheng-Zhong Xu and James Holle. “Predictive
Mark G. Lewis , “Topology broadcast based on reverse-path Directional Greedy Routing in Vehicular Ad hoc
forwarding (tbrpf),” Internet Draft, draft-ietf-manet-tbrpf- Networks”, (ICDCSW’ 07).
03.txt, work in progress, November 2001. [19]. F. Granelli, G. Boato, and D. Kliazovich. Mora: a movement-
[4]. S. R. Das, R. Castaneda, and J. Yan, “Simulation based
based routing algorithm for vehicle ad hoc networks. In IEEE
performance evaluation of mobile, ad hoc network routing
protocols,” ACM/Baltzer Mobile Networks and Applications Workshop on Automotive Networking and Applications
(MONET) Journal, pp. 179–189, July 2000. (AutoNet 2006), San Francisco, U.S.A., Dec.2006.
[5]. David B. Johnson and David A. Maltz, “Dynamic Source [20]. The Network Simulator: ns2, http: // /ns/."
routing in ad hoc wireless networks,” in Mobile Computing,
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 46

K.Prasanth received the B.E degree in

Computer Science from K.S.
Rangasamy College of Technology,
Tiruchengode in 2005, M.Tech in
Computer Science from SRM
University, Chennai in 2007, MBA
(System) in Periyar University, Salem in 2007. He worked
as Project Engineer in Wipro Technologies, Bangalore
from (2007-2008). He is currently working as lecturer in
Department of Information Technology, K.S.Rangasamy
College of Technology. His Current research interest
includes Mobile Computing, Mobile Ad Hoc Networks
and Vehicular Ad Hoc Networks.

Dr.K.Duraiswamy received the B.E.,

M.Sc. and Ph.D. degrees, from the
University of Madras and Anna
University in 1965, 1968 and 1987
respectively. He worked as a Lecturer in
the Department of Electrical
Engineering in Government College of Engineering,
Salem from 1968, as an Assistant professor in Government
College of Technology, Coimbatore from 1983 and as the
Principal at K.S.Rangasamy College of Technology from
1995. He is currently working as a Dean in the Department
of Computer Science and Engineering at K.S.Rangasamy
College of Technology (Autonomous Institution).His
research interest includes Mobile Computing, Soft
Computing, Computer Architecture and Data Mining. He
is a senior member of ISTE, IEEE and CSI.

K. Jayasudha received the B.Sc degree in

2002, MCA degree in 2005, M.E degree
in Computer Science in 2007, from
K.S.Rangasamy College Of Technology,
Thiruchengodu. She is currently working
as Lecturer in Department Of Computer
Applications, K.S.R.College Of Engineering,
Thiruchengodu. Her current research interest includes Data
Mining, Vehicular Networks.

Dr.C.Chandrasekar received the B.Sc

degree and M.C.A degree. He
completed PhD in periyar university
Salem at 2006. He worked as Head of
the Department, Department Of
Computer Applications at
K.S.R.College of Engineering from 2007. He is currently
working as Reader in the Department of Computer Science
at Periyar University, Salem. His research interest includes
Mobile computing, Networks, Image processing, Data
mining. He is a senior member of ISTE, CSI.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 47
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Framework for Visualizing Model-Driven Software Evolution

and its Application
Akepogu ANAND RAO and Karanam MADHAVI

Computer Science & Engineering Department

Jawaharlal Nehru Technological University,
Anantapur, Andhra Pradesh, India.

Abstract The introduction of Model Driven Engineering (MDE)

Software Visualization encompasses the development and needs a new style of evolution i.e. Model-driven Software
evaluation of methods for graphically representing Evolution. The first fundamental premise [1] for Model-
different aspects of methods of software, including its Driven Software Evolution (MoDSE) is that evolution
structure, execution and evolution. Creating visualizations should be a continuous process. The second premise is
helps the user to better understand complex phenomena. It that reengineering of legacy systems to the model-driven
is also found by the software engineering community that of the paradigm should be done incrementally. MDE
visualization is essential and important. In order to introduces a multitude of languages that are themselves
visualize the evolution of the models in Model-Driven artifacts of the development process. Due to these
Software Evolution, authors have proposed a framework multitude languages in MoDSE, there is a need to have the
which consists of 7 key areas (views) and 22 key features model interaction, integration, mapping and
for the assessment of Model Driven Software Evolution transformation. Further there should be possible views to
process and addresses a number of stakeholder concerns. capture this information about models during the
The framework is derived by the application of the Goal evolution. For this purpose multiple views for MoDSE
Question Metric Paradigm. This paper aims to describe an have been proposed in [9]. Stakeholder’s involvement in
application of the framework by considering different MoDSE typically has interests in, or concerns relevant to
visualization tools/CASE tools which are used to visualize that system. The ability of models to evolve gracefully is
the models in different views and to capture the becoming a concern for many stakeholders. Due to
information of models during their evolution. Comparison different and interrelated models used to design an entire
of such tools is also possible by using the framework. system in MoDSE, the concerns of stakeholders may
Keywords: Model-Driven Software Evolution, Software differ from one role to another role that a stakeholder play
Visualization, Visualization tools. during the life time of a software project. So, visualization
provides better solution to understand the complex
information during evolution of the models. This can be
1. Introduction done by using the existing visualization and/or CASE
tools. Software Visualization tools use graphical
Visualization is used to enhance information techniques to make software artifacts visible.
understanding by reducing cognitive overload. Using
visualization methodologies and tools, people are often Evaluating a particular visualization tool for MoDSE is
able to understand the information presented in a shorter essential. Common practice is that some set of guidelines
period of time or to a greater depth. The term are followed and a qualitative summary is produced.
“visualization” can refer to the activity that people However, these guidelines do not usually allow a
undertake when building an internal picture about real comparison of competing techniques or tools. A
world or abstract entities. Visualizing can also refer to the comparison is important because it identifies possible
process of determining the mappings between abstract or flaws in the research area or software development. Thus,
real-world objects and their graphical representation. This a framework for describing attributes of tools is needed.
work uses the term “visualization” in the later sense: the Once the tools have been assessed in this common
process of mapping framework, a comparison is possible. However, a
the evolution of models to the stakeholder concerns. framework can be used for comparison, discussion, and
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 48

formative evaluation of the tools. Such framework was simple graphs with containers. The GMF project currently
proposed in [8]. So, the major contribution of this paper is lacks the ability to specify “Query Result” visualizations.
to show how the framework can be applied to compare the
Visualization Tools which is presented in section 4. A An Open Framework for [10] visual mining of CVS based
Framework for visualizing Model Driven Software software repositories has three major aspects are data
Evolution falls into seven key areas (views): Context extraction, analysis and visualization. An approach was
View, Inter-model View, City View, Metric View, proposed for CVS data extraction and analysis. CVS data
Transformation View, Evolution View and Evaluation acquisition mediator used to extract the data from CVS
view [9] and 22 Key features are identified for all key repositories. Analysis techniques are used to analyze the
areas. The framework is used to evaluate visualization raw data retrieved from the CVS repositories from CVS
tools and it is also used to assess tool appropriateness from Querying. It also provides the comparison of the open
a variety of stakeholder perspectives. source projects. CVSgraph is a software tool used to
visualizing project at file level. This open framework does
This paper is structured as follows: Section 2 discusses the not provide the visualization of models, it provides for
related work. Section 3 summarizes the framework. program at file level only.
Section 4 discusses an application of the framework by
considering different Visualization tools/CASE tools. CVSscan[11] is a tool in which a new approach for
Section 5 outlines the conclusions and giving an outlook visualization of software evolution was developed. The
on future work. main audience targeted here is the software maintenance
community. The main goal is to provide support for
program and process understanding. This approach uses
2. Related Work multiple correlated views on the evolution of a software
project. The overall evolution of code structure, semantics,
This section reviews the literature related to the fields of and attributes are integrated into an orchestrated
Software Visualization, Software Evolution Visualization environment to offer detail-on-demand. And also provides
and Model Driven approaches. the code text display that gives a detailed view on both the
composition of a fragment of code and its evolution in
Source Viewer 3D (sv3D) [6] is a Software Visualization time. It is focused on the evolution of individual files.
framework that builds on the SeeSoft metaphor. sv3D can
show large amounts of source code in one view. Object
based manipulation methods and simultaneous alternative 2.1 Motivation for Framework and its Application
mappings are available to the user. The types of user tasks
and interactions that are supported by sv3D, is not directly There are number of frameworks exists in the literature for
related to solving/visualizing specific software comparison and assessment of the various CASE tools.
engineering tasks and it is a prerequisite for a software Comparison of these tools is essential to understand their
visualization tool. differences, to ease their replication studies, and to
discover what tools are lacking. Such a comparison is
Architecture to Support Model Driven Software difficult because there is no well-defined comprehensive
Visualization [7], borrows the field of Model Driven and common comparative study for different category of
Engineering (MDE) to assist with the creation of highly the tools. For design recovery tools a comparative
customizable interfaces for Software Visualization. In framework [14] was derived for comparison. This
order to validate the architecture, MDV framework for framework comprises eight concerns, which were further
Eclipse was developed. Model Driven Visualization divided into fifty three criteria and which were applied on
(MDV) is intended to address the customization of ten design recovery tools successfully. Another
information visualization tools, especially in the program framework [7] also exists in the literature for comparison
comprehension domain. The MDV architecture describes and assessment of the software architecture visualization
how to leverage the work done in the Model Driven tools. Software architecture is the gross structure of a
Engineering community and apply it to the problem of system; as
designing visualizations tools. such, it presents a different set of problems for
visualization than those of visualizing the software at a
The Graphical Modeling Framework(GMF)[12] project lower level of abstraction. Six visualization tools were
for evaluated in this framework. This framework consists of
eclipse has facilities to allow modelers to define graphical seven Key areas and 31 Key features, for the assessment
editors for their data. These graphical editors can be used of software architecture visualization tools. Both the
as viewers, however, the views they support are limited to
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 49
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

frameworks applied against the stakeholders concerns. Table 1: Framework Summary

From this analysis it is easy to know that how a selected Key Key Areas
tool satisfies the stakeholder concerns. Thus the Features
motivation for this work lies in above mentioned two Key Area 1 : Context View (CV)
frameworks. Combining the visualization approach with CV 1 Does the visualization provide context of a model?
CV 2 Does the visualization provide the scope of a model or
MoDSE is essential to understand the evolution of models model element?
in a better way. Large numbers of visualization tools are CV 3 Does the visualization express the model completely
available in the literature. Among them many tools support including all its surrounding elements?
the evolution at source code level, data level. This work Key Area 2: Inter-Model View (IMV)
IMV 1 Does the visualization provide the dependencies between
aims to find out the visualization tools which support the the models and model elements?
visualization at model level. As such there is no IMV 2 Does the visualization provide the indirect dependencies
framework exists in the literature to evaluate tools which between the models and model elements?
are useful for the MoDSE Visualization and also to IMV 3 Does the visualization provide the integration of the two
or more models?
understand the evolution of the models with respect to
Key Area 3 : City View (CiV)
stakeholder perspectives. Hence this paper aims to CiV 1 Does the visualization provide the extendibility of the
evaluate the already proposed framework for visualization models in a software system?
of MoDSE. CiV 2 Does the visualization provide the traceability of a model
or model element?
Key Area 4 : Metric View (MeV)
3. Framework Summary MeV 1 Does the visualization provide the metrics to estimate the
impact analysis of the models during evolution?
This section provides the summary of the already MeV 2 Does the visualization provide the visualization
techniques to know the evolution of the models?
proposed framework for Model-Driven Software MeV 3 Does the visualization provide the metric values to know
Evolution visualization in [8]. the evolution of the models?
MeV 4 Does the metrics provide the knowledge about the quality
The framework has seven key areas (views) for visualizing and complexity of the models during evolution?
Key Area 5 : Transformation View (TV)
MoDSE: Context View, Inter-Model View, City View,
TV 1 Does the visualization provide any kind of
Metric View, Transformation View, Evolution View and transformation?
Evaluation view. These seven views are derived based on TV 2 Does the visualization provide the knowledge about the
the viewpoints and were discussed in detail [8]. The transformation of the models?
dimensions proposed in the framework are not proposed TV 3 Does the visualization provide the mapping of models?
TV4 Does it provide transformation rules?
as formal representation of the characteristics of MoDSE, TV5 Does it provide transformation language?
but are necessary for discussion about, and evaluation of, Key Area 6 : Evolution View (EV)
such dimensions with respect to stakeholders and tools EV 1 Does the visualization provide trends and causes for
which they use. The Goal/Question/Metric (GQM) evolution of models?
EV 2 Does the visualization provide the dimension of
paradigm [9] is used to identify the questions and then to evolution?
enable the formation of framework features. Key Area 7 : Evaluation View (EaV)
EaV 1 Does the visualization provide the evolution trends and
The primary goal of the framework is to assess and techniques?
understand the evolution of the models in model driven EaV 2 Does the visualization causes for the evolution of models?
EaV 3 Does the visualization facilitate the stakeholders’
software evolution. The framework is derived from an feedback?
extensive analysis of the literature in the area of software
visualization with special emphasis on model driven
software evolution. Each of the seven views is a
conceptual goal which the framework must satisfy. It is 4. Application of the Framework
this that makes the application of the GQM Paradigm [13]
straightforward. This section describes an application of the framework.
For this purpose tools which are mainly research oriented
Framework summary with its goals, questions are given in and non commercial tools are considered. These tools are
Table.1 First column represents the key features also having the features which are necessary for
(questions) which are abbreviated with view names. visualization of models. The expensive commercial tools
Second column represents the key areas (views). The such IBM rational Rose Suite, Enterprise architect etc. are
responses for these questions will be the values used in the not considered here. The following sub sections briefly
Table 2.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 50

describe the features of the tools. And the comparison of is ‘NA’. Compare to other two tools, ArgoUML is
those tools and responses are shown in the Table 3. particularly inspired by the three theories within Cognitive
Psychology. So, the designers of a complex system do not
Table 2: Possible Responses (Metrics) conceive a design fully formed. Instead, they must
Response Meaning
construct a partial design, evaluate, reflect on, and revise
Y Full support it, until they are ready to extend it further. So, the
Y? Mainly supported responses for the features are shown in the Table 3 as EV1
N? Mainly not supported – Y? , EV2 – N, EaV1 – N, Eav2 – Y? , EaV3 – Y.
N No support
NA Not applicable (not in the scope)
? Unable to determine
4.2 MetricView Evolution Tool

4.1 Argo UML Tool MetricView Evolution (MVE) tool [2], [3], [15] is a
research activity within Empirical Analysis of
ArgoUML (ARUML) is a free UML diagramming tool Architecture and Design Quality Project (EmpAnAda).
[5], [17] released under the open source BSD License. It is This Project is an activity of the System Architecture and
a java based UML tool that helps users to design using Networking group at the Eindhoven University of
UML. It is able to create and save most of the nine Technology, Netherlands. MetricView Evolution tool
standard UML diagrams. ArgoUML not only a free UML provides features such as metrics calculations within the
modeling tool, it is also an open source project that any tool, several views to explore and navigate UML models,
one can contribute to extend or to customize the features visualization of evolution data. This is an extension of
of a tool. It is a powerful yet easy-to-use interactive, MetricView tool which includes more features.
graphical software design environment that supports the MetricView Evolution also supports analysis of model
design, development and documentation of Object- quality and model evolution. Due to some limitations in
Oriented software applications. The users of Argo UML this research activity and since the entire UML
are software designers, architects, software developers, specification is quite complex so, not all the information
business analysts, system analysts and other professionals available in each diagram. Only the necessary elements are
involved in the analysis, design and development of extracted and displayed in this tool. Even with limitations
software applications. First version released in April 1998 the reasons to select this tool is research activity, easily
and the recent version is 0.26.2 in November 2008.All downloadable and features are closer to the framework.
nine UML 1.4 diagrams supported and it also supports
many features but the major weakness is no support for MetricView Evolution tool has full support (Y) for the key
UML 2. The four key features that make ArgoUML features such as CV1, CV2, CV3, IMV1, IMV2, CiV1,
different from other tools are: it makes use of ideas from CiV2, MeV1, MeV2, MeV3, EV1, EaV1, and EaV2.
cognitive psychology, it is based on open standards and it Feature IMV3 (i.e. integration of models) is not supported
is 100% pure java is used. but there is a scope for integrating the models. Features
such as TV1, TV2, and TV3 are not applicable because
Explorer View in ArgoUML has 9 perspectives which these features are not in the scope of the tool. And the
satisfy the features of the framework such as CV1, CV2, purpose of the MetricView Evolution tool is for quality
CV3, IMV1, and IMV2. This is indicated with the and evolution of UML models not for transformation of
response ‘Y’ in the Table 3. Integration of the models models like model to model, code to model and model to
(IMV3) is not supported so, the response is ‘N’. Features code. EV2 and EaV2 features are not mainly supported
such as CiV1, CiV2 are not mainly supported because as (N?) because the purpose of the evolution view in the tool
such there is no geographical view of a complete project is to enable the user to spot the trends in the values of
but it provides all the models in a project in a hierarchal quality attributes and/or metrics at multiple abstraction
tree view. Hence the response is ‘N?’ in the Table 3. All levels not for multiple dimensions of evolution. The
the features in a (Mev1, MeV2, MeV3, and MeV4) Metric responses for the stakeholders concerns(key features or
View are not applicable here because it is not intended to questions) are shown in the Table 3 in terms of Y, N, Y?,
calculate the metric values of the models. This is shown as N? and NA.
‘NA’ response. The response for the Transformation
features such as TV1, TV2, and TV3 is ‘N?’ because 4.3 Visual Paradigm for UML
transformation from model to code is partially available
not the other kinds of transformation such as model to Visual Paradigm [16] for UML 6.4 (VP-UML) is a
model or code to model. Transformation rules and powerful visual UML CASE tool. It is designed for a wide
language (TV4 and TV5) is not applicable, so the response range of users, including software engineers, system
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 51
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

analysts, business analysts, and system architects like who design problems and many design issues and rules is also
are interested in building software systems reliably available. The remaining two features like ‘transformation
through the use of Object-Oriented approach. VP-UML rules’ and ‘transformation languages’ are not applicable
can run in different operating systems. It supports more and not supported by these three tools. By comparing the
than 20 diagram types including UML 2.1, BPMN, tools under this common framework a stakeholder can
SysML, ERD, DFD and more. Different editions are also easily understand and asses the tools and can find out the
available such as Enterprise, Professional, Standard, flaws in a particular tool.
Modeler, and Personal are commercial editions.
Community and Viewer are non-commercial editions. It From the comparison of various features of the
supports a rich array of tools. One special feature is three tools it is observed that still there is a need to
Resource-Centric interface, which lets the user access consider few more possible visualization/CASE tools
modeling tools easily without referring back and forth which are exists in the literature. It is possible to check the
from the workspace to various toolbars. Users can draw unsatisfied features of the three tools can be satisfied by
diagrams or models as with a pen and paper, executing the other tools and also possible to know the role of the
complicated modifications with just a click and drag, visualization tools in MoDSE. From the comparison of
creating completely visual environment number possible tools framework can be strengthen
further. Another application of the framework is to
It is observed that the names of the features in VP-UML evaluate stakeholder concerns considered in the
differ from the features of the framework. But the purpose framework against the concerns of the software
and intention of the features are same. So, they have full practitioners (stakeholders) from diverse organizations.
support for those features that labeled as ‘Y’ in the Table These are the subjects of the future work.
3. Transformation of the models such as model to model,
model to code and code to model available in the tool but
transformation rules and languages are not available.
Hence, features as TV4, TV5 are not applicable (NA).
MeV1,MeV2, MeV3, MeV4 features for metrics of the
models and which are not in the VP-UML tool that is
shown in the Table 3 as ‘ NA’. Features such as EV2,
EaV3 are not mainly supported in the tool i.e. shown in
the Table 3 as ‘N?’ Visualization of the models by using
different diagrams is possible but the techniques are not
available. So, the response is ‘N?’ for Mev2.
Stakeholder’s feedback (EaV3-N?) is not mainly provided,
but the user can store their opinions/ideas about the
evolution of the models.

5. Conclusions and Future work

An application of the framework for visualizing Model-
Driven Software Evolution has been presented. The
research oriented, non commercial tools such as
ArgoUML, MetricView Evolution, and Visual Paradigm
for UML are considered for the framework’s application.
These three tools have compared successfully under this
common framework. From this comparison it is observed
that a single tool does not consist of all the features of the
framework and each tool has its own intensions and
purposes. But, by using these three tools all the features
are satisfied except four features. Among these two
features such as ‘multiple dimensions of evolution’, ‘stake
holder’s feedback’ are partially supported by the two
tools. But, ArgoUML has provided the feature such as
‘Cognitive Psychology’ which provides freedom for a
stakeholder (designer) to make design decisions, to resolve
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 52

Table 3: Framework Application References

Key Key Areas ARUML MVE VP [1] Arie van Deursen, Eelco Visser, and Jos Warmer.: ‘Model-
Features U Driven Software Evolution: A Research Agenda’, In Dalila
ML Tamzalit (Eds.), Proceedings 1st International Workshop on
Context View (CV) Model-Driven Software Evolution, University of Nantes, 2007.
CV1 Context of a model Y Y Y pp. 41-49.
[2] Christian F.J. Lange, Martijn A.M. Wijins, Michel R.V.
CV2 Scope of a model or Y Y Y Chaudron.: ‘Metric View Evolution: UML-based Views for
model element Monitoring Model Evolution and Quality’, IEEE 11th European
CV3 Express the model Y Y Y Conference on Software maintenance and Reengineering
completely including all (CSMR’07), Amsterdam, the Netherlands, March 2007, pp 327-
its surrounding elements
Inter-Model View
[3] C.F.J. Lange, M.A.M Wijns, M.R.V Chaudron, ‘ A
IMV1 Dependencies between Y Y Y Visualization Framework for Task-Oriented Modeling using
the models and model UML’, Proceedings of 40th Hawaii International Conference on
elements System Sciences (HICSS’07), Hawaii, January 2007,pp 289a .
IMV2 Indirect dependencies Y Y Y [4] Jacques Saraydaryan, Fatiha BenalilI and Stéphane Ubeda:’
between the models and Comprehensive Security Framework for Global Threads
model elements analysis’, IJCSI International Journal of Computer Science
IMV3 Integration of the two or N N Y Issues, Vol. 2, 2009 pp18-32.
more models [5]Jason Elliot Robbins, ‘Cognitive Support Features for
City View (CiV) Software Development Tools’, Ph.D Thesis Report, University
CiV1 Extendibility of the N? Y Y of California, Irvine, 1999.
models [6] Jonathan I.Maletic, Andrian Marcus, Louis Feng., ‘Source
CiV2 Traceability of a model N? Y Y Viewer 3D (sv3D) - A Framework for Software Visualization’,
or model element? Proceedings of the 25th International Conference on Software
Metric View (MeV) engineering (ICSE’03), 2003.
[7] Keith Gallegher, Andrew Hatch, and Malcolm Munro,
MeV1 Metrics to estimate the NA Y NA
‘Software Architecture Visualization: An evaluation Framework
impact analysis of the
models during evolution
and its Application’, IEEE Transactions on Software
MeV2 Visualization techniques NA Y N?
for the evolution
engineering, Vol. 34, No.2, March/April 2008, pp. 260-270.
models [8] K.Madhavi, A.Anand Rao, ‘A Framework for Visualizing
MeV3 Metric values to know NA Y NA Model-driven Software Evolution’, IEEE International Advanced
the evolution of the Computing Conference (IACC’09), March 2009, Patiala, Punjab,
models India, pp 1785-1790. Published in IEEE Xplore.
MeV4 Metrics for quality and NA Y? NA [9] K.Madhavi, A.Anand Rao, ’Model-Driven Software
complexity of the model Evolution- The Multiple Views’, International MultiConference
Transformation View of Engineers and Computer Scientists (IMECS 2009), March
(TV) 2009, Hong Kong, pp 1089-1094.
TV1 Kind of transformation N? NA Y [10] Lucian Voinea, Alexandru Telea.:’An Open Framework for
CVS Repository querying, analysis and Visualization’,
TV2 Knowledge about the N? NA Y Proceedings of the international workshop on Mining software
transformation of the repositories, 2006, pp 33 – 39.
[11]Lucian Voinea, Alex Telea and Jarke J.Van wijk, ‘CVSscan:
TV3 Mapping of the models N? N Y?
Visualization of code Evolution’, Proceedings of the ACM
symposium on Software visualization, 2005, pp 47 – 56.
TV4 Transformation Rules NA NA NA
[12]Richard Granback.: ‘Graphical Modeling framework’,
TV5 Transformation NA NA NA Borland Developer Conference Proceedings, 2005.
Language [13]V. Basili, G. Caldiera, and H.D Rombach, ‘The Goal
Evolution View (EV) Question Metric Paradigm’, Encyclopedia of Software Eng.,
EV1 Trends and causes for Y? Y NA
vol.2, John Wiley & Sons, 1994, pp.528-532.
evolution of models [14]Yann-Gael Gueheneuc, Kim Mens, Roel Wuyts,
EV2 Multiple dimensions of N N? N? ‘Comparative Framework for Design recovery Tools’, 10th
evolution European conference on Software Maintenance and
Evaluation View (EaV) Reengineering (CSMR 2006), Bari, Italy, March 206, pp
EaV1 Evolution trends and N Y NA
techniques [15]
EaV2 Causes for the evolution Y? Y NA
of models
EaV3 Stakeholders feedback Y N? N?
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 53
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

(Accessed in March 2006).

[16] accessed October 2008
[17] accessed 2000.

Dr. Anand Rao Akepogu. recieved B.Sc (M.P.C) degree from Sri
VENKATESWARA University, Andhra Pradesh, India. He received
B.Tech degree in Computer Science & Engineering from University
of Hyderabad, Andhra Pradesh, India and M.Tech degree in A.I &
Robotics from University of Hyderabad, Andhra Pradesh, India. He
received PhD degree from Indian Institute of Technology, Madras,
India. He is currently working as a Professor & HOD of Computer
Science & Engineering Department and also as a Vice-Principal of
JNTU College of Engineering, Anantapur, Jawaharlal Nehru
technological University, Andhra Pradesh, India. Dr. Rao published
more than twenty research papers in international journals and
conferences. His main research interest includes software
engineering and data mining.

Madhavi karanam. recieved B.E degree in Computer Science and

Engineering from Kuvempu University, Karanataka, India and
M.Tech degree in Software Engineering from Jawaharlal Nehru
Technological University, Andhra Pradesh, India. Currrently she is
pursuing PhD from Jawaharlal Nehru technological University,
Andhra Pradesh, India. Madhavi published six research papers in
national and international conferences. Her main research interest
includes software visualization, model driven development and
software engineering. She is a graduate member of the IEEE
Computer Society.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010
ISSN (Online): 1694-0784 54
ISSN (Print): 1694-0814

Unidirectional Error Correcting Codes for Memory

Systems: A Comparative Study
Muzhir AL-ANI 1 and Qeethara AL-SHAYEA 2
Faculty of IT, Amman Arab University
Amman, Jordan

MIS Department, Al-Zaytoonah University
Amman, Jordan

Abstract S. Al-Bassam [5] presents an improved method to

In order to achieve fault tolerance, highly reliable system often construct t error-correcting and all unidirectional error
require the ability to detect errors as soon as they occur and detecting codes (t-EC/AUED).
prevent the speared of erroneous information throughout the Umanesan and Fujiwara [6] propose a class of codes
system. Thus, the need for codes capable of detecting and called Single t/b-error Correcting—Single b-bit byte
correcting byte errors are extremely important since many
Error Detecting codes which have the capability of
memory systems use b-bit-per-chip organization. Redundancy
on the chip must be put to make fault-tolerant design available. correcting random t-bit errors occurring within a single
This paper examined several methods of computer memory b-bit byte and simultaneously indicating single b-bit byte
systems, and then a proposed technique is designed to choose a errors.
suitable method depending on the organization of memory Bose, Elmougy and Tallin [7] design some new classes
systems. The constructed codes require a minimum number of of t-unidirectional error-detecting codes over Zm.
check bits with respect to codes used previously, then it is Krishnan, Panigrahy and Parthasarathy [8] develop the
optimized to fit the organization of memory systems according error-correcting codes necessary to implement error-
to the requirements for data and byte lengths. resilient ternary content addressable memories. They
Keywords: Unidirctional Error Coding, Correcting Codes prove that the rate (ratio of data bits to total number of
Design, Error Detection and Correcting and Error
bits in the codewords) of the specialized error-correcting
Constructing Codes.
codes necessary for ternary content addressable
memories cannot exceed 1/t, where t is the number of bit
1. Introduction errors the code can correct.
Naydenova and Kløve [9] study codes that can correct up
In recent years, there has an increasing demand for to t symmetric errors and detect all unidirectional errors.
efficient and reliable data transmission and storage Biiinck and van Tilborg gave a bound on the length of
systems. Fujiwara [1] insists that before designing a binary such codes. They gave a generalization of this
dependable system, we need to have enough knowledge bound to arbitrary alphabet size. This generalized
of the system’s faults, errors, and failures of the Biiinck-van Tilborg bound, combined with constructions,
dependable techniques including coding techniques, and is used to determine some optimal binary and ternary
of the design process for practical codes. codes for correcting t symmetric errors and detecting all
Saitoh and Imai [2] represent codes that are capable of unidirectional errors.
correcting byte and detecting multiple unidirectional In computer memory, when data are stored in a byte-per-
bytes, but it is efficient code when b≤8. They also chip, byte errors may be occurring. When both one to
propose in [3] a code, but it is not efficient code for b≤8. zero and zero to one error may occur, but they do not
Zhang and Tu [4] propose a systematic t-EC/AUED occur simultaneously in a single byte, the errors are
codes which it's encoding and decoding is relatively called a unidirectional byte error, which is a kind of byte
easy, but it is efficient in the cases of t=1 and 2 and when error [10].
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 55

2. Coding Theory These memories are normally packaged with multiple bit
(or byte) per chip organization [13].
The theory and practice of error-correction coding is Coding techniques play a major role in segment the
concerned with protection of digital information against information in to m blocks each block of k-bit or it may
the errors that occur during data transmission or storage. be taken as a single block of length k (k=256, 512, 1024,
Many ingenious error correcting techniques based on a 2048, 8192, 16384, 32768, 65536, 131072, 262144,
vigorous mathematical theory have been developed and 524288) according to the organized memory system in
have many important and frequent applications. The our research. BCH and RS code are two powerful
current problem with any high-speed data approaches to error control coding in memory systems.
communication system such as storage medium is how to The information segmented is the first step when
control the errors that occur during storing data in information in a computer memory is written. Then this
storage medium. In order to achieve reliable k-bit encoded in to n-bit called code word which consist
communication, designers should develop good codes of k-bit and r-bit parity check (n=k+r). This code word
and efficient decoding algorithms [11]. stored in memory.
There are three types of faults transient, intermittent, and The decoding method used to obtain the information k
permanent faults. Transient faults are likely to cause a with no errors according to the coding technique when a
limited number of symmetric errors or multiple code word fetched from the storage.
unidirectional errors. Also, intermittent faults, because of
short duration, are expected to cause a limited number of 2.2 Reed-Solomon Codes (RS Codes)
errors. On the other hand, permanent faults cause either
symmetric or unidirectional errors, depending on the A RS code is a class of non binary BCH codes. It is also
nature of the faults. The most likely faults in some of the a cyclic symbol error-correcting code. The RS code
recently developed LSI/VLSI, ROM, and RAM represent a very important class of algebraic error-
memories (such as the faults that affect address decoders, correcting codes, which has been used for improving the
word lines, power supply, and stuck-fault in a serial bus, reliability of compact disc, digital audio tape and other
etc.) cause unidirectional errors. The number of data storage systems [14]. Secure communications
unidirectional errors cause by the above mentioned faults systems commonly use RS code as one method for
can be fairly large [12]. protection against jamming. RS codes are also used for
The errors that can occur because of the noise are many error control in the data storage systems, such as
and varied. However, they can be classified into three magnetic drums and photo digital storage systems.
main types: symmetric, asymmetric, and unidirectional A RS code is block sequence of finite field GF (2m) of 2m
errors [7]. binary symbols, where m is the number of bits per
symbol. This sequence of symbols can be viewed as the
2.1 Error Control for Computer Main Memories coefficients of code polynomial C(x)=c0+c1x+c2x²+…+cn-
1x where the field elements Ci are from GF(2m) [10].
Error correcting codes have been used to enhance the A t-error-correcting RS code with symbols from FG(2m)
reliability and data integrity of computer memory has the following parameters:
systems. The error correction can be incorporated in to Code length : n=2m-1
the hardware. Number of information : k=n-2t
In particular the class of single error-correcting and Number of parity-check digits : n-k=2t
double error-detecting (SEC-DED) binary codes has Minimum distance : dmin=2t+1
been successfully used to correct and detect errors In the following, we shall consider Reed-Solomon codes
associated with failures in semiconductor memories. The with code symbols from the Galois field GF(2m). The
most effective organization is the so-called 1 bit per chip generator polynomial of a t-error-correcting Reed-
organization. In this organization, all bits of a code word Solomon code of length 2m-1 is
2 2t
are stored in different chips. Any type of failures in a g(x)=(x+α)(x+α )…(x+α ), where α is a primitive
chip can corrupt at the most 1 bit of the code word. As element of GF(2m), and the coefficients gi, 0≤ l ≤2t are
long as the errors do not line up in the same code word, also from GF(2m). An (n,k) RS code generated by g(x) is
multiple errors in the memory are correctable. Large an (n,n-2t) cyclic code whose code vectors are multiples
scale integration (LSI) and very large scale integration of g(x) [14,15].
(VLSI) memory systems offer significant advantages in Consider RS codes with symbols from GF(2m), where m
size, speed, and weight over earlier memory systems. is the number of bits per symbol.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010
ISSN (Online): 1694-0784 56
ISSN (Print): 1694-0814

Let d(x)=cn-kxn-k+cn-k+1xn-k+1+…+cn-1xn-1 be the This code is optimal, thus it is the only SbEC-DbEC
information polynomial and p(x)=c0+c1x+…+cn-k-1xn-k-1 code with three check bytes but for a given size b(b<16)
be the check polynomial. Then the encoded RS code there is only one or two value of information.
polynomial is expressed by:
c( x) = p ( x) + d ( x) (1)
where ci,0≤ l ≤n-1, are field elements in GF(2 ). Thus, a Table 1: The parameters of shortened modified RS code
vector of n symbols, (c0,c1,…,cn-1) is a code word if and b n n k
only if its corresponding polynomial c(x) is a multiple of 5 16 79 64
the generator polynomial g(x). The common method of 6 46 274 256
encoding a cyclic code is to find p(x) from d(x) and g(x), 7 77 533 512
which results in an irrelevant quotient q(x) and an 8 131 1048 1024
important remainder y(x). That is, 9 231 2075 2048
(2) 10 823 8222 8192
d ( x) = q ( x) g ( x) + y ( x)
11 1493 16417 16384
Substituting Eq. (1) in to (2) gives:
12 2734 32804 32768
c( x) = p ( x) + q ( x) g ( x) + y ( x) (3)
13 5045 65575 65536
If we define the check digits as the negatives of the 14 9366 131114 131072
coefficients of y(x), i.e, p(x) = -y(x), it follows that: 15 17480 262189 262144
c( x) = q ( x) g ( x) (4)
This ensures that the code polynomial c(x) is multiple of Let the two codes whose H0 matrices are denoted as Hv
g(x). Thus, the RS encoder will perform the above and Hw have minimum Hamming distance dmin=4
division process to obtain the check polynomial p(x) GF(2b), let vi, i=0,1,…,n-1, denote a column vector in the
[14]. matrix Hv. Preserving minimum distance, matrix Hw
Theorem 1: A Reed-Solomon code is a maximum converted to matrix Hw having an all 'I' row vector.
distance code, and the minimum distance is n-k+1. Next, this all 'I' row vector is removed from the matrix
This tells us that for fixed (n,k), no code can have a Hw, whose resultant matrix is now called Hw. Let vj,
larger minimum distance than a RS code. This is often a j=0,1,…,m-1, denote a column vector of matrix Hw. The
strong justification for using RS codes. RS codes always new code has a parity check matrix H1 of the form that
have relatively short block length as compared to other each column in it is defined by the following equation:
cyclic codes over the same alphabet [16]. (6)
(C ) T = (V W )
In decoding a RS code (or any non binary BCH code), ij i j

the same three steps used for decoding a binary BCH i=0,1,…,n-1, and j=0,1,…,m-1.The dmin of this code is
code are required, in addition a fourth step involving four over GF(2b).
calculation of the error value is required. The error value For example, let b=2, and Hw equal to
at the location corresponding to B1 is given by the
following equation: ⎡Ι Ι Ι Ι 0 0⎤
Z (βL−1 ) (5) Ηw = ⎢⎢Ι Τ Τ2 0 Ι 0⎥⎥
ei1 =
π ν Ι≈ ⎢⎣Ι Τ 2 Τ 0 0 Ι ⎥⎦
Where z(x) = 1 + (s1+σ1)x + (s2+σ1s1 + σ2)x2+…+
(sv+σsv-1+σ2sv-2+…+σv)xv where
The decoding method of RS code is worth mentioning
because of its considerable theoretical interest, even ⎡1 0 ⎤ ⎡0 1⎤ ⎡1 1⎤ ⎡0 0 ⎤
Ι=⎢ ⎥ Τ1 = ⎢ ⎥ Τ2 = ⎢ ⎥ 0=⎢ ⎥
though it is impractical [15]. ⎣0 1 ⎦ ⎣1 1⎦ ⎣1 0⎦ ⎣0 0 ⎦
This matrix can be converted to the new form that has
3. Byte-Per-Chip Memory Organization top row vector which has all 'I' elements. This conversion
can be carried out in the following manner.
In many computer memory and VLSI circuits The second row of Hw is multiplied by an arbitrary non
unidirectional errors are known to be predominant zero element Т^a. The multiplied result and the third row
protection must be against combinations of unidirectional vector are added to the first row vector in Hw. If the
and random errors because random byte errors also added row vector has non zero element, each column can
appear from intermittent faults in memories. Thus it is be normalized so that the first row element has a 'I'
very important to have such codes for protection of byte element. It can be derived that the number of Т^a
organized memories. Table (1) shows the parameters of elements is 2b-1. If Тa=Т' is chosen then:
modified RS code after shortening.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 57

⎡Ι Ι Ι Ι 0 0⎤ Theorem 2 [17]: Let H be the parity check matrix of a

Η w′ = ⎢⎢ Ι Τ2 Τ2 0 Ι 0⎥⎥ (n,n-r) linear SbEC-DbED code over GF(2b). The (2n,
2n-r-1) linear code over GF(2b) defined by the parity
⎢⎣Τ 2 Ι Τ2 0 0 Ι ⎥⎦
check matrix H'.
⎡Ι Τ2 Τ2 0 Ι 0⎤ (9)
Η w′′ = ⎢ 2 ⎥ ⎡0 0 K 0 1 1K 1⎤ (13)
⎣Τ Ι Τ2 0 0 Ι⎦ Η′ = ⎢ ⎥
⎣ Η Η ⎦
Here Hv as the H0 matrix shown in Eq. (7) is adapted. An
Eq. (13) is a SbEC-DbED code.
S2EC-D2ED code, whose H1 matrix has five rows, can
Table (4) is obtained after applying theorem 2 to the
be constructed from Hv and Hw''.
parameters in table (2). Table (5) shows the results when
In the same manner, the SbEC-DbED codes whose H0
the parameters are shorten.
matrices have odd number are obtained in the same way.
After obvious comparison between the parameters in
If even number of rows is required (in this example),
table (1) and parameters in table (3), we observe that
matrix Hw' can be shown as follows: there is no table with the best parameters for all value of
⎡ Ι Ι⎤ (10) k, so for the best parameters obtained table (6) is
Η w′ = ⎢ ⎥
⎣0 Ι ⎦ presented.
Since the two chip failure no longer takes place at the
The code length N for the proposed codes is given as
same time, these parameters can be used. Codes for only
follows: SbEC-DbED are proposed. So these codes can not
recognize all the unidirectional errors which occur in b-
N = ( 2 b + 2) ( r −1) / 2 r:odd (≥3) (11)
bit-per-chip memory organization. Wherefore code that
r:even (≥4) (12) fits memory organized in b-bit-per-chip fashion, and
N = 2(2 b + 2) ( r − 2) / 2
4<b<16 is constructed.
Table 2: The parameters of SbEC-DbED RS codes
b=5 b=6 b=7
r n k n k n k 4. Conclusions
3 170 155 396 378 910 889
4 340 320 792 768 1820 1792 The most likely faults in many computer memories cause
5 5780 5755 26136 26106 11300 118265 unidirectional errors, thus a detection of unidirectional
6 11560 11530 52272 52236 236600 236558 errors is required. In addition, byte-error-
b=8 b=9 b=10 correcting/detecting codes are useful for protection
r n k n k n k against byte errors which tend to occur when data are
3 2064 2040 4626 4599 10260 10230 stored in byte-per-chip memory organization. A
4 4128 4096 9252 9216 20520 20480 proposed technique for constructing SbEC-DbED codes
b=11 b=12 b=13 is presented in this paper that can be practically applied
r n k n k n k to large capacity memory units. The obtained results
3 22550 22517 49176 49140 106522 106483 indicate that the proposed technique is suitable and
4 45100 45056 98352 98304 213044 212992 efficient for memory system to recognize unidirectional
b=14 b=15 errors that occur in bit-per chip memory organization.
r n k n k
3 229404 229362 491550 491505
4 458808 458752
[1] E. Fujiwara, Code Design for Development Systems:
Theory and Practical Applications, John Wiley and Sons, Inc.,
It is important to know that r is a parity check digits in 2006.
bits, n is the code word length and k is the information [2] Y. Saitoh and H. Imai, Multiple Unidirectional Byte Errors-
length in all tables observer in this paper. Correcting Codes, IEEE Transactions on Information Theory,
The Parameters of SbEC-DbED RS codes are illustrated Vol. 37, No. 3, May, 1991, pp.903-908.
in table (2). When the code is shortening table (3) is [3] Y. Saitoh and H. Imai, Generalized Concatenated Codes for
obtained. Channels where Unidirectional Byte Errors are Predominant,
It is obvious from comparing the parameters in table (1) IEEE Transactions on Information Theory, Vol. 39, No. 3,
May, 1993, pp.1014-1022,.
with the parameters in table (3) that the parameters in
table (3) are more efficient than the parameters in table
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010
ISSN (Online): 1694-0784 58
ISSN (Print): 1694-0814

[4] Z. Zhang and C. Tu, On The Construction of Systematic IEEE Transactions on Information Theory, Vol. 38, No. 6,
TEC/AUED Codes, IEEE Transactions on Information Theory, Nov., 1992, pp. 1801-1807.
Vol. 39, No. 5, Sep., 1993, pp.1662-1669. [15] S. Lin, An Introduction to Error-Correcting Codes,
[5] S. Al-Bassam, Another Method for Constructing t- Prentice-Hall, Electrical Engineering Series, New Jersey, 1970.
EC/AUED Codes, IEEE Transactions on Computers, Vol. 49, [16] R. E. Blahut, Theory and Practice of Error Control Codes,
No. 9, Sep., 2000, pp. 964-967. Addison-Wesley, 1983.
[6] G. Umanesan and E. Fujiwara, class of Random Multiple [17] C. L. Chen, Byte-Oriented Error-Correcting Codes for
Bits in a Byte Error Correcting and Single Byte Error Detecting Semiconductor Memory Systems, IEEE Transactions on
(S[sub t/b]EC-S[sub b]ED) Codes, IEEE Transactions on Computers, Vol. C-35, No. 7, July, 1986, pp.646-648.
Computers, Vol. 52, No. 7, Jul., 2003, pp. 835-848.
[7] B. Bose, S. Elmougy, and L. G. Tallini, Systematic t- Muzhir. Shaban Al-Ani has received Ph. D. in Computer &
Unidirectional Error-Detecting Codes over Zm, IEEE Communication Engineering Technology, ETSII, Valladolid
Transactions on Computers, Vol. 56, No. 7, July, 2007, pp. University, Spain, 1994. Assistant of Dean at Al-Anbar Technical
Institute (1985). Head of Electrical Department at Al-Anbar
Technical Institute, Iraq (1985-1988), Head of Computer and
[8] S. Krishnan, R. Panigrahy and S. Parthasarathy, Error- Software Engineering Department at Al-Mustansyria University,
Correcting Codes for Ternary Content Addressable Memories, Iraq (1997-2001), Dean of Computer Science (CS) & Information
IEEE Transactions on Computers, Vol. 58, No. 2, Feb., 2009, System (IS) faculty at University of Technology, Iraq (2001-
pp. 275-279. 2003). He joined in 15 September 2003 Electrical and Computer
[9] I. Naydenova and T. Kløve, Some Optimal Binary and Engineering Department, College of Engineering, Applied
Ternary t-EC-AUED Codes, IEEE Transactions on Computers, Science University, Amman, Jordan, as Associated Professor.
Vol. 55, No. 11, Nov., 2009, pp. 4898-4904. He joined in 15 September 2005 Management Information
System Department, Amman Arab University, Amman, Jordan,
[10] G. Fang and H. C. A. Van Tilborg, "Bounds and
as Associated Professor, then he joined computer science
Constructions of Asymmetric or Unidirectional Error Code", department in 15 September 2008 at the same university.
Ablicable Algebra in Engineering Communication and
Computing (AAECC), 3, 1992, pp. 269-300. Qeethara Kadhim Abdul Rahman Al-Shayea has received Ph.
[11] M. Y. Rhee, Error Correcting Coding Theory, McGraw- D. in Computer Science, Computer Science Department,
Hill, New York, 1989. University of Technology, Iraq, 2005. She received her M.Sc
[12] D. J. Lin and B. Bose, Theory and Design of t-Error degree in Computer Science, Computer Science Department
Correcting and d (d>t)- Unidirectional Error Detecting (t-EC/d- from University of Technology, Iraq, 2000. She has received her
High Diploma degree in information Security from Computer
UED) Codes, IEEE Transactions on Computers, Vol. 37, No. 4,
Science Department, University of Technology, Iraq, 1997. She
Apri., 1988, pp.433-439. joined in 15 September (2001-2006), Computer Science
[13] C. L. Chen, Error-Correcting Codes with Byte Error- Department, University of Technology, Iraq as assistant
Detection Capability, IEEE Transactions on Computers, Vol. professor. She joined in 15 September 2006, Department of
C-32, No. 7, July, 1983, pp.615-621. Management Information Systems Faculty of Economics &
[14] M. Morii and M. Kasahara, Generalized Key Equation of Administrative Sciences Al-Zaytoonah University of Jordan as
Remainder Decoding Algorithm for Reed-Solomon Codes, assistant professor. She is interested in Coding Theory,
Computer Vision and Artificial Intelligence.

Table 3: The parameters of shortened SbEC-DbED RS codes

b=5 b=6 b=7
r n k r n k r n k
3 47 32 3 82 64 3 149 128
79 64 146 128 277 256
143 128 274 256 533 512
4 276 256 4 536 512 4 1052 1024
5 537 512 5 1054 1024 5 2083 2048
1049 1024 2078 2048 4131 4096
2073 2048 4126 4096 8227 8192
4121 4096 8222 8192 16419 16384
6 8222 8192 16414 16384 32803 32768
7 16419 16384 6 32804 32768 65571 65536
32803 32768 7 65578 65536 6 131114 131072
65571 65536 131114 131072 7 262193 262144
131107 131072 262186 262144 524337 524288
8 262184 262144 524330 524288
9 524333 524288
b=8 b=9 b=10
r n k r n k r n k
3 280 256 3 539 512 3 1054 1024
536 512 1051 1024 2078 2048
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 59

1048 1024 2075 2048 4126 4096

4 2080 2048 4123 4096 8222 8192
4128 4096 4 8228 8192 4 16424 16384
5 8232 8192 5 16429 16384 5 32818 32768
16424 16384 32813 32768 65586 65536
32808 32768 65581 65536 131122 131072
65576 65536 131117 131072 262194 262144
131112 131072 262189 262144 524338 524288
262184 262144 524333 524288
524328 524288
b=11 b=12 b=13
r n k r N k r n K
3 2081 2048 3 4132 4096 3 8231 8192
4129 4096 8228 8192 16423 16384
8225 8192 16420 16384 32807 32768
16417 16384 32804 32768 65575 65536
4 32812 32768 4 65584 65536 4 131124 131072
5 65591 65536 131132 131072 5 262209 262144
131127 131072 262204 262144 524353 524288
262199 262144 524348 524288
524343 524288
b=14 b=15
r n k r N k
3 16426 16384 3 32813 32768
32810 32768 65581 65536
65578 65536 131117 131072
131114 131072 262189 262144
4 262200 262144 4 524348 524288
524344 524288

Table 4: The parameters of new SbEC-DbED RS codes

b=5 b=6 b=7
r n k r n K r n k
16 340 324 19 792 773 22 1820 1798
21 680 659 25 1584 1559 29 3640 3611
26 11560 11534 31 52272 52241 36 236600 236564
b=8 b=9 b=10
r n k r n K r n k
25 4128 4103 28 9252 9224 31 20520 20489
33 8256 8223 37 18504 18467 41 41040 40999
b=11 b=12 b=13
r n k r n K r n k
34 45100 45066 37 98352 98315 40 213044 213004
45 90200 90155 49 196704 196655 53 426088 426035
b=14 b=15
r n k r n K
43 458808 458765 46 983100 983054

Table 5: The parameters of shortened new SbEC-DbED RS codes

b=5 b=6 b=7
r n k r n k r n k
16 48 32 19 83 64 22 150 128
80 64 147 128 278 256
144 128 275 256 534 512
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010
ISSN (Online): 1694-0784 60
ISSN (Print): 1694-0814

272 256 531 512 1046 1024

21 533 512 25 1049 1024 29 2077 2048
26 1050 1024 31 2079 2048 36 4132 4096
2074 2048 4127 4096 8228 8192
4122 4096 8223 8192 16420 16384
8218 8192 16415 16384 32804 32768
31 16415 16384 326799 32768 65572 65536
36 32804 32768 37 65573 65536 131104 131072
65572 65536 43 131115 131072 43 262187 262144
131108 131072 262187 262144 50 524338 524288
262180 262144 524331 524288
524329 524288
b=8 b=9 b=10
r n k r N K r n k
25 281 256 28 540 512 31 1055 1024
537 512 1052 1024 2079 2048
1049 1024 2076 2048 4127 4096
2073 2048 4124 4096 8223 8192
4121 4096 8220 8192 16415 16384
33 8225 8192 37 16421 16384 41 32809 32768
41 16425 16384 46 32814 32768 51 65587 65536
32809 32768 65582 65536 131123 131072
65577 65536 131118 131072 262195 262144
131113 131072 262190 262144 524339 524288
262185 262144 524334 524288
524329 524288
b=11 b=12 b=13
r n k r n k r n k
34 2080 2048 37 4133 4096 40 8232 8192
4130 4096 8229 8192 16424 16384
8226 8192 16421 16384 32808 32768
16418 16384 32805 32768 65576 65536
32802 32768 65573 65536 131112 31072
45 65581 65536 49 131121 131072 53 262197 262144
56 131128 131072 61 262205 262144 66 524354 524288
262200 262144 524349 524288
524344 524288
b=14 b=15
r n k r n k
43 16427 16384 46 32814 32768
32811 32768 65582 65536
65579 65536 131118 131072
131115 131072 262190 262144
262187 262144 524334 524288
524345 524288

Table 6: Best parameters obtained

b=5 b=6 b=7
(n,k) Table (n,k) Table no. (n,k) Table no.
(47,32) (3) (82,64) (3) (149,128) (3)
(79,64) (1,3) (146,128) (3) (277,256) (3)
(143,128) (1,3) (274,256) (1,3) (533,512) (1,3)
(272,256) (5) (531,512) (1) (1046,1024) (5)
(533,512) (5) (1049,1024) (1) (2077,2048) (5)
(1049,1024) (3) (2078,2048) (3) (4131,4096) (3)
(2073,2048) (3) (4126,4096) (3) (8227,8192) (3)
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 61

(4121,4096) (3) (8222,8192) (3) (10419,16384) (3)

(8218,8192) (5) (16414,16384) (3) (32803,32768) (3)
(16415,16384) (5) (32799,32768) (5) (65571,65536) (3)
(32803,32768) (3) (65573,65536) (5) (131108,131072) (5)
(65571,65536) (3) (131114,131072) (3) (262187,262144) (5)
(131107,131072) (3) (262186,262144) (3) (524377,524288) (3)
(262180,262144) (5) (524330,524288) (3)
(524329,524288) (5)
b=8 b=9 b=10
(n,k) Table no. (n,k) Table no. (n,k) Table no.
(280,256) (3) (539,512) (3) (1054,1024) (3)
(536,512) (3) (1051,1024) (3) (2078,2048) (3)
(1048,1024) (1,3) (2075,2048) (1,3) (4126,4096) (3)
(2073,2048) (5) (4123,4096) (1,3) (8222,8192) (1,3)
(4121,4096) (5) (8220,8192) (5) (16415,16384) (5)
(8225,8192) (5) (16421,16384) (5) (32809,32768) (5)
(16424,16384) (3) (32813,32768) (3) (65586,65536) (3)
(32808,32768) (3) (65581,65536) (3) (131122,131072) (3)
(65576,65536) (3) (131117,131072) (3) (262194,262144) (3)
(131112,131072) (3) (262189,262144) (3) (524338,524288) (3)
(262184,262144) (3) (524333,524288) (3)
(524328,524288) (3)
b=11 b=12 b=13
(n,k) Table no. (n,k) Table no. (n,k) Table no.
(2081,2048) (3) (4132,4096) (3) (8231,8192) (3)
(4129,4096) (3) (8228,8192) (3) (16423,16384) (3)
(8225,8192) (3) (16420,16384) (3) (32807,32768) (3)
(16417,16384) (1,3) (32804,32768) (1,3) (65575,65536) (1,3)
(32802,32768) (5) (65573,65536) (5) (131112,131072) (5)
(65581,65536) (5) (131121,131072) (5) (262197,262144) (5)
(131127,131072) (3) (262204,262144) (3) (524353,524288) (3)
(262199,262144) (3) (524348,524288) (3)
(524343,524288) (3)
b=14 b=15
(n,k) Table no. (n,k) Table no.
(16426,16384) (3) (32813,32768) (3)
(32810,32768) (3) (65581,65536) (3)
(65578,65536) (3) (131117,131072) (3)
(131114,131072) (1,3) (262189,262144) (3)
(262187,262144) (5) (524334,524288) (5)
(524344,524288) (3)
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 62
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

ICT in Universities of the Western Himalayan Region of India II:

A Comparative SWOT Analysis
Dhirendra Sharma1 and Vikram Singh2
University Institute of Information Technology, Himachal Pradesh University,
Shimla, Himachal Pradesh 171 005, India

Department of Computer Science and Engg, Ch. Devi Lal University,
Sirsa, Haryana 125 055, India

Abstract priority which would produce skilled professionals as

This study presents a comparative SWOT analysis to good teachers and should be adopted by the universities as
comprehend the pattern of development of ICT within six their sacred responsibility. Only then the challenges faced
universities of western Himalayan region of India. With the by the country due to widening of the gap between the
objective of achieving quality and excellence in higher education haves and have-nots, arising from the technological
system in the region, this study provides a basis to decision
development, may be bridged.
makers to exploit opportunities and minimize the external threats.
The SWOT analysis of different universities, placed under three
categories, has been undertaken within the four-tier framework
used earlier by the authors. Guided by the initiatives of National 1.1 SWOT as an Analytical Technique
Mission on Education through ICT (NMEICT) for SWOT
analysis, findings of this paper reveal, relative consistency of The origin of the SWOT, as an analytical technique, lies in
these three categories of universities, with the earlier study. A connection with the growth of strategic planning which
few suggestions, as opportunities, with an emphasis on problem dates back to the decade of 1960s. The concept was
solving orientation in higher education, have been made to developed later, to address possible shortcomings in the
strengthen the leadership of universities in the field of ICT. outcome of strategic planning [2,9,10].
Keywords: SWOT Analysis, Strategic Planning, ICT,
Information System, Four-tier Framework, NMEICT Initiatives, SWOT has established itself as a framework for analyzing
Enterprise Resource Planning (ERP).
strengths, weaknesses, opportunities and threats. Strengths
and weaknesses are mainly based on internal audit, as a
result of introspection of a universities/organization. The
1. Introduction
opportunities are related with the internal as well as
SWOT (Strength, weakness, opportunities and threats) external environmental factors. Threats are concerned
analysis has proved to be a general tool at the preliminary mainly with the external environment factors. The external
stages of policy making and strategic planning of an factors imply economy, competition, sources of funding,
organization and at a later stage as well, while analyzing demographics and culture. These are needed to be taken
the performance and planning for further development and care in strategic planning and activities. Opportunities
progress of the organization. For the latter, the SWOT represent factors that can be beneficially exploited.
analysis serves double purpose of getting the answers to Threats need to be considered because of the potential of
some very relevant questions from the performance and damaging the organization/ institute. SWOT analysis
planning for future development, success and failure/ normally reflects a viewpoint which can be used by others.
difficulties faced by the universities/organization. In view It has to be positive so that the analysis is exploited for the
of that experience, one looks for opportunities provided by benefit of the organisation. Different variants of SWOT
such an analysis. A saying is that opportunities knock the [6,16] provided a structure based planning [3,14] and
door at least once which is to be promptly identified and implementation. This technique is used to develop a
utilized. project or find a solution to a problem that takes care of
several different internal and external factors and
In this connection, Narayana Murthy [11], non-executive maximizes the potential of strengths and opportunities
chairman of Infosys Technologies, Bangalore, India, while while minimizing the effect of weakness and threats.
addressing a convocation in Jammu, admonished that In a recent paper [5], presented a comparative SWOT
problem- solving oriented education system should be the analysis of four universities in the pacific Asian region in
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 63

the context of information system. Supported by their country with a view to achieve critical mass of skilled
historical perspective, they elaborated on policies, their human resource/ researchers in any given field.
implementation and strategic management in those
universities. The historical perspective gives a bird’s eye These guiding principles are expected to lead to various
view of evolution of the university in ICT field, which important steps in planning and implementation as
can be quite revealing. A retrospective analysis of this follows:
perspective may be useful in documenting key changes
over time. With this, the SWOT analysis may provide • ICT Technology should reach to each learner
directions to assist in making decisions and strategies • Generation of quality e-content, questions bank
about the relative merits of different activities in the ICT as modules-based learning.
universities. • Development of interface modules for physically
challenged learners.
In order to undertake a SWOT analysis with rigor, an • Facility of Geographical Information System
essential pre- requisite is that the primary data collected (GIS) for planning upto the village level.
should be through persons who have a deep understanding • Improvement in course curriculum and teachers
of the organization, including its historical perspective. training programs.
This would enable one to identify its strengths, • Efficient and effective knowledge transfer to
weaknesses and opportunities as well as a sound learner with proper interaction
understanding of internal and external environment, which • Voice over Internet Protocol (VOIP) supported
may effect positively as opportunities and negatively as communication between learner and teacher
harmful effects and threat [8].
• Enterprise Resource Planning (ERP) and e-
governance for education, coordination &
synergy for implementation of the policies,
1.2. National Mission on Education through ICT
setting up virtual laboratories and support for
creation of virtual technical universities.
One of the most crucial challenges facing Indian higher • Performance optimization of e-resources
education is its quality for which Government of India has • Certification of attainments of any kind at any
an ambitious goal for eleventh plan. Recently, level.
Government of India through Ministry of Human
Resource Development (MHRD) has developed a holistic All these factors are supposed to contribute towards the
approach on National Mission on Education through ICT SWOT analysis of any higher educational institution.
(NME ICT)[15]. NMEICT has brought out a document, In the present paper SWOT analysis will be carried out
which has already been triggered during the period of through a 2x2 matrix worksheet as given in Table I, of
tenth plan (phase I). As per its strategy, its future vision, ICT in six universities of western Himalayan region of
planning and developmental activities will form phase II India.
and phase III during the eleventh five year plan period.
Table I

It has an ambitious vision of providing one stop solution Strengths Weakness

for the learning community. The working document of the
mission is concerned with the education from school/ Opportunities Threats
college (regular & engineering) level to university level. It
has three guiding principles.
Such a work sheet for SWOT analysis is particularly
Human resource development: Talent in the higher suited in providing a structure, objectivity, transparency, if
education should be identified, trained and utilized in the it is carved out by a small visionary group. Strengths and
service of the country. weaknesses may be visualized from the success and failure
E- content/ resource development. Quality e- content of an organization at the level of implementing policies
should be developed and delivered through the network and its performance thereafter. Further, one has to identify
connectivity of NME ICT. promptly the most attractive opportunities arising from
Building connectivity and knowledge network: In order internal factors to convert the weaknesses/ failures and
to provide maximum benefit to the learners, the maximum then the external environmental factors, to the advantage
possible inter- connectivity should remain available of the organization. In general, internal weaknesses must
among and within institutions of higher learning in the
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 64
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

be tackled first, in the form of opportunities, before

looking at the external environmental factors.

It is always advisable to analyses these four ingredients of

SWOT in a systematic manner. One of the ways, the
authors feel, is to follow a four- tier framework containing
vision and planning, infrastructure, activities, performance
& impact. Higher is the aim & the mission more intense is
the vision. Accordingly the infrastructure is acquired,
which is quite easy, given the financial resources.
However, it is really a passive ingredient without the
skilled human resource. The real strength of the
organization lies in programming and organizing the
activities more and more efficiently using the skilled
human resource which is only possible through the values
of self- discipline and dedication that inculcates a work
culture in them. Without these values one can not dream
of quality in higher education, howsoever good the
planning or the infrastructure may be. These value- based Fig.1. Four Tier Framework for ICT. The number in bracket denotes the
qualities in skilled human resource provide strength to the number of questions related to the item/ group in the questionnaire.
organization in the form of better performance and their
dilution makes the organization weaker and weaker. The
better performance through its product, in turn, leads to The focus of that analysis was on the performance
have an impact at the national and international level. For assessment based on the vision & planning, dynamic
the university system the four- tier framework is given curriculum, good infrastructure with prompt technical
below. support, technically skilled professionals & better
interaction with the students and integration of ICT with
1.3. Four- Tier Framework all its activities particularly the problem solving ability at
the academic level. It was further emphasized that
In a recent paper Sharma and Singh (2009) presented, a universities are meant to create knowledge to operate at
detailed analysis of initiative & planning, status and the universal level where quality remains the defining
performance, obtained from the primary data based on the element. According to that study six universities of this
questionnaire, in the field of ICT in the universities of region are placed in three categories as follows (Table II).
western Himalayan region of India. The analysis was
carried out within a four– tier framework, Fig.1, Table II: Details of six universities in three categories, A, B and C
containing vision & planning, infrastructure, activities and
performance. This study will form the basis for the SWOT
analysis discussed here.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 65

The paper is organized in five sections as follows. Having Both of these technical universities have single faculty of
presented the introduction in section 1, the SWOT analysis engineering with different engineering departments
of technical universities will be presented in section 2. including computer science, electronics and
Section 3, deals with the SWOT analysis of agriculture communication and have been running B.Tech, M.Tech
and horticulture universities. Section III will be devoted to and Ph.D programs.
regular multi-faculty universities and their SWOT
analysis. The conclusion will be given in the final Section. J.P.University in addition has a department of Information
Technology and Bio-informatics. These programs have
been going on in respective departments since 2002,
2. Technical Universities: whereas in NIT Hamirpur, computer centre was
established in 1986, department of electronic and
There are two technical universities out of the six selected communication was started in 1988 and computer science
from the western Himalayan region. One is J.P.University department was setup in 1989. The department of
of Information Technology, Solan and another is National management came in existence in the year 2008.
Institute of Technology, Hamirpur (Deemed University).
In general, at both the universities the courses in computer
2.1. Historical perspective: science are compulsory to learn the ICT skills, for all the
Historical perspective of J.P.University and NIT Hamirpur students belonging to different branches of engineering.
are presented in Table III A and IIIB respectively. But emphasis on the ICT programs is given through the
curriculum in the departments of computer science,
Table III A. Historical Perspective of J.P. University electronics and communication, information technology
and bio-informatics. The department of management in
Faculty Department IT Course Year NIT Hamirpur has also been conducting courses on
Computer B.Tech 2002 information system.
Science & Engg,
Faculty of Information
Engineering Technology, M.Tech, 2006 2.2 SWOT Analysis:
Electronics and Ph.D
Communication, SWOT analysis of these two technical universities
Civil, Bio- (J.P.University and NIT Hamirpur) is given in Table IVA
Informatics and IV B.

Table III B. Historical Perspective of NIT Hamirpur

Table IV A. SWOT Analysis of J.P. University, Solan
Faculty Department IT Courses Year Internal strengths Internal weakness
Vision- Initiative and ICT Vision-
Computer Centre 1986
Planning ICT Governance,
Faculty of Electrical and B.Tech, 1986 ICT Infrastructure Leadership
Engineering Civil. Network internet and ICT Infrastructure-
Electronics & B.Tech, 1988 security, Information Mobile computing.
Communication system, ERP, Teaching / Video-Conferencing
Computer Science B.Tech, 1989 technical staff, ICT budget Location specific, lack of
allocation, E-library qualified faculty
Mechanical B.Tech, 1993 IP Telephony
system/ E- Content, E-
Architecture B.Arch. 2000 placement/alumni portal, Lack of redundancy
ICT based teaching and feature in campus wide
Electrical, Civil, M.Tech 2006 network backbone.
learning, Redundancy
Electronics & and Ph.D
feature in firewall,
Communication, Activities-
Activities-Greater Industry
Computer Weak student teacher
Science, Performance- Research, interaction
Mechanical Placement.
Management and MBA, 2009
Social Science. Ph.D
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 66
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

External Opportunities External Threats b) NIT Hamirpur draws better faculty and technical
Effective interaction with Government policy and staff in sufficient number. They are encouraged
Alumni and industries. norms to improve their technical skills and qualification
Collaboration with Growing digital divide. at the national level through quality improvement
universities and industries. Threat from other private programs organized by AICTE.
Entrepreneurship programs /foreign universities c) NIT Hamirpur has better facility of IP telephony,
Solution to environmental players. wireless network and student counseling.
disasters. Development of d) NIT Hamirpur organizes application and training/
skilled professionals. extension programs using ICT facilities, for the
faculty of different universities and engineering
Table IV B. SWOT Analysis of NIT Hamirpur colleges.
Internal strengths Internal e) Better ICT facilities including the concentration
weaknesses on problem- based learning makes teaching
Vision- programs better, improving the quality of
Effective Leadership & ICT students.
Governance. ICT Infrastructure-
ICT Infrastructure Lack of Both the universities, due to greater industry interaction,
LAN Wired and Wireless, redundancy feature have better opportunities of placement for outgoing
Effect website, Internet firewall in campus wide students at the national and international level.
security, video Conferencing, IP network backbone Nevertheless, J.P. University encourages students to join
Telephony within campus, and firewall. J.P group itself. In overall performance, NIT Hamirpur has
Maintenance of networks, an edge over the other.
Informative website, IS, ERP,
Alumni portal/association, ICT Weaknesses: Both the technical universities lack in
technical staff, Maintenance, ICT campus wide network backbone redundancy features. NIT
budget, E library, E- content, Well Hamirpur also lacks in redundancy of firewall. J.P.
qualified faculty University is supposed to be weak in student- teacher
Activities interaction due to the large number students in the class
Problem based teaching & room as compared to NIT Hamirpur. Whereas
learning, Greater industry J.P.University lacks in mobile computing, IP telephony
interaction. and Video-conferencing.
Research collaboration, Actual Opportunities: Opportunities can be divided into two
placement, Training for faculties. groups; one coming from internal factors and another from
Opportunities Threats external/environmental factors. Most of the internal
More ICT based training Threat from security, weaknesses of an institution can become attractive
programs for professionals. government policy, opportunities manifested mainly in the form of
More effective contact with Private foreign performance.
alumni. universities.
Develop new ICT tools for In this respect, redundancy feature of the campus wide
teaching and e-learning network backbone, close contact with its alumni which
J.P. University may be able to handle in a better way, are
Strengths: As per our general framework, the vision and some of the opportunities for both of them. In particular
ICT planning and various initiatives belonging to tier I, these are the alumni who always prove very helpful to the
are reflected through good ICT infrastructure, networking, organizations and for providing placement to outgoing
internet security, information system/ ERP, e-library students. These technical universities are also having a
system, effective websites and video conferencing facility. greater responsibility towards finding solutions to
Both the universities are having good maintenance environmental disasters and development of skilled
network and of computers, close academic industry professionals along with the development of
interaction, e-placement and alumni association/portal. It entrepreneurs.
is interesting to point out that NIT Hamirpur has an edge
over J.P.University because of: Threats: Being financially sound, there is no internal
threat as such at the level of ICT infrastructure and
a) The availability of best financial resources from activities. At the performance level both the institutions
Government of India (GOI).
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 67

have a threat from the security in social context, which Computer

may be external. Further, there may be a threat from Applications
private/ foreign universities in future. Growing digital Agriculture Agriculture Introduction 1997
divide, again in social context, may be another threat. For Business Business to Computer,
a private university like J.P.University there may be threat Studies Studies MIS and
from the government policies framed from time to time. Computer
in MBA.
3. Agriculture and Horticulture Universities:
The computer centres as central facility, was established in
both the universities in 1988. Since then they offer
In this category there are two universities. One is compulsory courses on professional skills on computers
Agriculture University, Palampur and another is and its applications including programming, in each
Horticulture University, Nauni, and Solan. Initially, branch of studies at the UG and PG level.
Palampur campus for agriculture and Nauni campus for
Horticulture/ Forestry were part of the regular multi- Horticulture University, Nauni was the first to have
faculty, Himachal Pradesh University, Shimla, which network facility and internet connectivity provided by
trifurcated in the year 1977 resulting in a full fledged ICAR in the year 1996. VSAT was commissioned in the
Agriculture University Palampur and Horticulture year 1999 and OFB was laid in the following year.
University, Nauni, Solan Horticulture University, Nauni has two faculties i.e. the
faculty of Horticulture & Forestry and the faculty of
agriculture business studies. In the faculty of agriculture
3.1 Historical Perspective. business studies, the course on computer application was
started in the year 1997 for MBA students. They also
The historical perspective of Agriculture University,
organize program on Geographic Information System
Palampur and Horticulture University, Nauni, Solan are
(GIS) frequently using ICT facilities. Whereas, agriculture
displayed in Table VA and Table VB, respectively
university, Palampur has three faculties namely the faculty
of agriculture, the faculty of veterinary & animal science
Table V A. Historical Perspective Agriculture University,
and the faculty of physical science. Both these universities
organise extension programs with the help of ICT
Faculty Department IT Courses Year
Computer UG and PG 1988 3.2. SWOT Analysis.
Centre Level
(Central SWOT analysis of these two universities is given in VIA
Facility) and VI B.
Agriculture, Agriculture, Fortran 1991
Veterinary & Veterinary & Programming Table VIA. SWOT Analysis of Agriculture University,
Animal Animal Introduction to 2000 Palampur
Science and Science and computers and Internal strengths Internal weaknesses
Physical Physical ICT, Vision-ICT planning Vision- ICT Motivated
Science Science C ICT Infrastructure- leadership
Programming LAN facility ICT Infrastructure-
Internet and security Lack of redundancy feature
Table VB. Historical Perspective of Horticulture Mobile computing in campus backbone.
University, Nauni E library ICT support system.
Activities: IS and ERP
Faculty Department IT Courses Year ICT based teaching ICT Budget
Computer UG and PG 1988 and training, Maintenance of computers
Centre Level Extension programs Activities
(Central Performance: E placement and
Facility) Information from alumni portal
Horticulture Horticulture & Computer 1988 effective Website. ICT technologies in
& Forestry Forestry Science and Research. teaching
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 68
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Opportunities Threats supported by business management skills which are quite

Effective University Threat from government advantageous to the students in being good entrepreneurs.
Industry Interaction policies.
Close Contact with Alumni. Threat from other private Agriculture University, Palampur has a relative advantage
Effective Research players.
Collaboration Bandwidth Faculty Department Courses Year
upgradation Faculty of Computer Centre CIC, DCA 1987
Physical Computer Science MCA, PGDCA 1989
Table VI B. SWOT Analysis of Horticulture University, Science
Internal strengths Internal weaknesses Computer Science M.Tech.(CS) 2006
Vision-ICT Planning Vision- ICT Motivated Faculty of UIIT B. Tech. (IT) 2000
ICT infrastructure Leadership Engg &
LAN facility ICT Infrastructure- Technology
E-library system Internet, security, Faculty of Institute of
Activities- redundancy in OFB Commerce management of MBA 2002
ICT based teaching, learning IP Telephony, Video & Mgt Studies. (Information
Training/Extension Conferencing, Mobile Systems)
programs computing , Information of their better informative, effective and functional
Performance- system and ERP, ICT website with internal mail system. They have also
Actual placement. Teaching technologies, provided the internet connectivity with better backbone on
Budget allocation, ICT the campus.
support system/Placement
Portal, Teaching and Weakness: These two universities are lacking in
Technical Staff information system, ERP, E-placement and alumni portal,
Activities- redundancy feature in campus backbone and firewall.
Research using ICT They also lack in e-library system and e-content
Opportunities Threats management, IP telephony, video- conferencing, mobile
Close Contact with Alumni. Powerful private and computing. Being state universities, both the universities
More Problem oriented foreign Universities. are having financial constraints.
programs.Bandwidth Government policies.
upgradation Network is not secured. Though Horticulture University, Nauni is better in
planning and ICT infrastructure, however, it got leveled
Strength: Both these universities share similar vision and due to weakness in activities and performance as
planning in their respective fields of agriculture, compared to that in agriculture university, Palampur,
horticulture and forestry. The ICT infrastructures like which organises better ICT activities through more
network, internet connectivity, e-library system are also informative web site and intranet services.
similar and both of them organize extension programs in
order to train the professionals. These two universities Opportunities: The above weaknesses may be converted
have exclusive academic faculties in respective into opportunities for development. Both the universities
disciplines; as a result they have independent programs have opportunity in the form of upgrading the bandwidth
and activities. and other desired infrastructure alongwith the more
Horticulture University, Nauni has an additional faculty of activities.
agriculture business studies in which they have special
feature of organizing activities related to agricultural Threats: The network is not secure, in social context, in
business management as per industry requirement. They either of the universities. There may be a threat from
also train professionals in the field of Geographical private and foreign players.
Information System (GIS), an application oriented
program which gives information upto the village level.
Due to this innovative program the students from this
university have better placements. Further, they also
conduct various special courses, such as floriculture,
sericulture, mushroom cultivation. These activities are
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 69

4. Regular Multi- faculty Universities: ejournals, for academic community of the university. This
facility was centrally located on the university campus.
There are two universities in this category. These are In the year 2007, terra-byte optical fibre backbone
Himachal Pradesh University, Shimla and Jammu connectivity was commissioned. All the teaching faculty
University, Jammu. members, various teaching laboratories and administrative
officers got the internet facility right in their offices. Total
4.1 Historical perspective- number of users on campus became 810. This internet
facility provided access to all e-journals through the
Historical perspective of these two universities is given in Inflibnet (INFLIBNET), for teachers, researchers of this
VIIA and VII B respectively: university. In the year 2008, connectivity of 512 Kbps
was upgraded to 2 Mbps (1:1) leased line. The ICT
Table VIIA. Historical Perspective of H. P. University, infrastructure developed on the campus is being used by
Shimla whole of the university.

Table VII B. Historical Perspective of Jammu University, B.Tech. program in information technology was started in
Jammu the year 2000, with the University Institute of Information
Technology. MBA (Information System) was triggered in
the year 2002 under the faculty of commerce and
Faculty Department Courses Year management. M.Tech. in computer science began in the
Computer Centre 1987 year 2006, under the faculty of physical sciences.
Physical (Central facility)
Science Computer Diploma 1987 University of Jammu came into existence in 1969.
Science in Computer centre was established in the year 1987.
Computer University Optical Fibre backbone was established in
Science 2003. In the year 2005, a comprehensive website of this
Physical Computer MCA 1995 university became functional. The DCA program was
Science Science launched in 1987 followed by MCA program in 1995
Physical Computer Ph.D under the department of computer science. The department
Science Science of management studies has been conducting MBA (IT)
Management PG program on the campus. It is pertinent to mention that
Management Studies Diploma whole of the academic community is being benefited by
in Mgt these ICT facilities.
Management MBA 4.2 SWOT Analysis:
Himachal Pradesh University and Jammu University are
state universities of the state of Himachal Pradesh and
Jammu & Kashmir respectively in the western Himalayan
Himachal Pradesh University, Shimla started functioning
region. As a result, they have constraints on the financial
in 1971 and had been running various PG programs in
resources. Both are multi-faculty universities and are
more than a dozen faculties including faculty of
working for diverse disciplines. The SWOT analysis of
agriculture and horticulture/forestry. As a result of
these two universities is given Table VIII A and VIII B.
trifurcation in the year 1977, full fledged agriculture
university at Palampur and Horticulture University at
Table VIII A. SWOT Analysis of H. P. University,
Nauni, Solan, came into existence in the state of Himachal
Pradesh. Presently, Himachal Pradesh University is having
more than 30 teaching departments, on its campus. Internal Internal weaknesses
The Computer centre at Himachal Pradesh University, strengths Vision – ICT oriented leadership
Vision: ICT ICT Infrastructure-
Shimla was established 1987, under faculty of physical
Planning Non availability web
science, with diploma course in computer applications. In
the year 1989, MCA program was started and DCA was
ICT server/mailserver/e-content delivery
Infrastructure system/DNS Services. Redundancy
upgraded to PG diploma in computer applications
LAN Facility feature in firewall.
(PGDCA). In the year 2004, VSAT connectivity (512
Internet and Videoconferencing
kbps) was installed with bouquet of more than 4000
firewall IP Telephony Mobile
Security, computing(Wireless) ERP and E-
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 70
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Redundancy Governance ICT Technologies for Strength: In vision and planning, Himachal Pradesh
feature in teaching University had a little advantage over Jammu University
OFB. Qualified & sufficient Technical in ICT infrastructure, networking and security; both the
Staff universities are at par. In addition to this, Jammu
Activities: Maintenance of ICT infrastructure University is having its own Web server, mail server, DNS
Training/Extensi ICT budget facility, better mobile computing, and more internet
on E-library system/automation/ e- bandwidth as compared to Himachal Pradesh University
programs(ASC) Contents.
Activities- Weakness: Himachal Pradesh University is lacking with
Performance E-placement and alumni association. the facility of web server, mail server, DNS facilities, and
using ICT Problem based learning/teaching video conferencing. Mobile computing facilities, ERP, e-
Research approach. content, e-governance are not available in both the
Collaboration with other universities. universities. They lack sufficient ICT technologies for
Synergy with multi disciplinary teaching, e-library, e-placement, ICT support system,
activities maintenance of computers along with sufficient well
Training of faculty/skilled qualified teaching faculty. Nevertheless, Jammu
professionals, Problem oriented University has certain advantage in respect of video
education. conferencing and better bandwidth connectivity over
Opportunities Threats Himachal Pradesh University.
Collaboration with Migration of students to other
industries. universities, Presence of private Opportunities: The weaknesses in respect of ICT are to
Entrepreneurship foreign universities. be converted into opportunities in infrastructure and
Solution of ICT Threat from government policies. activities. The more crucial is to adopt the technology of
disaster. Close problem solving orientation in learning as internal factors
contact with and to face the challenges due to external factors like
alumni collaboration with other universities, industries and
development of ICT applications at the advanced and
professional level. Establishment of close contact with
Table VIII B: SWOT Analysis of Jammu University, alumni will also be helpful for both the universities.
Internal strengths Internal weaknesses Threats: Threats come from government policies, private
Vision- ICT Planning Vision Motivating & foreign universities. ICT security is the major threat in
ICT Infrastructure Leadership. both the universities.
Impact of ICT, research ICT Infrastructure-
and placement. Redundancy feature in
LAN facility backbone and Firewall, 5. Conclusions and Suggestions
Internet and firewall video conference facility
security IP Telephony, IS/ ERP We have presented a comparative SWOT analysis in
Mobile Computing ICT Technologies in respect of ICT, of six universities placed in three
Activities- Teaching, ICT teaching, categories, supported by their historical perspectives. This
Extension technical staff, E-library has been done within the four- tier framework of ICT[15]
programs(ASC) system, E content, ICT and on the basis of primary data/ feedback obtained from
Performance support System different universities. Findings of this paper are along the
Research Activities – lines to those of National Accreditation and Assessment
Industry university of Committee (NAAC), an autonomous body of University
interaction Grants Commission as far as regular multi faculty
Research Performance. universities are concerned. ICT activities have a crucial
Problem Oriented role to play as per NMEICT directions/ policies to be
Training/Faculty adopted by the universities in order to achieve quality and
Opportunities Threats excellence in higher education system in the region.
Close Contact with Wi-fi is not secured.
Alumni. Threat from other private On the basis of this SWOT analysis, answers to some of
University Industry players. Threat from the glaring questions regarding ICT ingredients may be
Interaction. Collaboration government policies. briefly mentioned as follows:
with Foreign universities
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 71

¾ More and more collaboration with industries

ICT Vision and Planning (TIER I): be encouraged.
¾ A motivating ICT oriented leadership should ¾ Classroom teachings are strengthened
be provided to universities. through ICT.
¾ There should be academic flexibility with a
dynamic curriculum through the feedback There seems to be a downturn in IT field which may
obtained from alumni, teachers and affect all the universities alike and should be taken as
industries. a challenge. This can be overcome by improving the
¾ Interdisciplinary courses based on ICT quality of teaching through problem solving
should be encouraged. orientation/ training programs as per the need of the
country and the global scenario.

All these suggestions may be seen as opportunities for

the development of the system of various higher
educational institutions/ universities.
ICT Infrastructure (Tier II)
¾ Adequate internet bandwidth and number of
e-journals should be made available. The authors thank Dr. S.P. Saraswat, Agro- Economic
¾ The network system should be best secured Centre, Himachal Pradesh University, Shimla for his
against anti- national elements. valuable discussions during the SWOT analysis.
¾ Latest technology should be adopted from
time to time.
¾ ERP systems are implemented so as to have References
paperless communication along with an [1] Balamurali krishna Radha and Dugger, J.C (1991),
interface with the academic community. SWOT Analysis, a Management Tool for Initiating
¾ There should be ICT based connectivity New Programs in Vocational Schools J. Vocational
among various universities and institutions and Technical Education, Vol 12, 1, Iowa State
so as to share knowledgeable e- resources. University.
[2] Bartol, K.M and Martin D.C, (1991), Management,
Activities (Tier III) New York, Mc Grawhill.
¾ ICT has yet to make an impact in classrooms. [3] Bourgeois, L.J.(1996). Strategic Management from
¾ Problem- solving oriented education system Concept to Implementation, Fort Worth, TX: The
is followed in the universities, so as to train Dryden Press.
skilled professionals, providing responsible [4] Gable, G.G (2007) The Information System Academic
leadership through qualified and dedicated Discipline in Pacific Asia- a Contextual Analysis 21
teachers. pp.1-22.
¾ Interaction with alumni (with a strong data [5] Gable G.G, Lee.J.N., Kwahk, K.Y and Green, P.
base) and industries needs to be enhanced. (2007) Administrative Placement of the Information
¾ All university administrative wings be systems academic Discipline: A comparative SWOT
computerized and integrated with ERP. Analysis. Communication of the Association for
¾ Faculty members are motivated to take the Information Systems (21), pp 137-165.
projects/ consultancy. [6] Houben, G., Lenie, K. and Vanhoof, K. (1999), “ A
¾ More effective counseling is made available Knowledge-Based SWOT Analysis System as an
to the academic community. Instrument for Strategic Planning in Small and
Medium Sized Enterprises,” Decision Support
Performance (Tier IV) System, (26), pp 125-135
¾ Internal Quality Audit Cell (IQAC) should [7] INFLIBNET,
be strengthened for quality enhancement [8] Jackson, S.E, Joshi, A. and Erhardt, N.L. (2003),
¾ University should explore placements Recent Research on Team and Organization
through their alumni and corporate sector. Diversity: SWOT analysis and implication” Journal of
¾ Collaboration among different universities/ Management 29, 6 pp 801-830.
institutions in order to share e- resources and [9] Johnson G, Scholes, K and Sexty, R.W (1989),
research projects at the national and Exploring Strategic Management. Prentice Hall.
international level.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 72
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

[10] Mintzberg, H.(1994), The Rise and Fall of Strategic (India), since 2004 onwards. Earlier he was working with
Planning, New York: Prentice Hall. Kurukshetra University, Kurukshetra. His areas of interest
[11]Narayana Murthy N.R. (2009), Tribune India are computer networks, e-Governance, and system
November 16. simulation tools. He is having more than 17 years of
20091117/j&k.htm#1 teaching/ research experience, has more than 30
[12] NAAC (2007), National Assessment & Accreditation publications in international and national
Council, New Methodology of Assessment & journals/conference proceedings, alongwith three books
Accreditation. on the subject.
[13] NME ICT 2009, National Mission on Education
through ICT, and
[14] Pearce, J. A. and Robinson, R.B. (1997), Strategic
Management Formulation, Implementation and
Control, 6th Edition, Chicago: Irwin.
[15] Sharma, D. and Singh, V. (2009), ICT in Universities
of Western Himalayan Region of India: Initiative,
Status and Performance- An Assessment,
International Journal of Computer Science Issues, Vol
6,No 2, pp 44-52
[16]Weihrick, H,(1982), “ The Two- Matrix- A Tool for
Situational Analysis, “ Long Range Planning (15)2,
pp 54-66.

Dhirendra Sharma has obtained his

MBA (1998) from Maastricht School of
Management, Maastricht, Netherlands,
M.S (Software System) (1996) from
BITS, Pilani, India, M.Sc.Physics
(1989) & M.Phil. (Physics) (1990) from
Himachal Pradesh University, Shimla,
India. His areas of interest are ERP and
its implementation in higher educational institutions,
computer networking (wired and wireless), Wireless
Sensor Networks (WSN), and open source web content
management. He played a very important role in Design
and Implementation of Campus Wide Optical Fibre
Network at Himachal Pradesh University, Shimla. He is
having more than 10 years of teaching experience in
addition to his 5 years in IT Industry. He is having more
than 07 publications in international/national
journals/conferences. At present he is pursuing Ph.D. from
the Department of Computer Science and Engineering,
Ch. Devi Lal University, Sirsa, Haryana, India.

Dr. Vikram Singh is Ph.D (2004) in

Computer Science from Kurukshetra
University, Kurukshetra, India.
Presently he is working as Professor
and Head in the Department
of Computer Science & Engg and Dean
Faulty of Engineering, Ch. Devi Lal
University, Sirsa – 125055. Haryana,
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 73
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Stochastic Model Based Proxy Servers Architecture for

VoD to Achieve Reduced Client Waiting Time

Dr T R GopalaKrishnan Nair1 and M Dakshayini2,

Director, Research and Industry Incubation Centre, DSI,
Bangalore, India

Research Scholar, Dr. MGR University. Working with Dept. of ISE, BMSCE, Bangalore.
Member, Multimedia Research Group, Research Centre, DSI,
Bangalore, India

Abstract the partial or the complete videos which has a high

In a video on demand system, the main video repository may be demand locally at the proxy servers solves all these
far away from the user and generally has limited streaming problems. This reduces the main server load by
capacities. Since a high quality video’s size is huge, it requires distributing the load across the network [3].
high bandwidth for streaming over the internet. In order to VoD system usually has several servers and distributed
achieve a higher video hit ratio, reduced client waiting time,
clients over the entire network. These servers contain
distributed server’s architecture can be used, in which multiple
local servers are placed close to clients and, based on their prerecorded videos and are streamed to the clients upon
regional demands video contents are cached dynamically from request from the clients. Proxy cache attempt to improve
the main server. As the cost of proxy server is decreasing and performance of the overall network communication in
demand for reduced waiting time is increasing day by day, newer three ways [9]:
architectures are explored, innovative schemes are arrived at. In i Reduce the request-service delay associated with
this paper we present novel 3 layer architecture, includes main obtaining documents (because the proxy cache is placed
multimedia server, a Tracker and Proxy servers. This architecture typically closer to the user).
targets to optimize the client waiting time. We also propose an ii. Lower the network traffic (the documents served
efficient prefix caching and load sharing algorithm at the proxy
already are available to the user for next time so less load
server to allocate the cache according to regional popularity of
the video. The simulation results demonstrate that it achieves on the network)
significantly lower client's waiting time, when compared to the iii. Reduce the Network cost.
other existing algorithms. In recent years, to reduce the request-service delay and
Keywords: Video Streaming, Proxy prefix caching, video bandwidth demand between the Main multimedia server
distribution, Load sharing, client waiting time. and the proxy servers, a number of caching and buffering
techniques have been proposed. Most of these techniques
use proxy servers with large storage space for caching
1. Introduction videos which are requested frequently. The cached data is
used to serve the future requests and only the un cached
The tremendous growth of World Wide Web has resulted portions of the video are downloaded from the Main
in an increase of bandwidth consumption throughout the servers [2, 12].
internet. Proxy caching has been recognized as an Proxy servers have been widely used for multimedia
effective technique to reduce network traffic. Caching is contents to decrease the startup delay and to reduce the
also an important mechanism for improving both the load of the Main multimedia server. Recent works
performance and operational cost of multimedia networks investigate the advantages of connected proxy servers
[10,13]. Recent web video access patterns show frequent within the same intranet [3, 4 and 8].
requests for a small number of popular objects at popular
sites. So a popular video can be streamed to the same
network link once per request. In the absence of caching, 2. Related work
this approach results in server over load, network
congestion, higher request-service delay, and even the This section briefly discusses the previous work as
higher possibility of rejection of a clients request. Caching follows, Tay and pang have proposed an algorithm in
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 74

Ref.[3] called GWQ (Global waiting queue) which reduces connected to MMS, is in turn connected to its left and right
the initial startup delay by sharing the videos in a neighboring LPSG in a ring fashion through its tracker.
distributed loosely coupled VoD system by balancing the We also propose an efficient regional popularity based
load between the lightly loaded proxy servers and heavily prefix caching and load sharing algorithm (RPPCL). This
loaded proxy servers in a distributed VoD. So whenever algorithm efficiently allocates the cache blocks to the
the local server is busy, the request will be serviced from video according to their local popularity and also shares
the remote server. This introduces the additional network the videos present among the PSs of the LPSG. Hence our
traffic that flows from remote servers. They have approach increases the video hit rate and reduces the client
replicated the videos evenly in all the servers, for which waiting time, network usage on MMS to PS path.
the storage capacity of individual proxy server should be The main aim of arranging the group of proxy servers in
very large to store all the videos. This may not allow each the form of LPSG is to provide the following advantages.
server to store replicas of more number of videos. Our • Reduced Client waiting time: replicating the
proposed scheme replicates only regionally (local and videos at PSs of Lp based on their local
global) popular videos using dynamic buffer allocation popularity, and sharing of these videos among the
algorithm[2] there by utilizing the proxy server storage
PSs of Lp can provide the service to the clients
space more efficiently to store replicas of more number of
videos. In [4] Sonia Gonzalez, Navarro, Zapata proposed a immediately as they request.
more realistic partial replication and load sharing • Increased aggregate storage space: by
algorithm PRLS to distribute the load in a distributed VoD distributing large number of videos across the
system. In their research, they have demonstrated that their PSs and TR of Lp, high cache hit rate can be
algorithm maintains a small initial start up delay using less achieved. For example, if 10 PSs within a LPSG
storage capacity servers by allowing partial replication of
managed 500 Mbytes each, total space available
the videos. They store the locally requested videos in each
server. Our work differs by caching the initial some is 5 GB. 200 proxies of LPSG could store about
portion of the video as prefix-1 at proxy and next part of 100 GB of movies.
the video as prefix-2 at tracker based on local and global • Load reduction: replication of the videos among
popularity using dynamic buffer allocation algorithm [2]. the PSs of Lp based on their regional popularity,
S.-H. Gary Chan, Fouad Tobagi in [7] considers the exchange allows more number of clients to get serviced
of cached contents with the neighboring proxy server from Lp. This reduces the communication with
without any coordinator. Our approach differs, in which
the main multimedia server and in turn its load.
we have made a group of proxy servers with a coordinator
(Tracker) to make the sharing of videos more efficient. • Scalability: by adding more number of PSs the
Another approach to reduce the aggregated transmission capacity of the system can be expanded.
cost has been discussed in [6] by caching the prefix and Interconnected TRs increases the system
prefix of suffix at proxy and client respectively. Since the throughput  
clients are not trustable, and can fail or may leave the The organization of rest of the paper is as follows: In
network at any time without any notice, they have adopted section 3 we present a Model of the problem, Section 4
an additional mechanism to verify the client and cached describes the proposed approach and algorithm in detail,
data at client, which increases the overhead of such In section 5 we present a simulation model, Section 6
verification. Both searching of the video in the whole presents the simulation results and comparison of RPPCL,
cluster of proxy servers, and the verification process GWQ and PRLS algorithms, Finally, in section 7 we
increases the client's waiting time. conclude the paper and refer to further work.
So in order to minimize the client waiting time and
network traffic in the VoD system, in this paper, we
present a novel 3 layer architecture of distributed proxy
servers, for serving videos with a target to optimize the
client waiting time. This architecture consists of a Main
multimedia server [MMS], which is very far away from the 3. Stochastic Model of the Problem
user and is connected to a set of trackers [TR]. Each
tracker is in turn connected to a group of proxy servers Let N be a stochastic variable representing the group of
[PSs] and these proxy servers are assumed to be videos. It may take the different values for (videos) Vi
interconnected in a ring pattern, this arrangement of (i=1,2 . . N) and the probability of the video Vi being
cluster of proxy servers is called as Local Proxy servers asked is p(Vi). Let the set of values p(Vi) be the
Group[LPSG(Lp)]. Each of such LPSG, which is probability mass function. Since the variable must take
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 75
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

∑ Server. After requesting for a video Vi at PSq, the

one of the values, it follows that i =1
p(vi ) = 1 . So the streaming of that video Vi may be delayed by
estimation of the probability of requesting Vi video, is
ni Wt i
= T(pref-1 )i where i=1..N, q=1...M
p(Vi) = .
Where I is the total number of observations and ni is the Where T is the time required to retrieve and initiate the
streaming of (pref-1)i from PS to the requested user(ps-
number of requests for ith video. A cumulative distribution
user). Subsequently by the end of w1 minutes (pref-2)i will
function denoted as P(Vi) is the function that gives the
be streamed from TR to user through PS ((TR-PS)(PS-
probability of a request (random variable’s) being less
user)). By the end of w2 minutes, (S-(pref-1)-(pref-2))Vi
than or equal to a given maximum value.
will be streamed from MMS to user through PS ((MMs-
We assume that client’s requests (X/hr) arrive according to
TR)(TR-PS)(PS-user)) in continuous to (pref-1)i.
Poisson process with λ as shown in Fig.2 of simulation
Another output stochastic variable y is the average
model. Let Si be the size (duration in minutes) of ith video
with mean arrival rates λ1 . . . λN respectively that are waiting time for all the clients. Thus y is a sample mean
being streamed to the users using M proxy servers (PSs) of of client delays Wt1,Wt2…WtN.
J LPSGs (Lp p=1..J). 1

Each TR and PSq(q=1..M), has a caching buffer large That is y = i =1
enough to cache total P and B minutes of H and K number Q
of videos respectively.
∑ ∑
i.e. p = i =1
(pref-2 )i and B= i =1
(pref-1 )i
Every i video Vi is divided into 3 parts, first W1 minutes
of each video Vi is referred to as prefix-1 (pref-1)i of Vi .
If Vi is globally popular then it is replicated at all M PSs
otherwise it is replicated across L PSs of Lp(p=1..J), in
which the frequency of accessing the video Vi is high.
Next W2 minutes of video Vi is referred to as prefix-2
(pref-2)i of Vi is cached at TR of Lp and the rest of the
video is referred to as suffix of the video and is stored at
MMS as shown in fig.1. This arrangement of replicating
the popularity based (pref-1) at L PSs, helps the system to
serve the request immediately as the request arrives. It also Let Q be another stochastic variable represents the number
keeps the queue length QL very small. of requests served immediately from PSq, and Wt( ), is the
non-linear function. The optimization problem is to
maximize the number of clients Q served from PSq
immediately, by replicating the popularity based (pref-1)
videos at L PSs using dynamic buffer allocation algorithm
Depending on the probability of occurrence of user [2]. Also to minimize the average user waiting time y at
requests to any video, the popularity and size of (pref-1) PS by sharing the videos cached among the PSs of Lp.
and (pref-2) of the videos to be cached at PS and TR This can be formulated as follows:
respectively are determined. That is size (W) of (pref-1) 1

and (pref-2) for ith video is determined as. Minimize Wtime is y = i =1
i.e. Q
Subject to
W (pref − 1 )i = xi×Si where 0<xi<1 ∑ ∑
So B = i =1
(pref − 1 )i , P =
i =1
(pref − 2 )i
W (pref − 2 )i = xi×(Si- (pref-1)i )
where 0<xi<1 ( pref − 1) > 0 and ( pref − 2) > 0
Where xi is the probability of arrival of requests for the ith
video from last t minutes, and ni is the total number of
requests for video Vi. Let bi be the available bandwidth
for Vi between the proxy server and Main multimedia
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 76

4. Proposed Architecture and Algorithm

4.1. Overview of the proposed Architecture

The proposed 3 layer architecture is as shown in Fig.4.
This architecture consists of a MMS, which is connected to
a group of trackers (TRs), Each TR has various modules.
As shown in the fig. 3, they are
1. Interaction Module (IMTR) – Interacts with the PS and
2. Service Manager (SMTR) – Handles the requests from
the PS.
3. Database – Stores the complete details of presence and
size of (pref-1) of videos at all the PSs.
4. Video distributing Manager(VDM) – Responsible for
deciding the videos, and sizes of (pref-1), (pref-2) of
videos to be cached. Also handles the distribution and
management of these videos to group of PSs, based on
video’s global and local popularity.
Each TR is in turn connected to a set of PSs. These PSs are
connected among themselves in a ring fashion. Each PS
also has various modules such as,
1. Interaction Module (IMPS) – Interacts with the user and
2. Service Manager (SMPS) – Handles the requests from 3. The MMS, the TR and the PSs of LPSG are assumed to
the user, be interconnected through high capacity optic fiber cables.
3. Popularity agent (PA) – Observes and updates the In the beginning, all the Nv videos are stored in the MMS.
popularity of videos at PS as well as at TR, The distribution of the selected N of Nv videos among M
4. Cache Allocator (CA) – Allocates the Cache blocks PSs of the LPSG is done by VDM as follows. First, all the
using dynamic buffer allocation algorithm [2]. Also to N videos are arranged with respect to their popularity at jth
each of these proxy servers a large number of users are LPSG . The popularity of a video is defined as the
connected [LPSG]. Each proxy server is called as a parent probability of frequency of requests to this video per
proxy server to its clients. All these LPSGs are threshold time t. Here, we assume that the frequency of
interconnected through their TR in a ring pattern as shown requests to a video follows Zipf law of distribution. The
in fig. 4. video distribution module divides N videos into two
The PS caches the (pref-1) of videos distributed by VDM, subgroups- the globally popular k(0 <= k <= N) videos
and streams this cached portion of the videos to the clients like Cartoons, and locally popular N – k videos –such that
upon the request through LAN using its less expensive former small subgroup is replicated in all the PSs and the
bandwidth. later subgroup is cached at PS of Lp based on the local
We assume that, demand for the videos 4.2. Proposed Algorithm
1. The TR is also a PS with high computational power and Since the storage cache space of both PS (CPSq )and TR
large storage compared to other proxy servers, to which (CTR ) is limited, the VDM of the TR first executes the
clients are connected. It has various modules, using which decision making algorithm to fix up the sizes(segments)
it coordinates and maintains a database that contains the of (pref-1) and (pref-2) of videos to be cached at CPSq and
information of the presence of videos, and also size of in its cache CTR respectively. Then caching is done using
(pref-1) and (pref-2) of video in each PS and TR dynamic buffer allocation algorithm [2]. The
respectively corresponding entry is updated in its database at TR.
2. Proxies and their clients are closely located with Whenever a client at PSq wishes to play a video Vi, it first
relatively low communication cost[1]. The Main server in sends a request to its parent proxy PSq, the SMPSq
which all the videos completely stored is placed far away immediately starts streaming the (pref-1) of video
from LPSG, which involves high cost remote requested to the client, if it is present in its cache. So
communication. waiting time is almost negligible. And informs the SMTR
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. x, January 2010 77
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Proposed algorithm
When there is a request for a video vi (at a particular proxy PSq of Lp , do the following:
If (Vreq € PSq)
(pref-1)Vreq is streamed immediately to the user ( y = time required to stream (pref-1) from proxy - user)
So wt Vreq = wt(p-1 )Vreq
else - pass the request to the TR(Lp)
if (Vreq € PS(Lp))
If (PS(Lp) is left or right NBR(PSq)
SMTR streams (pref-1)Vreq from NBR(psq), (pref-2) Vreq from its cache and the remaining portion from
wt Vreq = wt(p-1 )Vreq ( y = time required to stream (pref-1) from proxy- proxy & proxy - user)
SMTR streams the (pref-1)Vreq from OTR(PSq), (pref-2) Vreq from its cache and the remaining portion from
MMS to-User thru PSq using optimal path found
wt Vreq = wt(p-1 )Vreq ( y = time required to stream (pref-1) from proxy- proxy & proxy - user)
Pass the request to left or right TR(NBR(Lp))
if (Vreq € NBR(Lp))
TR(NBR(Lp)) streams the Vreq from NBR(Lp)-user thru TR(Lp)
(t −t)+(t − p)+(p −u)
wt Vreq = wt[(p-1 ) + (p-2 )]Vreq ( y = time required to stream (pref-1) from tracker –
Tracker , tracker – proxy & proxy - user)
TR(Lp) downloads the complete Vreq from MMS and streams to the user
wt Vreq = wt(S)Vreq ( y = time required to stream (pref-1) from MMS -TR, TR-PS & PS-user)
Also caches the (pref-1) and (pref-2) of Vreq at PSq using Dynamic Buffer allocation algorithm[ 2].
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. x, January 2010 78
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

corresponding entry is updated in its database at TR. Whenever the sufficient buffer and bandwidth is not
Whenever a client at PSq wishes to play a video Vi, it first available in the above operation the user request is
sends a request to its parent proxy PSq, the SMPSq rejected.
immediately starts streaming the (pref-1) of video
requested to the client, if it is present in its cache. So
waiting time is almost negligeble. And informs the SMTR
to initiate the streaming of (pref-2) of Vi, then the IMTR 5. Simulation Model
coordinates with MMS to download the remaining portion
(S-(pref-1)-(pref-2))Vi of the video Vi. In our simulation model we have a single MMS and a
If it is not present in its cache, the IMPSq forwards the group of 6 TRs. All these TRs are interconnected among
request to its parent TR, VDM at TR searches its database themselves in a ring fashion. Each of these TR is in turn
using perfect hashing to see whether it is present in any of connected to a set of 6 PSs. These PSs are again
the PSs in that Lp. If the Vi is present in any of the PSs in interconnected among themselves in a ring fashion. To
that Lp, then the VDM checks whether the PS in which the each of this PS, 25 clients are connected. We use the video
Vi found is neighbor to the requested PSq [NBR(PSq)]. hit ratio (VHR), the average client waiting time y
If so, the VDM intimates the same to SMTR which initiates
the streaming of the (pref-1)Vi from that NBR(PSq), and Table 1: Simulation Values
(pref-2)Vi from its cache, to the requested PSq and the same
is intimated to the requested PSq. Then the IMTR Notation System Parameters US Letter Paper
coordinate with MMS to download the remaining portion S Video Size 25 to 1120 min
(S-(pref-1)-(pref-2))Vi, and hence the client waiting time is CMMS Cache Size (MMS) 2000blocks
very small .
CTR Cache Size(TR) 800(40%)
Otherwise, if it is not [NBR(PSq)] and is present in more
than one PS of Lp then SMTR selects one PS such that, the CPS Cache Size(PS) 300(15%)
path from selected PS to PSq should be optimum and λ Mean request arrival rate 45 reqs/hr
initiates the streaming of the (pref-1)Vi from the selected
and network usage as parameters to measure the
PS, and (pref-2)Vi from its cache, to the requested PSq
performance of our proposed approach more correctly by
through the optimal path found by the SMTR and the same
comparing the results of RPPCL, GWQ and PRLS
is intimated to the requested PSq and hence the client
algorithms. In addition we also use the WAN bandwidth
waiting time is relatively higher, but acceptable with high
usage on MMS-PS path and probability of accessing the
main server as the performance metrics.
If the Vi is not present in any of the PSs in that Lp, then
We assume that the request distribution of the videos
the IMTR Passes the request to the tracker of NBR(Lp).
follows a zipf-like distribution. The user request rate at
Then the VDM(NBR(Lp)) checks its database using perfect
each PS is 35-50 requests per hour. The ratio of cache
hashing, to see whether the Vi is present in any of the PSs
sizes at different elements like MMS, TR and PS is set to
of its Lp. If it is present in one or more PSs, then the
CMMS : CTR : CPS = 10:4:2 and transmission delay between
SM(NBR(Lp)) selects the optimal streaming path from the
the proxy and the client, proxy to proxy and TR to PS as
selected PS(NBR(Lp)) to the requested PSq and intimates
120sec, transmission delay between the main server and
the same to IM(Lp). Then the SM(Lp) in turn initiates the
the proxy as 480 to 600sec, transmission delay between
streaming of Vi to the requested PSq through the optimal
tracker to tracker 240sec, the size of the cached [(pref-
path, and the same is intimated to the requested PSq and
1)+(pref-2)] video as 280MB to 1120MB(25-min-1hr) in
hence client waiting time is comparatively high but
proportion to its popularity.
acceptable because it bypasses the downloading of the
complete video from MMS using MMS-PS WAN
bandwidth. 6. Simulation Results
If the Vi is not present in any of the PSs of its NBR(Lp)
also, then the TR(Lp) modules decides to download the The simulation results presented below are an average of
Vi from MMS to PSq. So the IMTR coordinates with MMS several simulations conducted on the model
to download the Vi, and hence the waiting time is very Our main focus was to minimize the client waiting time
high, but the probability of downloading the complete via exploiting load sharing among the PSs of Lp. Fig.9
video from MMS is very less as shown by our simulation shows the total number of requests served from the
results. system, the average number of requests served
immediately at PSq as 51%, the average number of
requests served from (Lp+NBR[Lp]) as 34%, and the
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 79
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

average number of requests served from MMS, that is only MMS has been contacted for very few (15-25% of the
15% which is very less. The corresponding average videos) number of videos, when the Vi is neither present
waiting time required for serving (pref-1) immediately in that Lp, nor in NBR(Lp). Even though the initial startup
from PS, from other PS of Lp (Lp+NBR[Lp]) and from delay and transmission cost seems to be more it is
MMS is shown in the fig. 5. acceptable because on an average (pref-1) and (pref-2) of
As the (pref-1) of most frequently asked videos have been
cached and streamed from the PSq of Lp and NBR[Lp],
with the cooperation of various modules of PSs, and the
coordination of modules of TR of Lp, Our scheme has
achieved a very high video hit ratio ( 86%) as shown in
Fig 6. So the local and global popularity based replication
of mostly accessed videos at the respective

nearly 85% of the videos are cached and streamed from Lp

and NBR(Lp) by assuring high QoS as shown in Fig.9

PSs of LPSG has significantly reduced the waiting time

for the user when compared to GWQ and PRLS as shown
in Fig.6 and Fig.7.
Thus more (80% - 86% of the video) number of blocks of
requested videos are cached and streamed from Lp, by
sharing the videos among the proxies and TR of Lp. So
when there is a request for any of these ith video, streaming
starts from one of the PS immediately and hence client
waiting time, network usage from MMS to proxy is very and only about 15% of the videos are downloaded from
less as shown in fig. 6 and 8, and in turn transmission cost, MMS which has drastically reduced the client waiting
transmission time is also reduced. GWQ also reduces the time. Hence our proposed approach has successfully
waiting time by balancing the load between heavily loaded achieved the load balance among the interconnected PSs
and lightly loaded proxy servers. But it still introduces the of Lp and [NBR[Lp]].
unnecessary network traffic flows from remote servers.
If the requested videos are present at NBR(PSq) of Lp, then
these videos are streamed from NBR(PSq) to the client 7. Conclusions
through PSq, so the waiting time for these videos is very
small. If the requested videos are present in Lp-NBR(PSq), In this paper we have proposed an efficient regional (local
then these videos are streamed from Lp- NBR(PSq) to the and global) popularity based replication, prefix caching
client through PSq, so the waiting time for these videos is and load sharing (RPPCL) algorithm with the architecture.
relatively higher, Otherwise also, some good number of In which all PSs cooperate with each other to achieve
videos are served from NBR(Lp), which reduces frequent reduced client waiting time and increased video hit ratio,
downloading of requested videos from MMS to the PSq by caching (replicating) and streaming maximum portion
which in turn reduces the initial play out delay for the ((pref-1)+(pref+2)) of most frequently requested videos
clients for the requested videos which are not which are among the proxies of Lp. Our simulation results
not present at PSq as shown in Fig.5 and 9. demonstrated that our proposed approach has reduced the
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 80

client waiting time for the videos requested at PSq, [4] S. González, A. Navarro, J. López and E.L. Zapata, “Load
average network traffic of the system, and also the load of Sharing in Distributed VoD Systems”, Int'l Conf. on
MMS by the regional popularity based replication of most Advances in Infrastructure for e- Education, e-Science, and
popular videos at appropriate PSs of Lp. And sharing of e-Medicine on the Internet (SSGRR 2002w), L'Aquila, Italy,
January 21-27, 2002.
these videos among the proxies of the system with the
[5] Yuewei Wang, Zhi-Li Zhang, David H.C. Du, and Dongli Su
“A Network-Conscious Approach to End-to-End Video
Delivery over Wide Area Networks Using Proxy Servers”,
IEEE INFOCOM, pp 660-667, April 1998.
[6] Alan T.S lp, Jiangchuan Liu, John C.S.Lui COPACC: An
Architectureof cooperative proxy-client caching System for
On-Demand Media Streaming. A technical report.
[7] S.-H. Gary Chan, Fouad Tobagi, “distributed Servers
Architecture for Networked Video Services”, IEEE.
Transactions on networking, vol.9, No. 2, Aprol.
[8] S. Acharya and B. C. Smith, “Middleman: A video caching
proxy server”. in Proc. of NOSSDAV, June 2000.
[9] A. Feldmann, R. Caceres, “Performance of Web Proxy
Caching in Heterogeneous Bandwidth Environments”, In
Proc. Of IEEE INFOCOM ’99, March 1999.
[10]P. A. Chou and Z. Miao. Rate-distortion optimized streaming
of packetized media. Technical Report MSR-TR-2001-35,
Microsoft Research Center, February 2001.
[11]Dr.Mahmood Ashraf Khan, Prf.Go-Hasegawa, Yoshiaki
Taniguchi “QoS Multimedia Network Architecture” murata
Laboratory, Osaka University, Japan.
[12] Lian Shen, Wei Tu, and Eckehard Steinbach “A Flexible
Starting Point based Partial Caching Algorithm For Video
On Demand”, 1-4244-1017-7/07@2007 IEEE.
[12] B. Wang, S. Sen, M. adler, and D. Towsley, “ Optmal Proxy
Cache Allocation For Efficient Streaming Media
Distribution,” In IEEE INFOCOM, june 2002.IEEE.

Dr.T R Gopalakrishnan Nair holds M.Tech. (IISc, Bangalore) and

Ph.D. degree in Computer Science. He has 3 decades experience
in Computer Science and Engineering through research, industry
and education. He has published several papers and holds patents
coordination of Tracker also reduces the server-to-client in multi domains. He won the PARAM Award for technology
innovation. Currently he is the Director of Research and Industry in
network usage, transmission cost and time, maintains high Dayananda Sagar Institutions, Bangalore, India.
QoS for the users when compare to GWQ and PRLS M Dakshayini. holds M.Tech (VTU Belgaum) in computer science
algorithms. The future work is being carried out to securing second rank . She has one and a half decades
improve the performance of the system by writing an experience in teaching field. She has published many papers.
Currently she is working as a teaching faculty in the department of
algorithm to handling the failure of the coordinator Information science and engineering at BMS College Of
tracker. Engineering, Bangalore ,India.

[1] Bing Wang, Subhabrata Sen, Micah Adler and don Towsley
“Optmal Proxy cache Alloction for Efficient Streaming
Media Distribution” IEEE Transaction on multimedia, vol. 6,
No. 2, April 2004.
[2] H.S.Guruprasad,M Dakshayini et. el “Dynamic Buffer
Allocation for VoD System Based on Popularity”
proceedings of NCIICT 2006, PSG College of Technology,
Coibatore, 13- 17, WWW. psgtech. edu/ NCIICT/ files/
[3] Y.C Tay and HweeHwa Pang, “Load Sharing in Distributed
Multimedia-On-Demand Systems”, IEEE Transactions on
Knowledge and data Engineering, Vol.12, No.3, May/June
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 81
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Modified EESM Based Link Adaptation Algorithm for

Multimedia Transmission in Multicarrier Systems
Department of Electronics and Communication Engineering
Pondicherry Engineering College
Pondicherry, India

Department of Electrical & Electronics Engineering
Pondicherry Engineering College
Pondicherry, India

Abstract and a mean SNR metric is only sufficient to obtain a

The previous link adaptation algorithms on ofdm based
long term performance metric of the channel. On the other
systems use equal modulation order for all sub carrier index hand, short term performance metrics, which are also key
within a block. For multimedia transmission using ofdm as the to obtaining performance enhancements with feedback,
modulation technique, unequal constellation is used within one are obtained from the actual instantaneous channel
ofdm subcarrier block, a set of subcarriers for audio and realization.
another set for video transmissions. A generic model has been
shown for such a transmission and link adaptation algorithm A well-known approach to link performance
has been proposed using EESM (Effective Exponential SNR modeling and link quality prediction is the Effective
mapping) method as basic method. Mathematical model has Exponential SINR Metric (EESM) method, which
been derived for the channel based on bivariate Gaussian computes an effective SNR (also referred to as
distribution in which the amplitude varies two dimensionally AWGN equivalent SNR) metric by taking as input
in the same envelope. From the Moment generating function the individual sub carrier SNRs and using an
of bivariate distribution, Probability of error has been exponential combining function. Once computed, the
theoretically derived. Results have been shown for BER block error rate is obtained from looking up an
performance of an ofdm system using unequal constellation.
AWGN performance curve. This approach has been
BER performances have been shown for different values of
correlation parameter and fading figure. widely applied to OFDM link layers and is based on the
performance approximation by asymptotic union bounds
Keywords: 802.16, OFDM, link adaptation EESM [2]. It has been shown in literature survey [5, 6] that the
Doppler spreading of OFDM systems follows joint
probability distribution function implies the transmission
1. Introduction will also be dependent on Bivariate Gaussian
Next generation cellular systems support multiple distribution and gives better PDF, CDF. Under the same
transmission m o d e s , which can be used to improve assumption the amplitude variations of the channel
the performance of such systems by adapting to current under two different constellations has been modeled as
bivariate Gaussian distribution. The paper has been
channel conditions. This p r o c e s s i s r e f e r r e d to as
organized as follows; Chapter 1 explains the link
link adaptation. Typically, these transmission modes
adaptation based on unequal constellation. Chapter II
include different modulation and coding, schemes (MCS)
derives the mathematical model based pn bivariate
and different multiple antenna arrangements modes –
Gaussian distribution, and Chapter III produces the
such as beam-forming, space- time coding and spatial
simulated results.
multiplexing – as the transmission becomes
multidimensional in space, time, and frequency domain.
Orthogonal Frequency Division Multiplexing (OFDM) is
the air interface for 802.11, 802.16 (WiMAX), and
3GPP Long Term Evaluation (LTE) systems. The
resources typically referred to as subcarriers, available in
an OFDM frame, can be defined on a time frequency
grid [1]. The performance of a binary code depends on
the channel condition obtained over the allocated
subcarriers. Typically, the channel is frequency selective,
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 82

2. EESM for Unequal Constellation in OFDM the calibration parameter that is typically different for each
Block MCS. The procedure used to link adaptation with EESM is
listed herewith

Determine quadratic approximation of SINR Vs β.

1. Obtain the parameters a,b,c from quadratic equation
SNR eff (β) = a + b βdb + c β2db
Where a,b,c are y-intercept ,linear and quadratic
2. Sent a,b,c as Channel quality indicator to BTS.
3. BTS receiving the parameters will update it and select
the appropriate β for the required SNReff .
4. Based on current β value the modulation order is

unequal modulation eff.snr vs beta


Fig1.Unequal ofdm transceiver structure

16 y = - 0.042*x 2 + 1.3*x + 7.8

The proposed structure [15] transceiver structure consists

of data compressor, unequal adaptive modulator and 14
channel estimator. Multimedia data is compressed using a
source coder based on discrete wavelet transformation, in
eff.S NR in dB

wavelet analysis we represent low frequency information
by approximation coefficients cAn an high pass spatial 10
frequency data by horizontal vertical an diagonal signal by
cHn,cVn,cDn. To achieve a target SNR we allocate 8
approximate coefficients with low modulation order and 16QAM
detail coefficients with high modulation order based on the quadratic
channel estimation feedback message from the receiver.
EESM method has been identified as one of the fast link
adaptation technique for multicarrier based systems. For a 4
0 5 10 15
fast link adaptation of multimedia transmission where the beta
audio and video are transmitted in different constellations Fig. 2. Instantaneous SNReff vs. β1 and SNReff vs. β2 curve with the
on the same ofdm block, the EESM method has been corresponding quadratic approximations.
modified and used for performance prediction (e.g., the
FER metric) for the current channel conditions. In the
subcarrier block with two modulation order are taken such
a way that low frequency components (Audio) are
transmitted with low modulation order and high frequency 3. Vertical Shift Method
components (video) with high modulation order. This Enabling the method described above it would require the
method maps a set of per subcarrier SNRs { γ 1 …γN2} to approximation of the SNReff vs. β1 and curve to SNReff vs.
a single effective SNR (SNReff) : β2 have to be sent potentially as often as every frame, in
order to track changes in the SNR due to fading. This
represents less feedback than sending the entire channel
⎡ N1 − γ i N 2 − λi ⎤ response {γ 1, …, γN}, but is still a significant amount of
⎢∑e + ∑e β2
⎥ …. (1)
γ eff = − β . ln ⎢ i =1 i = N 1+1 ⎥ (1)
⎢ N1 + N 2 ⎥
⎢ ⎥
⎢⎣ ⎥⎦

Where N1&N2 is the number of sub-carriers used by a

codeword , for 1 to N1 subcarriers β1 is the calibration
parameter and from N1+1 to N2 subcarriers β2 is used as
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 83

More specifically, this method consists of sending

infrequently the instantaneous SNReff vs. β1 curve only as a
reference SNR value (reference curve) instead of
transmitting two curves. The SNReff vs. β2 curve for the
Compute current channel is then obtained by a simple vertical (1-
Snr Eff Vs β dimensional) shift of the reference curve by 3dB. This first
order approximation will be referred to as the vertical shift
A,B,C Update
3. Unequal Modulations – Bivariate
The two dimensional amplitude variations within same
channel envelope can be modeled as bivariate Gaussian
Check If distribution. The purpose of such a multilevel modulation
B&C Is is that the carriers in an ofdm block which can withstand
NO deep fading can be allotted a low modulation order and the
others high level modulation and as the result the SNR
performance can be improved much better than equal
modulation. From theory two univariate marginal
Select distributions following Gaussian distribution can be
order for the modeled as under the same roof as bivariate joint
new distribution given [3] by
( )
( xx
1 2)
α−1 /2
⎛ x /β +x /β ⎞ ⎛2 ρ xx ⎞
F= xexp⎜− 1 1 2 2 ⎟xIα−1 ⎜ 1 2
⎟⎟U( x1)U( x2 )
Γ( α)(ββ
1 2)
( )
α+1 /2
(1−ρ) ρ( α−1) /2
⎝ 1−ρ ⎠ ⎜⎝1−ρ ββ
1 2⎠
Vertically shift (2)
the curve and
obtain the Where г(.)is gamma function and I(.) is a modified
Bessel function of first kind ,β1 and β1 are scaling
parameters,x1 and x2 are random variables for different
marginal distributions, and ρ is the correlation coefficient
,When ρ → 0 the joint pdf tends to product of two
univariate gamma distribution . Consider two OFDM
Assign the modulation techniques 16QAM and 64QAM and the joint
modulation for the probability distribution for them is given below,
subcarrier block
corresponding to
low modulation


Figure 3. Modified EESM link adaptation technique for unequal

Fig. (4) PDF for Bivariate normal distribution for 16QAM and
information. For the quadratic approximation, three 64QAM
coefficients (a, b, c) for two different modulation orders
Since the channel is characterized by two different
would have to be fed back every frame. A first possible
constellations, the channel has two different
approach to eliminate the frame-by-frame feedback of curve
parameterized fading envelope. Hence, the envelope of
parameters is to assume that the shape of the β curve does
the channel response can be modeled as nakagami-m
not change significantly with changes in band-average SNR
model which is characterized by two parameters, fading
as long as the channel power-delay profile does not change.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 84

figure and moments. It has the inherent advantage of

versatility and covers a wide range of fading channel m m −1/ 2
scenarios. The m-distributed pdf of the envelope is taken π ⎡ m2 ⎤ ⎛ γ1 ⎞
pa ( γ1 ) = ⎢ ⎥ ⎜ ⎟ I m −1 / 2 ( β' γ1 ) e − a ' γ1
as [2]. Γ ( m ) ⎣⎢ γ1 γ2 (1 − ρ ) ⎦⎥ ⎝ 2β' ⎠
2m m R 2 m −1 mR 2
P( R) = exp( − ), R ≥ 0 γ1 ≥ 0 (6)
Γ ( m)Ω m Ω
whereΩ = E ( R 2 ), m = , m ≥ 1/ 2 Where the parameters α′ and β′ are normalized versions
E[( R 2 − Ω) 2 ]
of the parameters α and β, and are given by

α m(γ 1 + γ 2 )
α' = =
E( ) represents the average and Γ ( ) represents the Es / N 0 2γ 1γ 2 (1 − ρ )
gamma function, Ω /2 is the average power of the
signal, ‘m’ is named fading figure, parameter related to
fading range. In the case of M-distributed correlated
nonidentical fading Nakagami-m channels PDF of the
β ='
β (
m (γ 1 + γ 2 ) − 4γ 1γ 2 (1 − ρ )
) 1/ 2

combined signal envelope[14]- pa(rt), is given by Es / N 0 2γ 1γ 2 (1 − ρ )

m −1/ 2
2rt π ⎡ rt 2 ⎤
pa ( rt ) = m ⎢ ⎥
Γ ( m ) ⎡⎣σ1σ 2 (1 − ρ ) ⎤⎦ ⎣ 2β ⎦ (7)
and finally the equation reduces to
(βrt 2 ) e−αrt , rt ≥ 0
I (4)
1 ⎧⎪ ⎡ λ1 ⎤ ⎡ γ1 ⎤⎫⎪
m −1/ 2

p a (γ 1 ) = ⎨exp⎢− ⎥ − exp⎢−
where Iυ(·) denotes the υth-order modified Bessel
function (
2 ργ ⎪⎩ ⎢⎣ 1+ ρ γ ⎥⎦ ) ⎥⎬ ,
⎢⎣ 1− ρ γ ⎥⎦⎪⎭ ( )
cov r1 , r2
2 2
) ,0 ≤ ρ < 1 (5)
γ1 ≥ 0 (8)

( ) ( )
var r1 var r2

Using the Laplace transform it can be shown after some

is the envelope correlation coefficient [3] between the
two signals[6] and the parameters σd (d = 1, 2), α, and β manipulations that the MGF p a (γ 1 ) of is given by [5]
are defined as follows:
σd = , d = 1, 2 ⎡ (γ + γ ) (1 − ρ ) γ1 γ2 2 ⎤

m M a ( s : γ1 . γ2 : m : ρ ) ≅ M a ( s ) = ⎢1 − 1 2 s + s ⎥
⎣ m m2 ⎦
σ1 + σ 2 s≥ 0 .
α= (9)
2σ 1 σ 2 (1 − ρ )

( σ1 − σ2 )
+ 4σ1σ 2 ρ
β2 =
4σ12 σ 2 2 (1 − ρ )


where Ωd, d = 1, 2, is the average fading power of the

uneven sub-carrier blocks .The PDF of the combined
SNR per symbol, p a (γ 1 ) is given by
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 85


Order of the system (M) 4,16 and 64
The average error probability [14] in general is given by
Types of Modulation QPSK, QAM
Pe = 1 − ∫
−π / m
p(θ)d θ
Data rate (Max.) 50 Mbps

whereP(θ) = ∫ f (θ / R) po ( R) dR. Coding rates 1/2, 2/3,3/4
Subcarriers for audio transmission 52
⎡ ⎤ Subcarriers for video transmission 256
Γ(2m + 1/ 2) 1 ⎢ 1 ⎥
Pe = . .⎢ ⎥
Number of pilot carriers 4,8
ε m π Γ(2m + 1) (1 − k ) ⎢ 2 ⎛ π ⎞⎥
2 m

⎢ ( γ / m) sin ⎜⎝ m ⎟⎠ ⎥
⎣ ⎦ Guard interval 800 ns
OFDM symbol duration 4 µs
f (θ / R ) is the p.d.f of the detection error, po(R) is the pdf
of the envelope of the fading signal. After approximation Channel bandwidth 20MHz
[14] Table.1 Simulation parameters for unequal modulation
⎡ ⎤
Γ(2m + 1/ 2) 1 ⎢ 1 ⎥ For the above mentioned parameters in a rayleigh channel
Pe = . .⎢ ⎥ (11) the simulation has been conducted for unequal modulation
ε m π Γ(2m + 1) (1 − k ) ⎢ 2 ⎛ π ⎞⎥
2 m

⎢ ( γ / m) sin ⎜⎝ m ⎟⎠ ⎥ orders. The resultant shows the unequal ber performance is

⎣ ⎦ better compare to high modulation order to all subcarriers.

K is the power correlation coefficient of the two set of

fading from the above formulas the moment generating
function is dependent upon the correlation coefficient

4. Simulation Results
4QAM,unequal modulation ,16QAM



4 qam 4qam an 16 qam

Fig. 6 Performance of SNR vs BER for different values of ρ
0 5 10 15
SNR in dB

Fig. 5 Ber results for equal and unequal modulation

The above graph is plotted for different values of ρ for 16

qam modulation. The plot shows that as the ρ value increases
the ber increases as the fading profile increases with ρ
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 86
5. Conclusion methodology based on convex metric. Lucent Technologies
Jan 2007.
[8] IEEE 802.16m-07/031, IEEE 802.16 Broadband Wireless
The paper proposes an EESM based link adaptation algorithm Access Working Group: RBIR MLD PHY Abstraction for
in which unequal constellation orders has been used for the HARQ IR/CC,03/10/2007.
same sub-carrier block of an ofdm system. Simulation results [9] Simon MK, Alouin MS. Digital communication fading
have been shown for ber and fading margin with unequal channels –A unified approach to performance analysis. John
constellation size. The result shows performance enhancement Wiley and Sons;2000.
over equal constellation orders in ofdm system. [10] Sayana K, Zhuang J. Link performance abstraction based
on mean mutual information per bit (MMIB) of the
LLR channel. IEEE Contribution, C802.16m-07/097; May
References [11] Ericsson, System-level evaluation of OFDM –
[1] Sayana K, Zhuang J, Stewart K, Motorola Inc. Short further considerations, 3GPP TSG-RAN WG1, No. 35 R1-
term link performance modeling for ml receivers with 031303; 2003. [12] Baum, Kevin, L., Blankenship,
mutual information per bit metrics June 2002. Yufei, W., Classon, Brian, K.,Sartori. : US251180 (2006).
[2] 3GPP2, TSG-C WG3.3GPP TS 25.321 v 5.5.0, Medium [13] EEROLAINEN, LAURI.: US455436 (2005).
Access Control (MAC) protocol specification. [14] Yoshiya miyagaki, Norihiko, Toshihiko namekawa “Error
[3] IEEE 802.16 Broadband Wireless Access Working Group, probability characteristics for cpsk signal Through m-
Sayana K, Zhuang J, Stewart K, Motorola Inc. Link distributed fading channel” IEEE Transactions on
performance abstraction based on mean mutual information communication,vol.com26,N0.1,January 1978
per bit(mmib) of the llr channel Nov 2006. [15] Tasso Athanasiais,Kevin H.Lin, and Zahir M.Hussain, “An
[4] Brueninghaus K, Astely D, Salzer T, Visuri S. Link Unequal modulation scheme for the transmission of
performance models for system level simulations of compresse multimedia data over aaptive mimo ofdm
broadband radio access systems, IEEE 16th International systems”,TENCON 2006.2006 IEEE Region 10th conference
Symposium on Personal, Indoor and Mobile Radio 14-17 Nov 2006 pages:1-4
Communications (PIMRC) Sept 2005.
[5] Wan L, Tsai S, Almgren M. A fading- insensitive R. Sandanalakshmi has been teaching for 9 years in the field of
performance metric for a united link quality model. electronics and communication Engineering. She is currently
Proceedings of WCNC 2006;2110-2114. Asst.Professor, Department of Electronics and Communication
[6] Wang TR, Proakis JG, Masry E, Zeidler JR. engineering, Pondicherry engineering college. She finished her
Performance Degradation of OFDM Systems Due to masters in the area communication engineering at pondicherry
engineering college in the year 2000. She is currently pursuing her
Doppler Spreading Sept 2004.
research in the field of link adaptation algorithms for multicarrier
[7] Kim J, Ashkhimin A, Wijngaarden A, Soljanin E. systems for next generation wireless networks.
Reverse link hybrid ARQ link error prediction
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 87
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Reliable Mining of Automatically Generated Test

Cases from Software Requirements Specification
Lilly Raamesh1 and G. V. Uma2
Research Scholar, Anna University,
Chennai 25, India

Asst. Professor/CSE, Anna University,
Chennai-25, India

Writing requirements is a two-way process. In this paper we 2. Our Approach
use to classify Functional Requirements (FR) and Non
Functional Requirements (NFR) statements from Software
Requirements Specification (SRS) documents. This is Classifier
systematically transformed into state charts considering all
relevant information. The current paper outlines how test cases Training set
can be automatically generated from these state charts. The
application of the states yields the different test cases as
solutions to a planning problem. The test cases can be used for User
automated or manual software testing on system level. And requirements Rule
also the paper presents a method for reduction of test suite by generator
using mining methods thereby facilitating the mining and
knowledge extraction from test cases.
Keywords: SRS, FR, NFR, State model, Test case, Test suite,
Mining. Classification
1. Introduction

The systematic production of high-quality software,

which meets its specification, is still a major problem. Functional and non-
Although formal specification methods have been around functional
for a long time, only a few safety-critical domains justify requirements
the enormous effort of their application. The state of the
practice, which relies on testing to force the quality into
the product at the end of the development process, is also Generation of State
unsatisfactory. The need for effective test automation Diagram
adds to this problem, because the creation and
maintenance of the test ware is a source of inconsistency
itself and is becoming a task of comparable complexity
as the construction of the code. Clustering
techniques Test Suite
Data mining algorithms can be applied at different levels
of abstraction and help the user discover more
meaningful patterns. Data mining will create patterns
from the existing database. Using well-established data
mining techniques, practitioners and researchers can Mined test cases
explore the potential of this valuable data in order to
manage their project and to produce higher quality Fig 1. Automatically Generated Test Cases From Software
software systems that are delivered on time and within Requirements Specification mining System
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 88

Our approach is as follows tester. Once the test data corresponding to a particular
i. Generation of classification rules. predicate are determined, the steps are repeated by
ii. Generate test cases from the UML state machine. selecting the next predicate on the state machine diagram.
iii .Finally data mining techniques are applied on the The process is repeated until all Predicates on the state
generated test cases in order to further reduce the machine diagram have been considered.
test suite size.
3.1. Predicate selection:
2. Generation of Classification Rules
For selecting a predicate, a traversal of the state diagram
In the current paper, we provide the Software is performed using depth first (DFS) traversal or breadth
Requirements Specification to the classifier system. For first (BFS) traversal to see that every transition is
classifying we use Weka. The Weka classifier is initially considered for predicate selection. DFS traversal is used
trained with a training set. Later it is provided with the here. During traversal, conditional predicates on each of
SRS. It classifies SRS in to functional and non functional the transitions are looked. Corresponding to each
requirements by generating a classification rules. The conditional predicate, test data are generated.
classification rules are applied to the SRS to get FR and
NFR. The test data are generated for each predicate
From NFR we derive the state machine. State machines corresponding to the true or false values of the
specify the behaviour of a system/subsystem. conditional predicate satisfying the prefix path condition.

3.2. Predicate transformation:

3. Generation of Test Cases
Let I0 consists of all variables that affect a predicate q in
This section briefly describes a transformation from State the path P of a state machine diagram, then two points
diagrams in to Test cases. State machines and state named ON and OFF for a given border satisfying the
diagrams have a long history in computer science. boundary-testing criterion are created. The relational
Recent versions of UML include an expressive state expressions of the predicates are transformed into a
diagrams concept. Especially the abstraction mechanisms function F called predicate function. If the predicate q is
in the UML state machine formalism, i.e. nesting of of the form (E1 op E2), where E1 and E2 are arithmetic
states and stubs, allow us to map all the important Expressions and op is a relational operator; then F = (E1
elements of our use case documents to State machines. -E2) or (E2 - E1) depending on whichever is positive for
the data I0. Next, the input data I0 is modified such that
From State Machines to Test Cases the function F decreases and finally turns negative. When
F turns negative, it corresponds to the alternation of the
Using state models to derive test cases has been common outcome of the Predicate. Hence, as a result of the
practice in the software testing world for some time .The predicate transformation, the point at which the outcome
final goal of model-based testing is to automate the test of a predicate q changes, corresponds to the problem of
case generation from test models as much as possible. minimization of the corresponding function F. This
Our approach generates a set of valid test sequences, minimization is achieved through repeated modification
where the preconditions of all transitions are established of the input data values.
either by previous actions or by properties of the test data.
The scope of our method is the generation of test 3.3. Test data generation:
sequences supplemented by constraints on the test data,
as far as these can be derived from the information The basic search Procedure we use for finding the
presenting the state machine. minimum of a predicate function F is the alternating
variable method. This method is based on minimizing F
The method given in [3] can be used to create test cases with respect to each input variable in turn. An initial set
from state machine diagrams which is as follows: of inputs can be randomly generated by instantiating the
data variables. Each input data variable xi is increased/
There are three main steps in test case generation, in the decreased in steps of Sxi, while keeping all other data
first step a predicate is selected on a transition from a variables unchanged. Here, Sxi refers to a unit step of the
UML state machine diagram. In the next step, the variable xi. The exact value of unit step can be defined
selected predicate is transformed into a predicate conveniently. For example, unit step of 1 is normally
function. In the third step, test data are generated used for integer values. Unit step can easily be defined
corresponding to the transformed predicate function. The
generated test data are stored for use with an automatic
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 89
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

for many other types of data such as float, double, array, 4. Mining Techniques For Test Suite
and pointer and so on. Reduction
However, the method may not work when the variable
assumes only a discrete set of values. Each predicate in a Data mining is the process of extracting patterns from
path can be considered to be a constraint. A path will not data. As more data are gathered, with the amount of data
be traversed for some input data value, if the doubling every three years, data mining is becoming an
corresponding constraint is not satisfied. If a path P is not increasingly important tool to transform these data into
traversed for some data value, then we say that a information. It is commonly used in a wide range of
constraint violation has taken place for that data value. profiling practices, such as marketing, surveillance, fraud
We compute the value of F when each input datum is detection and scientific discovery.
modified by Sxi. If the function F decreases for the
modified data, and Constraint violation does not occur, While data mining can be used to uncover patterns in
then the given data variable and the appropriate direction data samples, it is important to be aware that the use of
is selected for minimising F further. Here, appropriate non-representative samples of data may produce results
direction refers to whether we increase or decrease the that are not indicative of the domain. Similarly, data
data variable xi so that F is minimised. We start mining will not find patterns that may be present in the
searching for a minimum with one input variable, while domain, if those patterns are not present in the sample
keeping all other input variables constant, until the being "mined". There is a tendency for insufficiently
solution is found (the predicate function becomes knowledgeable "consumers" of the results to attribute
negative) or the positive minimum of the predicate "magical abilities" to data mining, treating the technique
function is located. In the latter case, the search as a sort of all-seeing crystal ball. Like any other tool, it
continues from this minimum value with the next input only functions in conjunction with the appropriate raw
variable. Two data values Iin (inside boundary) and Iout material: in this case, indicative and representative data
(outside boundary) are generated using the search that the user must first collect. Further, the discovery of a
procedure mentioned. These two points are on different particular pattern in a particular set of data does not
sides of the boundary. necessarily mean that pattern is representative of the
whole population from which that data was drawn.
For finding these two data points, a Series of moves is Hence, an important part of the process is the verification
made in the same direction determined by the search and validation of patterns on other samples of data.
procedure mentioned above and the value of F is
computed after each move. The size of the step is The term data mining has also been used in a related but
doubled after each successful move. This makes the negative sense, to mean the deliberate searching for
search for the test data quick. A successful move is one apparent but not necessarily representative patterns in
where the value computed by the predicate function F is large numbers of data. To avoid confusion with the other
reduced. When the Minimisation function becomes sense, the terms data dredging and data snooping are
negative (or zero), the required data values Iin and Iout often used. Note, however, that dredging and snooping
are noted. These Points are refined further to generate a can be (and sometimes are) used as exploratory tools
data value, which corresponds to a minimum value of the when developing and clarifying hypotheses.[6]
minimisation function along the last processed Direction.
This refinement is done by reducing the size of the step 4.1. Applying data mining concepts
and comparing the value of F with the previous value.
Also, the distance between the data points is minimised There are many methods available for mining different
by reducing the step size. For each Conditional predicate kinds of data, including association rule, characterization,
in the state machine diagram, we generate the test data. classification, clustering, etc.
The generated test data are stored in a file. A test
executor can use these test cases later for automatic We can utilize any of these techniques based on
• What kind of data bases to work on
The above said procedure produces a test suit that is of
• What kind of knowledge to be mined
some what smaller size. But we can further reduce the
• What kind of techniques to be utilized
size by using mining techniques.
We can apply association or clustering
techniques for test case mining.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 90

4.1.1. ASSOCIATION o Ability to deal with noisy data:

o Insensitivity to the order of input records:
Association rules describe the association among items o High dimensionality:
in the large data base. For example, one may find, from
a large set of transaction data, such as association rule as 5. Experiments
if a customer buys(one brand of) milk, he/ she usually
buys(another brand of) bread in the same transaction.
Using these association rules, we can derive the Total 365 sentences
association patterns from large databases. • 235 annotated as “NFR”
• 130 annotated as “FR”.
Training Correctly In-Correctly
set no classified classified sentences
Data classification is the process which finds the sentences
common properties among a set of objects in a database 1 215 150
and classifies them into different classes, according to a 2 259 106
classification model. 3 356 9

4.1.3. CLUSTERING 6. Conclusions

Clustering is the process of grouping the data into classes In this paper, a new approach to automatically generate
or clusters so that object within a cluster has high test cases from SRS and mining of test cases has been
similarity in comparison to another, but is dissimilar to discussed. Firstly a formal transformation of a detailed
object in other clusters. It doesn’t require the class label SRS to a UML state model, secondly the generation of
information about the data set because it is inherently a test cases from the state model and lastly mining of Test
data driven approach. It is the process of grouping or cases. The introduction of agents can bring enhancement.
abstract object into classes of similar object.
Among all the mining techniques, clustering is the most
effective technique which we are going to use for test 1.Zhijie Xu, Laisheng Wang, Jiancheng Luo, Jianqin Zhang,
case mining. “A Modified clustering algorithm for data mining”, Beijing
100101, China.
2. Ming-Syan Chen, Senior Member, IEEE Jiawei Han, Senior
Clustering analysis helps constant meaningful Member, IEEE, and Philip S.Yu, Fellow, IEEE. “ Data Mining:
partitioning of a large set of object based on a “divide An Overvies from a Database Perspective”.
and conquer” methodology which decomposes a large 3. Automatic test case generation using unified Modeling
scale system into smaller components to simplify design language (UML) state diagrams, P. Samuel R. Mall A.K.
and implementation. As a data mining task, data Bothra, Department of Computer Science and Engineering,
clustering identifies cluster or densely populated regions, Indian Institute of Technology, Kharagpur 721302, West
according to some distance measurement, in a large, Bengal, India
multidimensional data. Given a large set of E-mail:
multidimensional data points, the data space is usually 4.Dae-Kyoo Kim, Jon Whittle, ”Generating UML Models from
Domain Patterns”, USA.
not uniformly occupied by the data points. Data 5. Tao Xie,USA. Jian Pei, Canada, Ahmed E.Hassan,Canada.
clustering identifies the sparse and the crowded places, “Mining software Engineering Data”.
and hence discovers the overall distributions patterns of 6.”A Comparative Evaluation of Tests Generated from
the data set. Different UML diagrams”, Supaporn Kansomkeat, Department
of Computer Science Faculty of Science, Prince of Songkla
For cluster analysis to work efficiently and effectively as University Hat Yai, Songkhla, 90112, Thailand”.
many literatures have presented, there are the following 7. Sarma M., "System State Model Generation from UML 2.0
typical requirements of clustering in data mining. Design", Technical Report TR-04-07, Department of Computer
Science and Engineering, Indian Institute Of Technology,
Kharagpur, April 2007
o Scalability: 8. Castejon H. N. "Synthesizing State-machine Behavior from
o Ability to deal with different types of attributes: UML Collaboration and Use Case Maps", Lecture Notes in
o Discovery of clusters with arbitrary shape: Computer Science, Vol. 3530, Springer, June 2005.
o Minimal requirements for domain knowledge to 9. Gupta A., "Automated Object's Statechart Generation and
determine input parameters: Testing from Class Method Contracts", 3rd Intl Workshop on
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 91
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Model Development, Validation, and Verification (MoDeV2a

2006), Genova, Italy, October 2006.
10. Vasilache S., and Tanaka J., "Synthesis of State Machines
from Multiple Interrelated Scenarios Using Dependency
Diagrams," Journal of Systemics, Cybernetics and Informatics,
Vol.3, No.3, 2006.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 92
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Understanding Formulation of Social Capital in Online Social

Network Sites (SNS)
S.S.Phulari,1, Dr.S.D.Khamitkar2 , N.K.Deshmukh3, P.U.Bhalchandra4 , S.N.Lokhande5 and A.R.Shinde6
School of Computational Sciences,
Swami Ramanand Teerth Marathwada University, Nanded, MS, India, 431606

Department of Computer Science,
N.S.B. College, Nanded, MS, India, 431601

Abstract Influence: An individual performing an action can cause

some contacts to do the same by providing information or
Online communities are the gatherings of like-minded people, by increasing the value of the action to them.
brought together in cyberspace by shared interests. The shared
interest has hidden social capital aspects and can be of bridging
or bonding type .Creating such communities is not a big Homophily: Similar individuals are more likely to be
challenge but sustaining member’s participation is. This study friends. For example two geologists are more likely to
examines the formation and maintenance of social capital in
know each other than two random people.
social network sites. In addition to assessing bonding and
bridging social capital, we explore a dimension of social capital
that assesses one's ability to stay connected with members of a
previously inhabited community, which we call maintained Environment: Influence from external elements. Friends
social capital. Such dimension is enacted here in terms of are more likely to live in the same area, thus attend and
Hypothesis. take pictures of similar events, and tag them with similar
tag. This focus on a particular action like buying a
Keywords: Social Area Networks, Social Capital, Analysis. product, joining a community, publishing in a conference,
using a particular tag, getting premier membership, etc

1. Introduction Social network analysis goes back to almost half a

century of work in sociology and social psychology.
Social network are the graph that represents Milgram in 1967 attempted to define it in concise form
relationships between users. These can take any shape, and stated that two random people were connected on
including average by a path of six acquaintances. This is called as
• Offline Social connections (friends, clubs, six degrees of separation theorem. Similarly, Zachary in
groups, associations) 1972 discussed social relationships and rivalries in a
• Online social connections (Face book, University Karate Club (via graph min cuts). In recent
MySpace, Live Journal, Orkut, LinkedIn) years, as the world is migrating to global village or hyper
• Messaging (IM, chat, address book) world, large number of social networks found activated,
• Social book marking (Digg, Delicious) with thousands of users associated? For theorist computer
• Content sharing (Flickr, YouTube) scientists, this exponential growth or the resulting data is a
goldmine for giving a new perspective via sociological
It is evident from the exponential growth of Internet studies, data mining studies, pattern matching studies, etc.
users in recent year that, more and more of people’s This data is of massive amounts, as hundreds of millions
interactions are happening online. Hence, now days, social of users are there. This data has a study phenomenon at
networks are commonly seen to be associated with different scales, for example interaction of people in
Internet. Such networks must get studied with the type of focused groups of different sizes, overall structure of the
social correlation the user community has. We found three network including regional and outside-region support,
sources of social correlation as defined below, etc. This data has ability to measure, record, and track
individual activities at the finest resolution also. For
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 93

example, user befriending another, user buying a DVD, systems would be connecting with others outside their pre-
user tagging a photo, etc. existing social group or location, liberating them to form
communities around shared interests, as opposed to shared
This study is important today in order to find out geography [2, 5].This could be interesting to predict about
behavior pattern of users connected via social networks. social capital. A hallmark of this early research is the
This helps us to predict the dynamics of the network presumption that when online and offline social networks
system also. At the same time , this study help us to talk overlapped, the directionality was online to offline-online
about new norms of behavior, technology, influences or connections resulted in face-to-face meetings. Although
idea , we can discover after analyzing data by questioning early work on SNS addressed many interesting findings,
and analyzing how similar is the behavior of connected we find no traces of social capital. This is because there is
users ? , how similar the structure of services and little empirical research that addresses whether members
technologies social network sites have? .We can also use SNSs to maintain existing ties or to form new ones,
device key indictors of social networks from such study. the social capital implications of these services is
A popular Social Network researcher, Wellman [1]
argued that the concept of "social networks" is difficult to The SNS like Face book constitutes a rich site for
define. Usually a social network is defined as relations researchers interested in the affordances of social
among people who deem other network members to be networks due to its heavy usage patterns and technological
important or relevant to them in some way [2]. Using capacities that bridge on line and offline connections.
media to develop and maintain, social networks is Research suggests that Face book users engage in
established in practice. Now days, this concept is "searching" for people with whom they have an offline
popularly implemented in term of Internet sites. These connection more than they "browse" for complete
websites allow participants to construct a public or semi- strangers to meet [4]. We believe that this approach in
public profile within the system and formally articulate Face book represents an understudied offline to online
their relationship to other users in a way that is visible to trend where it originally, primarily served a
anyone who can access their profile [3, 4] We feel that geographically-bound community (the campus). However,
this definition does not specify the closeness of any given as social capital researchers we have another perspective
connection, but only that participants are linked in some to look into – findings of social capital as place-based
fashion. community facilitate the generation of social capital.

Beside, being just a gatherings of like-minded people, We are optimistic to get many social capitals and if a
brought together in cyberspace by shared interests linkage, Regression analysis is conducted on them then it will
Social Network Sites (SNSs) such as such as MySpace definitely suggest a strong association between use of SNS
allow individuals to present themselves, articulate their and the type of social capital, with the strongest
social networks, and establish or maintain connections relationship being to bridging social capital. In addition
with others. These sites can be oriented towards work- we could also find hidden aspects to interact with the
related contexts like, romantic relationship measures of psychological well-being. This discovery may
initiation like, and for connecting those help in suggesting that social capital and SNS can provide
with shared interests such as music or politics, or the greater benefits for users experiencing low self-esteem and
college student population like the low life satisfaction. This is because a depressed person
has more tendencies to express feeling than a normal
The SNS operational style can be online or offline. person. [6].Such study brings a requirement to statistically,
The offline style has no face to face communications. The analytically process SNS data. However a social network
online social network application enables its users to will definitely look different depending upon how one
present themselves in an online profile, accumulate measures it, like, counting the number of interactions
"friends" who can post comments on each other's pages, between members (say “a”) or rating the closeness of
and view each other's profiles. For example, Face book relationships (say “b”) or “a versus b”, etc .
members can also join virtual groups based on common
interests, see what classes they have in common, and learn
each others' hobbies, interests, musical tastes, and 2. An Overview of Face book
romantic relationship status through the profiles. Online
SNSs support both the maintenance of existing social ties Created in 2004, by 2007 Face book was reported to
and the formation of new connections. Early research on have more than 21 million registered members generating
online communities assumed that individuals using these 1.6 billion page views each day The site is tightly
integrated into the daily media practices of its users: The
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 94
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

typical user spends about 20 minutes a day on the site, and increases and decreases in social capital [1]. We can probe
two-thirds of users log in at least once a day [7]. into social capital in two ways,
Capitalizing on its success it created separate versions for
high school students, communities for commercial Bridging (weak ties): loose connections between
organizations. individuals which may provide useful information or new
perspectives for one another
Face book s widely used for SNA due to its openness
and large customer base. Much of the existing academic Bonding (strong ties): strong connection between
research on Face book has focused on identity individuals. Usually emotionally close relationships, such
presentation and privacy concerns [8]. Looking at the as family and close friends.
amount of information Face book participants provide
about themselves, the relatively open nature of the Recently, researchers have emphasized the
information, and the lack of privacy controls enacted by importance of Internet-based linkages for the formation of
the users, argue that users may be putting themselves at weak ties, which serve as the foundation of bridging social
risk both offline (e.g., stalking) and online (e.g., identify capital. Because online relationships may be supported by
theft). Other recent Face book research examines student technologies like distribution lists, photo directories, and
perceptions of instructor presence and self-disclosure [9] search capabilities [15], it is possible that new forms of
temporal patterns of use [10] and the relationship between social capital and relationship building will occur in online
profile structure and friendship articulation [4]. social network sites. Bridging social capital might be
augmented by such sites, which support loose social ties,
The literature survey [7] shows that Face book is allowing users to create and maintain larger, diffuse
used more by college-age students and was significantly networks of relationships from which they could
associated with measures of social capital. We use Face potentially draw resources [10].
book as a research context in order to determine whether
knowledge about social capital patterns can be generated
by enacting some theorems or hypothesis. 4. Understanding Formulation of Social Capital
After briefly describing, to the extant of literature [9,
3. Defining Social Capital 11, 16, 17, and 18] regarding the forms of social capital
and the impact of Internet, we introduce some basic
Social capital is an elastic term. It broadly refers to hypothesis for social capital which speaks to the ability to
the resources accumulated through the relationships maintain valuable connections as one progress through life
among people. It has variety of definitions [11]. It is often changes.
conceived as a cause and an effect for social networking.
For individuals, social capital allows a person to draw on 4.1. Hypothesis 1: Use of SNS will be proportional
resources from other members of the networks to which he to user’s perceived bonding social capital.
or she belongs. These resources can take the form of
useful information, personal relationships, or the capacity Explanation: Bonding social capital reflects strong
to organize groups [12].This also give capabilities to ties with family and close friends. Day by day, internet is
access to individuals outside one's close circle and help to maturing and providing new means, connections to
get non-redundant information, resulting in benefits such connect and come closer. These new connections may
as employment connections [13] . result increase in social capital as users might be in a
position to provide emotional support whenever needed.
Social capital has been linked to a variety of positive Thus, as the tendency to share interests or regional goals
social outcomes, such as better public health, lower crime increases, user starts bonding the social capitals and use of
rates, and more efficient financial markets [13].When SNS will increase subsequently. This is the reason why
social capital declines, a community experiences increased terrorists frequently use SNS like Orkut and Face Book.
social disorder, reduced participation in civic activities, Online users are more likely to have a larger network of
and potentially more distrust among community members. close ties than non-Internet users, and Internet users are
Greater social capital increases commitment to a more likely than non-users to receive help from core
community and the ability to mobilize collective actions, network members. This proves our hypothesis.
among other benefits. Social capital may also be used for
negative purposes, but in general social capital is seen as a
positive effect of interaction among participants in a social
network [14]. The Internet has been linked both to
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 95

4.2. Hypothesis 2: Bridging social capital is largely

affected by the psychological status of users across
SNS 4.4. Hypothesis 4: A bridging social capital always
maintain the social capital
Explanation: Usually SNS help an individual who
have difficulties .Consider a research scholar having less Explanation: Literature review suggests that the
economical affordability. For completion of the project he weak ties provide more benefit when the weak tie is not
or she has undertaken, he or she will definitely need associated with stronger ties, as may be the case for
availability of full size citation papers. Considering the maintained high school relationships. This is because the
time constraint associated with the research project and high school relations are friendship relations and have no
because of sever economical backwardness, he or she may much emotional, business gravities. However this is
become panic and get attached to contemporary research subject to psychological aspects, which rarely play a role.
platforms via SNS where he or she could find some of For example, a child of divorced parents or a child of
research literature by bridging social capital. We will not working parents needs a bonding relationship with his
observe bonding social capital because the connection friend. Such ties are only maintained in high school
established is temporary and does not reflect long term, relationships even after leaving schools .We call this
core and sturdy relationships. If his or her economical concept "maintained social capital."
conditions were good or there was no narrow time
constraint with the project, then the person would not
became panic and would never had accessed SNS / 5. Conclusions
internet. Thus SNS may be useful for individuals who
The findings of this exploratory study on social
have difficulties or individuals with low psychological
capital suggest that research efforts are required in the area
levels .However bridging social capital will vary
of user behavior and SNS modeling, with a special
depending on the degree of a person's self esteem or
attention to the definition social capital. We have
depending on the degree of a person's satisfaction with
elaborated three measures of social capital including
bridging, bonding, maintained social capital and analyzed
user’s behavior with respect to them. It is evident from our
4.3. Hypothesis 3: Change in user’s life style work that internet is linked to both increase and decrease
in social capital. After briefly describing to the extant of
adversely affects social capital
literature on these three forms of social capital and the
Explanation: Social networks change over time as Internet, we introduced an additional dimension of social
relationships are formed or abandoned. Particularly capital that speaks to the ability to maintain valuable
significant changes in social networks may affect one's connections as one progress through life changes. The
social capital, as when a person moves from a geographic conclusions put here are mainly for most popular SNS-
location in which his network was formed and thus loses Face Book, as much of the existing academic research has
access to those social resources. The possible causes of been focused on it. This could be due to its relatively open
decreased social capital can be many including increase in nature. Basic theorems are also mentioned to support our
families moving for job reasons, natural disasters, findings.
migrations, or even discrete internet connectivity which
results increasing tendency to rely on emails for long References
distance communications. The time zone, working timings [1] Wellman, B., Haase A., Witte J. and Hampton K “Does
also affect the social capital .The connectivity between the Internet Increase, Decrease, or Supplement Social
Capital? , appeared in Social Networks, Participation and
parents and children pursuing their education abroad have
Community Commitment, American Behavioral Scientist,
affected social capital as timing zones are different and it issue 45, November 2001.
is highly impossible to remain online. We found that [2] Wellman B, Computer Networks as Social Networks,
services like email and instant messaging help college Science press,2001, pp 231-234.
students remain close to their high school friends after [3] Danah Michele Boyd, Friendster and publicly articulated
they leave college for jobs. This is known as social networking, extended abstracts on Human factors in
“Fiendsickness”. Young adults moving from high schools computing systems, April 2004, pp 24-29, Vienna,
often leave friends from high school with whom they may Austria [DOI-10.1145/985921.986043]
have established rich networks; completely abandoning [4] Ellison N., Lampe C. and Steinfield C.,Spatially Bounded
Online Social Networks and Social Capital: The Role of Face
these high school networks would mean a loss of social
book, International Conference on Communication
capital. Association, Dresden, 2006.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, No. 3, January 2010 96
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

[5] Adam N. Joinson,Looking at, looking up or keeping up with

people?: motives and use of face book, Proceeding of the
twenty-sixth annual SIGCHI conference on Human factors in S.S.Phulari: He is a research student working on social area
computing systems, 2008, pp 05-10, Florence, Italy networking under the supervision of Dr.S. D. Khamitkar. He has
[6] Utz S,the Development of Friendships in Virtual Worlds, completed his M.Phil and has 2+ papers in international
Social Information Processing in MUDs: Journal of Online conferences/ journals
Behavior, 2000.
Dr.S.D.Khamitkar : He has 14+ years experience in teaching and
[7] Jennefer Hart , Charlene Ridley , Faisal Taher , Corina Sas ,
research. He is supervising ten research students currently and
Alan Dix,Exploring the face book experience: a new has 8+ papers in international conferences and journals
approach to usability, Proceedings of the 5th Nordic
conference on Human-computer interaction: building N.K.Deshmukh,S.N.Lokhande and P.U.Bhalchandra : These are
bridges, October 2008, Lund, Sweden faculties at University Department and have 5+ research papers in
[8] Cliff A. C. Lampe, A familiar face (book): profile elements international conferences and journals .
as signals in an online social network, Conference on Human A.R.Shinde : He is a faculty at one of the affiliated college and
Factors in Computing Systems proceedings of the SIGCHI has ten years teaching experience. He is currently planning for
conference on Human factors in computing systems, San research work and actively involved in understanding research
Jose, California, USA process.
[9] Stutzman, F, An Evaluation of Identity-Sharing Behavior in
Social Network Communities, appeared in iDMAa and IMS
Code Conference, Oxford, 2005.
[10]Sunnafrank M, Predicted outcome value during initial
interactions: A reformulation of uncertainty reduction theory,
(1986), Journal of Human Communication Research, 13. pp
[11]Joan DiMicco , David R. Millen , Werner Geyer , Casey
Dugan , Beth Brownholtz , Michael Muller,Motivations for
social networking at work, Proceedings of the ACM 2008
conference on Computer supported cooperative work,
November 2008, San Diego, CA, USA
[12]John M. Carroll , Mary Beth Rosson , Theorizing mobility in
community networks, , International Journal of Human-
Computer Studies, v.66 n.12, , Dec. 2008 ,pp.944-962 .
[13]Jenny Preece , Diane Maloney-Krichmar, Focusing on
sociability and usability, Online communities: The human-
computer interaction handbook: fundamentals, evolving
technologies and emerging applications, Lawrence Erlbaum
Associates, Inc., Mahwah, NJ, 2002
[14]Adam N. Joinson,Looking at, looking up or keeping up with
people?: motives and use of facebook, Proceeding of the
twenty-sixth annual SIGCHI conference on Human factors in
computing systems, April 2008, Florence, Italy
[15]Leysia Palen , Sarah Vieweg,the emergence of online wide
scale interaction in unexpected events: assistance, alliance &
retreat, Proceedings of the ACM 2008 conference on
Computer supported cooperative work, November 2008, San
Diego, CA, USA
[16]Charles Steinfield , Joan M. DiMicco , Nicole B. Ellison ,
Cliff Lampe,Bowling online: social networking and social
capital within the organization, Proceedings of the fourth
international conference on Communities and technologies,
June 2009, University Park, PA, USA
[17]Ellison, N., Lampe, C. and Steinfield, C, Spatially Bounded
Online Social Networks and Social Capital: The Role of Face
book. Proceedings of Journal of International
Communication Association, Dresden, 2006.
[18]Brain Knudsen, Richard Florida and Denise Rousseau ,
Dissertation Report on Bridging and Bonding – A multi
Dimensional Approach to Regional Social Capital ,the
Martin Prosperity Institute , University of Toronto , 2007
Volume 7, Issue 4

The topics suggested by this issue can be discussed in term of concepts, surveys, state of the
art, research, standards, implementations, running experiments, applications, and industrial
case studies. Authors are invited to submit complete unpublished papers, which are not under
review in any other conference or journal in the following, but not limited to, topic areas.
See authors guide for manuscript preparation and submission guidelines.

Accepted papers will be published online and authors will be provided with printed
copies and indexed by Google Scholar, Cornell’s University Library,
ScientificCommons, CiteSeerX, Bielefeld Academic Search Engine (BASE), SCIRUS
and more.

Deadline: 31st May 2010

Notification: 30th June 2010
Revision: 10th July 2010
Online Publication: 31st July 2010

• Evolutionary computation • Software development and

• Industrial systems deployment
• Evolutionary computation • Knowledge virtualization
• Autonomic and autonomous systems • Systems and networks on the chip
• Bio-technologies • Context-aware systems
• Knowledge data systems • Networking technologies
• Mobile and distance education • Security in network, systems, and
• Intelligent techniques, logics, and applications
systems • Knowledge for global defense
• Knowledge processing • Information Systems [IS]
• Information technologies • IPv6 Today - Technology and
• Internet and web technologies deployment
• Digital information processing • Modeling
• Cognitive science and knowledge • Optimization
agent-based systems • Complexity
• Mobility and multimedia systems • Natural Language Processing
• Systems performance • Speech Synthesis
• Networking and telecommunications • Data Mining

For more topics, please see

All submitted papers will be judged based on their quality by the technical committee and
reviewers. Papers that describe research and experimentation are encouraged.
All paper submissions will be handled electronically and detailed instructions on submission
procedure are available on IJCSI website (

For more information, please visit the journal website (

The International Journal of Computer Science Issues (IJCSI) is a refereed journal for
scientific papers dealing with any area of computer science research. The purpose of
establishing the scientific journal is the assistance in development of science, fast
operative publication and storage of materials and results of scientific researches and
representation of the scientific conception of the society.

It also provides a venue for researchers, students and professionals to submit on-
going research and developments in these areas. Authors are encouraged to
contribute to the journal by submitting articles that illustrate new research results,
projects, surveying works and industrial experiences that describe significant advances
in field of computer science.

Indexing of IJCSI:

1. Google Scholar
2. Bielefeld Academic Search Engine (BASE)
3. CiteSeerX
5. Docstoc
6. Scribd
7. Cornell’s University Library
8. SciRate
9. ScientificCommons