Editors Column
Great ideas, it is said, come into the world as gently as doves. Perhaps, then, if we listen attentively, we shall hear amid the uproar of empires and nations a faint flutter of wings; the gentle stirring of life and hope. --Albert Camus Great environments are definitely breeders of great ideas but everybody needs a break from the tight timelines to nurture such ideas. As we enjoy the last few days of the mid-sem break, let us refresh ourselves with the latest updates of the activities of the department and the people associated with it. In this issue we heartily welcome two new faculty Dr. Maya Ramanath and Dr. Smruti Ranjan Sarangi to the department. We also thank Dr. Sarangi for his highly motivating article on innovation in an Indian context. A live example of this is discussed in an interview with an Alumini, also the creator of Flipkart, one of the most popular online shopping ventures in India. Ideas flow from one plain to another, be it through the open house, various seminars or other events organized in the department. The ACM Chapter organized many major events including Yahoo Hack U! and IKES 2011, a research based knowledge exchange symposium organized in collaboration with IBM IRL. The team offers its homage to one of the greatest innovators of the era, Steve Jobs, through his personality profile as well as some of his famous quotes. Switch on your thinking pads for the fun section that offers a new poem and a crossword with a difference. Read on for all this and more!
Innovation in an Indian Context Project article I2Tech 2011 Newsflashes Faculty Profile Personality Profile Faculty Profile Flipkart profile ACM activities Fun Section
1,2
3 4
5
Most undergraduate and gradu- -partments, an iPod, would not have ate students consider the idea of been possible. We can think of an creating a startup at some point iPod as the outcome of a massive of time in their life. They are system consisting of many many mostly inspired by the lives of gears, all operating in perfect harmony and synchrony. The loss of visionaries like Steve Jobs, and any single cog in the wheel can stall Bill Gates. Also, creating a sucthe entire system. cessful startup is a very superlative technical achievement, an For example, when I went to extremely attractive option both Ethiopia, I saw a man sitting by professionally as well as finan- the road side making new TVs. cially. Consequently, many stu- He had a couple of old CRT TV dents, academicians, and work- sets of different makes. He was ing professionals are drawn into cannibalizing different parts from this seemingly lucrative profes- them, and creating a brand new sion. However, the economics of TV set consisting of entirely reinnovation is extremely different furbished parts. This to me is, in India as compared to the Sili- innovation. Doing this surely con Valley. I would tend to ar- does require some amount of gue that the meaning of the term technical knowledge. Secondly, innovation means different you require a supply chain that things in different places. It provides old TVs with at least one needs to be interpreted in the part working, and then you need a market that can correct pay you well s o c i o Since we live in an age of for the invested econ omi c and c o n t e x t . innovation, a practical education t i m e This article must prepare a man for work that money. makes the does not yet exist and cannot yet Now, coming case for a be clearly defined back to the more bal~Peter F. Drucker Indian context, anced aplet us look at proach for the financial dealing with angle first. To make a product, we the issue of pursuing a career need at least three engineers. Two founding startups as compared to engineers need to be devoted to traditional jobs in India. development, and one engineer When we think of innovation, needs to focus on testing. Even if the first thing that strikes us is we assume that a couple of entresomething like an iPod, or iPad. preneurs dont draw any salary, We need to fundamentally revisit we will nonetheless need to pay this notion of invention in devel- the engineers at multinational oping countries in India. We company rates. Assuming an avneed to understand that to create erage figure of ten to fifteen lakh something like an iPod, we need rupees CTC per year, this hundreds of positive factors in amounts to about 40 lakhs per our socio-economic system like year for three engineers. Secthe availability of high quality ondly, we need some basic comtechnical manpower, designers, puting infrastructure, licenses for venture capitalists, marketing software, and money to pay utility managers, and extremely diligent bills, and rent. Assuming five sales personnel. If Apple were lakhs for rent per year, and the deficient in any one of these de- rest for computing facilities, we
CSE Newsletter
Here's to the crazy ones, the misfits, the rebels, the troublemakers, the round pegs in the square holes... the ones who see things differently -- they're not fond of rules... You can quote them, disagree with them, glorify or vilify them, but the only thing you can't do is ignore them because they change things... they push the human race forward, and while some may see them as the crazy ones, we see genius, because the ones who are crazy enough to think that they can change the world, are the ones who do. ~ Steve Jobs
Page 2
An interesting project article: Systematic Ranking of Thread Schedules for Testing Multithreaded Programs
Student : Sandeep Kumar Bindal Supervisor : Dr. Sorav Bansal Publication details : S. Bindal, S. Bansal, "Systematic ranking of Thread Schedules for Testing Multithreaded Programs," ACM SRC PLDI Student Research Competition, San Jose, June 2011. (Third Prize Winner in Undergraduate Research Category. ) Introduction Concurrency is used pervasively in the development of large programs. However, testing concurrent programs is notoriously hard due to the inherent non-determinism. Recently, many different techniques have been proposed to identify concurrency-related bugs early. An effective but expensive approach is modelchecking, where all possible schedules of a program are executed to ascertain the absence of a bug. Unfortunately, the space of all schedules is huge, and exhaustively enumerating it is usually infeasible. Therefore, an alternative is to try and maximize the probability of uncovering a bug rather than trying to ascertain its absence. Background Work Many different approaches have been proposed in this direction. Musuvathi and Qadeer recently proposed using context-bound to rank schedules, and show that it is an effective method to uncover most common bugs. A context-bound is the number of preemptive context-switches required to execute a schedule. The schedules are enumerated in increasing order of their contextbound. They had shown that all known bugs in real-world applications can be found with context-bound values of 2 or less. The space of schedules is reduced significantly at low contextboundvalues, but it is still too large to exhaustively enumerate. Tools like CHESS based on this algorithm reduces this search space by considering only explicit synchronization operations as possible pre-emption points. This approach is based on the assumption that all shared-memory accesses will be protected by a lock but many systems deliberately avoid explicit synchronization, often for performance reasons. Another recent tool, CTrigger, focuses on atomicity-violation bugs and preferentially searches the space of schedules that are likely to trigger these bugs. CTrigger first profiles executions of the program to determine the shared variables and their unprotectedaccesses. It then attempts to generate schedules that are thread schedules to be tested on four dimensions, namely, contextbound, memory locations (variables), threads, and loop iterations. We discuss all these ranking criteria as follows:
Ranking on Context Bound Ranking on Number of Memory Locations Ranking on Number of Threads Ranking on Loop Iteration Bound
Conclusions We have implemented this algorithm in a concurrency testing tool for Java, called RankChecker. Our tool does not require source-level annotations and instruments the binary class code of a Java program and associated libraries to insert appropriate schedule points. The instrumented test program is linked with the RankChecker library which spawns a scheduler thread to dictate the thread interleavings using semaphores. Table 1. shows the results of our experiments. The programs from MergeSort to AirLineTickets are taken from ConTest Benchmarks. We are able to find out all the known bugs in the ConTest Benchmarks. We have also tested our tool on cache4j, a fast threadsafe implementation of a cache for Java Objects, upto (c = 2, v = vtot, t = 2, l = 1) but we are not able to find out any bug. Other key contributions of our work include : Providing a systematic ranking of test schedules while testing concurrentprograms. Presenting new classifications of bugs based on the number of variables involved, the number of threads required to uncover it, and the number of loop iterations. Presenting a concurrency-testing tool for Java based on our ranking scheme. Our tool tests for many different types of concurrency bugs at once.
Page 3
likely to violate assumptions of atomicity. CTrigger is primarily interested in atomicity-violation bugs and often overlooks other concurrency bugs.
Methodology 3.1 Ranking We propose a ranking scheme of test schedules, which is applicable to all programs irrespective of the synchroni- zation disciplines being followed, and targets all types of bugs. We rank the
Did you spot the banner on the top of the first page of this edition?
We are proud to say that the ACM IITD Chapter has done it again! We won the ACM Student Chapter Excellence Award during 2010-11 for Outstanding School Service among ACM student chapters from all over the world based on the review by an Essay Competition. The ACM Student Chapter Excellence Awards Program recognizes
chapters that display considerable initiative during the academic year. All student chapters compete for this award in five key areas. Winning chapters in each of these five areas receive $500 and an award banner to proudly display on their chapter's Web page. These chapters are featured on the ACM Web site and in ACM MemberNet
Outstanding Community Service: University of Michigan/Dearborn ACM Student Chapter Outstanding Recruitment Program: New Jersey Institute of Technology ACM Student
Chapter
Outstanding Chapter Website: Northeastern University ACM Student Chapter
Everything that can be counted does not necessarily count; everything that counts cannot necessarily be counted. ~Albert Einstein
Page 4
wireless sensor networks. In Proceedings of the 7th IEEE Intl. Conference on Distributed Computing in Sensor Systems (DCOSS '11), pp 1-8, June 2011.
A SysML Profile for Development and Early Validation of TLM 2.0 Models Vaibhav Jain, Anshul Kumar, Preeti Panda In Robert France et al., editors, Proc. 7th Int. Conf. Modelling Foundations and Applications (ECMFA'2011), pages 299-311. Springer, Berlin, LNCS 6698, 2011 A UML based framework for efficient Validation of TLM 2 Models Vaibhav Jain, Anshul Kumar, Preeti Panda Forum on Specification and Design Languages (*FDL'11*) Oldenburg, Germany, September, 2011 M. Agrawal, M. Hanmandlu and K.K. Biswas, Generalized Intuitionistic Fuzzy Soft Set and its Application in Practical Medical Diagnosis Problem, presented at Fuzz IEEE 2011conference, June 27 30, 2011, Taipei, pp. 2972 2978. Richa Sharma and K.K. Biswas, Using Courteous Logic based representation for Requirements Specification, 4th International Workshop on Managing Requirements Knowledge (MaRK11), Aug. 30,2011, Torino M. Agrawal, K.K.Biswas and M. Hanmandlu, Relations in Generalized Intuitionistic Fuzzy Soft Set, IEEE CIMSA, Computational Intelligence for Measurement Systems and Analysis, Sept 18-21, 2011, Ottawa. Paper entitled "Reducing Dependency on Middleware for Pull Based Services in LBS Systems" by Saroj Kaushik, Shivendra Kumar and Priti Goplani was presented by Prof Saroj Kausshik, in an International Conference on Wireless Communications and Applications (ICWCA 2011) held during 13 August 2011, Haikou, China. P. Goyal, V. J. Ribeiro, H. Saran, A. Kumar, Strap-Down Pedestrian DeadReckoning System, IPIN, Guimaraes, 2011. A. K. Somani and A. Gumaste, "Light
Trails: Distributed Optical Grooming for Emerging Data Center, Cloud Computing, and Enterprize Applications," in Proc. of OFC, 2011, Invited Paper, March, 7-10, 2011.
A. Gupta, J. Kumar, D. Mathew, S. Bansal, S. Banerjee, H. Saran, "Design and Implementation of the Workflow of an Academic Cloud," Databases in Networked Information Systems, University of Aizu, Japan, December 2011. S. Bindal, S. Bansal, "Systematic ranking of Thread Schedules for Testing Multithreaded Programs," ACM SRC PLDI Student Research Competition, San Jose, June 2011. Third Prize Winner in Undergraduate 2010, SciTePress, ISBN 978-989-8425-294, pp 165 170.
IV Awards
Sandeep Kumar Bindal won third prize at PLDI SRC 2011 in the Undergraduate Category for his work on Systematic Ranking of the Threads Schedules for Testing MultiThreaded Programs.
The Trafficdroid team of Kanav Goyal, Kushal Dudani, Saurabh Gupta, and Shikhar Gupta won third place in the Ericsson Application Awards for 2011, for their Android application on road traffic monitoring. Dr. Sorav Bansal won IBM SUR Award, 2011 Dr. Amit Kumar awarded INSA Medal for Young Scientist 2011
October
Nisheeth Vishnoi, Microsoft Research, Bangalore, 18th October 2011 Elisa Celis, University of Washington, Seattle, 18th October 2011 III. Publications Amitabha Bagchi, Adit Madan, Achal Premi. A computationally inexpensive and power efficient fully distributed topology for data collection in heterogeneous
V. Events
IIT Delhis 1st Knowledge Exchange Symposium (IKES 2011) organized by ACM Student Chapter in collaboration with IBM IRL , 16th October 2011 Hack U! organized by ACM Student Chapter and Yahoo, 13-15th October 2011
Page 5
Q.1. Please tell us a bit about your student days. Ans. 1. My student days were pretty boring, I would say. I was a good student who studied hard and got good marks. That pretty much sums up my student life. Life is so much more interesting now .
Q.4. It seems that you continued for your PhD after your masters. What are the advantages/disadvantages of the same? In your opinion is it better to get experience and come back or continue and complete it in one run? Ans. 4. This is a difficult question to answer. On the one hand, sometimes it does matter what your age is for certain kind of things. For example, in some places, you will not be considered young enough for assistant professor positions if you are over 35. And you will not be eligible to apply for certain kinds of funding, etc. So, it really depends on what you think your goals will be, how long you would like to get experience in a job and then come back. The trend seems to suggest that if you can get a PhD by the time you are in your early 30s, the world is open to you. If I could do it again, I would take up a job for a couple of years after my BTech or Masters. The truth is that going through a PhD program is like being in the real world you will have to deal with all kinds of people, some good and others nasty, and all kind of pressures, mostly bad. And its your degree at stake if you dont know how to navigate in the right manner. So, I would say, its better to go out and get some experience in the world and come back with a little more maturity.
Q.7. What made you want to come back to India? Ans.7. I got a good job! It was the right move to advance my career.
Q.2. How did you get interested in Computer Science as a career? Ans. 2. I was very young (about 10 or so) when my father bought a computer -- a commodore PC 40 with a color monitor. That was fascinating for me. I started reading books about computers and learning and programming in BASIC. And of course, playing a lot of computer games (primitive though they were). This continued for many years and finally, it was just a natural choice to choose Computer Science as my preferred subject.
Q.8. Your advice to the students. Ans.8. Always have a plan. It could be a 6-month plan or 1-year plan, but never do anything without first thinking about your goals. For example, if you are in your final year of BTech, think about what your plan is for the final year. Is your goal to get a job if so, what kind of company? Or go abroad for higher studies if so, what kind of university? Either goal requires a different kind of planning. If you go in without a reasonable idea of what your goals are, you are most likely wasting time. More likely, you suddenly find at the end of the final year that you actually want to join Intel, and now wish you had taken up a project in computer architecture. Eventually you will have to make longer-term plans and it is good to get into the habit early.
Compiled by : Aditi Kapoor
Q.3. What created a passion for Database Systems in you? Ans. 3. The initial interest in database systems came when I was studying concurrency control and recovery during my B.E. days. I could really see the beauty of the subject how the theory and practical aspects intertwined. And I had a natural aptitude for understanding the subject I had to read something once or twice and I
Q.5. What would you call the turning point of your career so far? Ans. 5. Going to MPII has been a major event in my career so far. I had the option of either going for a job in an industrial research lab or as a postdoc at MPII. I
It's not that I'm so smart, it's just that I stay with problems longer. ~Albert Einstein
Page 6
Page 7
Q.1. Please take us through your career journey Ans. 1. I did my B.Tech from IIT Kharagpur in 2002 after which I finished my MS and PhD from University of Illinois at Urbana-Champaign, IL(UIUC). Then I worked in Synopsis research for almost 6 months from where I moved to IBM systems and technologies lab in Bangalore by late 2007. I moved again to IBM Research Labs in Early 2009 and then joined here as assistant professor.
Q.2. Please share some enlightening experiences as a student at IIT Kharagpur Ans. 2. I couldnt say I enjoyed my time there. My father was a professor in Kharagpur and so I was a dayscholar After missing a few classes I found it difficult to latch on to what was being taught and so I had to get the help of my classmates. Surprisingly helping hands were usually those who were among non topper students of the class. But still it was difficult to cop in next lectures after missing one or two of some subject. This is something I try to take care here. Even if people miss classes, I try to put in some effort to give them the current coordinates
Q.7. Was there any incident in life that Q.4. How did you decide Computer greatly influenced you? Architecture as your field of specializaAns. 7. Back during my PG days, an Intion? dustry leader from India had visited Ans. 4. Well, it was not by choice. Back UIUC. He was the CEO of one of the in the UG years, our seniors used to adbiggest software companies in India. A vice that there were two main streams of few of us wanted to know why none of Computer Science: Theoretical Computer the Indian companies attempted to bring Science and Networking. Also, Theoretiout their own products. He told us that as cal Computer Science was for smart peolong as there was money from services, ple. At that point in life you have too there was no need to get into products. much faith in your seniors. While in many This answer from a supposed visionary situations they were more informed than shocked us. What people fail to recognize we were, a lot of what they knew was is that even companies outside the US half-baked. At that time there was a lot of have brought out products that have been prestige associated with RA-ship. Since huge successes. the networks group didnt have a RA slot, approached a professor who at the time was well funded in his projects. He told Q.8. What are the important qualities me to do well in a paper and then he for someone going into teaching and would consider me. I got an A in that, but by then I had another offer for RA-ship. research? So I approached the professor again and Ans. 8. One should be dedicated, sensible he straightaway accepted me. I asked him and should be persist in trying to solve about the project only then. He gave me a problems. Additionally, he should be tough name and I was bewildered. So I thick skinned to deal with the reviews, asked him to explain in layman terms and many of which would be peer reviews. I was very surprised that the field was One should always be ready to learn new Computer Architecture. After that there things. This is something I found different was a lot of learning and finally I did my in India from the US. Education in the PhD in the area. society is less. In the US people read books magazines etc when they travel but I hardly see anyone reading in Delhi Q.5. How was your industry experi- Metro. ence? Ans. 5. I had made up my mind during my PhD period that I was going to join academics. However I didn't want to join academics at that stage and so I joined the industry to make up my mind. The pay package of academia was low at the time.
Q.9. Message to students Ans. 9. Just try to come with the best in you. Students should focus on both productivity and quality in their work.
Q.3. How was the transition from UG to PG and PhD? Ans. 3. It was an enlightening experience culturally and academically. The competition was tough. There were
Compiled by :
Q.6. What pulls you back to India? Ans. 6. I had always wanted to return to
Page 8
We invite contributions for articles and Your contributions can be in the form of a also membership into the ACM student variety of articles covering any noteworthy chapter. This could be for being part of achievements, interesting projects, publica-
Till I realize, enjoys his aloneness he does, Comforting me to bear with mine
Contributed by : Aditi Kapoor
Page 11
B R O U G H T OU T B Y A C M STUDENT CHAPTER
SponsorM Balakrishnan ACM Student Chapter Sponsor
Newsletter Team in-charge: Magazine in-charge Aditi Kapoor (aditi@cse.iitd.ernet.in) Assisted by ACM chapter members Chair - Savin Goyal Abhishek Gupta Swati Verma Pulkit Yadav Aayush Goel (Photos in above order left to right)