Anda di halaman 1dari 5

International Journal of Advance Foundation and Research in Computer (IJAFRC)

Volume 2, Issue 9, September - 2015. ISSN 2348 4853, Impact Factor 1.317

VMs Consolidation Using Ant Colony System And Clustering


Of PMs For Energy Efficient Network
Mr. Manoj V. Waghmode, Mr. Dayal B. Tilekar, Mr. Shubham D. Sonawane, Mr. Amol K. Sonavane
Imperial College of Engineering and Research
Pune,Maharashtra India
dayaltilekar@gmail.com, manojwaghmode44@gmail.com
ABSTRACT
In networking there is a big problem of energy consumption. So to improve the energy efficiency
and to reduce the energy consumption we can use distributed system architecture for
performing dynamic VM consolidation. This will help in reducing the energy consumption in
Network by maintaining the desired Quality of Services (QoS). We are planning to improve the
given system model using clustering of PMs and assigning them one consolidation manager to
each cluster. This manager manages the respective cluster which helps to monitor the all tasks
which are performed in that cluster of PMs. Due to monitoring the tasks we will be able to
improve the energy efficiency in a network.
Index Terms : Dynamic VM consolidation, VM migration, ant colony system, k-means clustering,
PMs clustering, energy efficiency.

I. INTRODUCTION
From last few years there are various ways to reduce the energy Consumption in Network. There are two
main techniques which are used that are Virtual Machine (VM) consolidation and dynamic server. We are
going to use Ant colony algorithm for virtual machine migration. In the field of computer science and
operation research field, Ant Colony Optimization (ACO) algorithm is used for finding shortest path
between various nodes. In this paper we are supposed to make the cluster of physical machines for
efficiently migration of virtual machine to other physical machine and assigning the monitor to
respective cluster for observing all the physical machine tasks and load balancing in network.
For clustering of physical machine we are using k-means clustering algorithm. K-means is the simple
unsupervised learning algorithms that is used to solve the clustering problem. K-means follows a simple
and easy way for classification of given data set through a certain number of clusters.

Explanation:
In Network large numbers of physical machine are in working state. In each physical machine some
virtual machines performing their task, for that VMs uses required resources from PMs. If VMs perform
another task and resources of that task are available in another PMs also for accessing those resources
VMs get migrate from one to other PMs.
For migration of VMs before it we should find proper physical machine from network. To find that PM it
consume more energy so to reduce that we are going to create the cluster of PMs. So due to that cluster of
PM finding energy of VM get reduce.
A. Ant colony Algorithm
The ant colony algorithm is used to find the optimal path that is based on the behavior of ants that are
going in search for food.
Initially, the ants search the food randomly. When an ant identifies a source of food, it goes back to its
colony leaving behind the "markers" (pheromones) that shows the path of the food source. When other
29 | 2015, IJAFRC All Rights Reserved

www.ijafrc.org

International Journal of Advance Foundation and Research in Computer (IJAFRC)


Volume 2, Issue 9, September - 2015. ISSN 2348 4853, Impact Factor 1.317

ants come across the markers, they are also assumed to follow the path with a certain probability. If they
do, then they populate the path with their own markers while bringing the food back to the colony. When
more ants find the path, the markers get stronger until there are a couple of lines of ants traveling to
different food sources that are near to the colony. This ant colony system algorithm is used for Virtual
machine migration.
B. Algorithm:
/* Initialization phase */
For each pair (r,s) (r,s):= End-for
For k:=1 to m do
Let
be the starting node for ant k
( ):= {1, ..., n} /* ( ) is the set of yet to be visited nodes for ant k in node
*/
:=
/* is the node where ant k is located */
End-for
2. /* This is the phase in which ants build their tours. The tour of ant k is stored in Tour k . */
For i:=1 to n do
If i<n
Then
For k:=1 to m do
Choose the next node according to Eq. (3) and Eq. (1)
( ):= ( )
(i):=( , )
End-for
Else
For k:=1 to m do
/* In this cycle all the ants go back to the initial node
*/
:=
(i):=( , )
End-for
End-if
/* In this phase local updating occurs and pheromone is updated using Eq. (5)*/
For k:=1 to m do
( , ):=(1-) ( , ) +
:=
/* New node for ant k */
End-for
End-for
3. /* In this phase global updating occurs and pheromone is updated */
For k:=1 to m do
Compute
/* is the length of the tour done by ant k*/
End-for
Compute
/* Update edges belonging to
using Eq. (4) */
For each edge (r,s)
( , ):=(1-)( , )+
End-for
4. If (End_condition = True)
then Print shortest of
else goto Phase 2
C. Clustering algorithm:
For clustering physical machine we are using k-means clustering algorithm.
30 | 2015, IJAFRC All Rights Reserved

www.ijafrc.org

International Journal of Advance Foundation and Research in Computer (IJAFRC)


Volume 2, Issue 9, September - 2015. ISSN 2348 4853, Impact Factor 1.317

K-means clustering is a method used for vector quantization, originally from signal processing, that is
popular for cluster analysis in data mining. K-means clustering partitions n observations into k clusters
in which each observation belongs to the cluster with the nearest mean serving as a prototype of the
cluster.
Algorithm steps:
Let X = {x1,x2,x3,..xn} be the data points set and V = {v1,v2,.,vc} be the set of centers.
1) Randomly select any c cluster centers.
2) Then calculate the distance between each data point and cluster centers.
3) Assign the data point to the cluster center whose distance from the cluster center is the
minimum of all the cluster centers.
4) Recalculate the new cluster center using:

Where ci represents the number of data points in ith cluster.


5) Recalculate the distance between each data point and new obtained cluster centers.
6) If no data point was reassigned then stop, otherwise repeat from step 3).

II. Architecture Diagram:

In above fig. Each PMs contains a CPU with multi-core. Basically CPU performance is in Millions of
Instructions Per Second (MIPS). Here, we Are using Ant colony alogorithm for VMs Migration. If one of
the PM gets overloaded , then we can migrate VM to the under load PM. To find that PM it consume more
energy so to reduce that we are going to create the cluster of PMs. So due to that cluster of PM finding
energy of VM get reduce. Here, we are going to migrate VM with the nearest PM by using Ant colony
algorithm. We are supposed to make the cluster of physical machines for efficiently migration of virtual
machine to other physical machine and assigning the monitor to respective cluster for observing all the
physical machine tasks and load balancing in network.
III. Flow Diagram:

31 | 2015, IJAFRC All Rights Reserved

www.ijafrc.org

International Journal of Advance Foundation and Research in Computer (IJAFRC)


Volume 2, Issue 9, September - 2015. ISSN 2348 4853, Impact Factor 1.317

Fig: Migration of Virtual Machine


In above figure, it consists of two modules:
A. Framework of VM migration
In above figure it is the migration of multiple virtual machines. It consists of four main modules:
Migration decision maker, Migration controller, Resource reservation controller and Resource monitor.
1. Migration Decision Maker:
It is responsible for making effective migration decision.
2. Migration Controller:
It control real migration process and it also choice right target machine.
3. Resource reservation controller:
It implements different resource from source and target machine. It also avoids migration failure.
4. Resource Monitor:
It can monitor resource both VM and PM. It used for workload stability for avoiding migration
threshing.

B. Resource reservation method:


In this method there are two parts
1. Source machine: Source has CPU and Memory

32 | 2015, IJAFRC All Rights Reserved

www.ijafrc.org

International Journal of Advance Foundation and Research in Computer (IJAFRC)


Volume 2, Issue 9, September - 2015. ISSN 2348 4853, Impact Factor 1.317

a. CPU: Resource reservation of source machine in xen virtualization platform by adjust CAP
b. Memory: Implemented dynamic adjust virtual machine memory

2. Target machine: If not enough resource for reserved virtual machine, the physical machine remove
target machine list.
IV. Conclusion:
Hence we are going to propose a system model by clustering PMs and assigning them a respective
consolidation manager. By using Ant colony algorithm we are migrating VMs. This manager manages the
respective cluster which helps to monitor the all tasks which are performed in that cluster of PMs. Due to
monitoring the tasks we will be able to improve the energy efficiency in a network.
V. REFERENCES
[1]

C. Clark, K. Fraser, S. Hand, J. G. Hansen, E. Jul, C. Limpach, I. Pratt, and A. Wareld, Live
migration of virtual machines, in Proc. 2nd Conf. Symp. Netw. Syst. Design Implementation,
2005, vol. 2, pp. 273286.

[2]

M. Dorigo and L. Gambardella, Ant colony system: A cooperative learning approach to the
traveling salesman problem, IEEE Trans. Evolutionary Comput., vol. 1, no. 1, pp. 5366, Apr.
1997.

[3]

F. Farahnakian, T. Pahikkala, P. Liljeberg, and J. Plosila, Energy aware consolidation algorithm


based on K-nearest neighbor regression for cloud data centers, in Proc. IEEE/ACM 6th Int. Conf.
Utility Cloud Comput., Dec. 2013, pp. 256259.

[4]

T. Cover and P. Hart, Nearest neighbor pattern classication, IEEE Trans. Inf. Theory, vol. IT-13,
no. 1, pp. 2127, Jan. 1967.

[5]

M. Mishra and A. Sahoo, On theory of VM placement: Anomalies in existing methodologies and


their mitigation using a novel vector based approach, in Proc. IEEE Int. Conf. Cloud Computing,
Jul. 2011, pp. 275282.

[6]

L. Deboosere , B. Vankeirsbilck, P. Simoens, F. Turck, B. Dhoedt, and P. Demeester, Efficient


resource management for virtual desktop cloud computing, J. Supercomput., vol. 62, no. 2, pp.
741767, 2012.

[7]

E. Feller, C. Morin, and A. Esnault, A case for fully decentralized dynamic VM consolidation in
cloud , in Proc. IEEE 4th Int. Conf. Cloud Comput. Technol. Sci., Dec. 2012, pp. 2633.

[8]

A. Beloglazov and R. Buyya, Optimal online deterministic algorithms and adaptive heuristics
for energy and performance efficient dynamic consolidation of virtual machines in cloud data
centers Concurrency Comput.: Prac. Exp., vol. 24, no. 13, pp. 13971420, 2012.

[9]

A. Ashraf and I. Porres, Using ant colony system to consolidate multiple web applications in a
cloud environment, in Proc. 22nd Euromicro Int. Conf. Parallel, Distrib. Netw.-Based Process.,
2014, pp. 482489.

[10]

T. Cover and P. Hart, Nearest neighbor pattern classification, IEEE Trans. Inf. Theory, vol. IT13, no. 1, pp. 2127, Jan. 1967.

33 | 2015, IJAFRC All Rights Reserved

www.ijafrc.org

Anda mungkin juga menyukai