28
“Ant Colony Optimization for Load Balancing of Nodes in Cloud.”

Ant Colony Optimization for Load Balancing in Cloud

Embed Size (px)

Citation preview

“Ant Colony Optimization for Load Balancing of

Nodes in Cloud.”

2

AgendaAbstractIntroductionAnt Colony OptimizationNatural behavior of antACO for load balancing in cloudLiterature surveyConclusion and Future scopeReferences

1

AbstractCloud computing refers to a parallel and distributed computing system

consisting of a collection of inter-connected and virtualized computers that are dynamically provisioned and presented as one or more unified computing resources based on service-level agreements (SLA) established through negotiation between the service provider and consumers.

Cloud load balancing is a type of load balancing  that is performed in cloud computing. Cloud load balancing is the process of distributing workloads across multiple computing resources.

Ant Colony Optimization is basic foraging behavior of an ant that encouraged them to find the optimal shortest path from their nest to food introduced by Dorigo and Gam-bardella

2

IntroductionWhat is Cloud Computing?

Internet-based on demand computing.Pay-as-you-goAccessing Computing resources of Third party

3

Characteristics of cloud computingLarge scale infrastructureVirtualizationHigh reliabilityUniversalityEasy scalabilityIn the form of demand requested service Low cost

4

Load Balancing in cloudLoad Balancing is a method to distribute workload across one or more

servers

Load balancing is used to make sure that none of your existing resources are idle while others are being utilized.

To balance load distribution, you can migrate the load from the heavy loaded nodes to the comparatively lightly loaded destination nodes.

Goals of Load balancing To improve the performance substantially To have a backup plan in case the system fails even partially To maintain the system stability 5

Ant Colony OptimizationThe inspiring source of ACO is the foraging behavior of real ants. When

searching for food, ants initially explore the area surrounding their nest in a random manner.

As soon as an ant finds a food source, it evaluates the quantity and the quality of the food and carries some of it back to the nest.

During the return trip, the ant deposits a chemical pheromone trail on the ground. The quantity of pheromone deposited, which may depend on the quantity and quality of the food, will guide other ants to the food source.

Indirect communication between the ants via pheromone trails enables them to find shortest paths between their nest and food sources. This characteristic of real ant colonies is exploited in ant colony optimization.

6

A Practical Scenario

There is a clear way….7

An obstacle has blocked the path.8

9

10

Con…Paths that have the highest pheromone intensity have the

shortest distance between the point and the best food source. The movements of these ants independently update a solution set.

The Traversal of ants in this system is generally of two types:

1) Forward movements-In this type of movement the ants move for extracting the food, or searching for the food sources.

2) Backward movements-In this type of movements the ants after picking up food from the food sources traverse back to the nest for storing their food. 1

1

ACO for load balancing in cloudThe ACO is used for load balancing. Ants Continuously originates from head node and traverse the width and

length of the network.These Ants along with their traversal will be updating a pheromone table.movement of ants in two ways similar to the classical ACO, which are as

follows:1) Forward movement-The ants continuously move in the forward direction

in the cloud encountering overloaded node or under loaded node.2) Backward movement-If an ant encounters an over-loaded node in its

movement when it has previously encountered an under loaded node then it will go backward to the under loaded node to check if the node is still under loaded or not and if it finds it still under loaded then it will redistribute the work to the under loaded node.

Main task is to redistribute the work among the nodes.Maintain a table for resource utilization. 1

2

Pheromone tablesSimilar to routing table of a network.The pheromone strengths are represented.Every node has a pheromone table for every possible destination in the

network, and each table has an entry for every neighbour.Their most frequent load type is denoted in the bracket (i.e. Overloaded-0,

Medium loaded-M, Under loaded-U) and simultaneously the calculated pheromone table which indicates the level of particular pheromone type (i.e. High-H, Low-L, Medium-M) between corresponding nodes.

13

Con….Here L,M,H shows the probability of moving.So the question is that probability in terms of what?Probability is based on pheromone concentration. The equation for the

moving probability is shown in next slide. (eq-2)And how the pheromone is calculated is based on eq-1.

13

What should be the pheromone??The value of pheromone here is,

τij(t=0)=f(MIPSJ,L,BWJ)………………eq-1

Pheromone value in between two node i and j at turn t=0, MIPSJ (Million Instructions per Second) is the maximum capacity of each processor of VMJ

The parameter BWJ is related to the communication bandwidth ability of the VMJ.

L is the delay cost is an estimate of penalty, which cloud service provider needs to pay to customer in the event of job finishing actual time being more than the dead-line advertised.

14

Con…..The ants traverse the cloud network, selecting nodes for their next step

through the classical formula given below, where the probability Pk of an ant,which is currently on node r selecting the neighboring nodes for traversal, is:

Pk(r, s) =[ τ(r, s)][η(r, s)]^β …………………….eq-2

[τ (r,u)][η(r, u)] ^β

where, r = Current node,

s = Next node,

τ= Pheromone concentration of the edge,

η = The desirability of the move for the ant (if the move is from an under loaded node to overloaded node or vice-versa the move will be highly desirable),

β= Depends upon the relevance of the pheromone con-centration with the move distance. 1

5

Pheromone UpdationThe ant will use two types of pheromone for its movement. The type of

pheromone being updated by the ant would signify the type of movements of the ant and would tell about the kind of node the ant is searching for (i.e.

overloaded or underloaded node). The two types of pheromones updated by the ants are as follows: 1) Foraging Pheromone (FP)

While moving from underloaded node to overloaded node, ant will update FP. Equation for updating FP pheromone is

FP( t+1 ) = ( 1 - βeva )FP(t) + ∆FP

Where,

βeva = Pheromone evaporation rate

FP = Foraging pheromone of the edge before the move

FP( t+1 ) = Foraging pheromone of the edge after the move

∆FP = Change in FP16

Con…. 2) Trailing Pheromone (TP)

While moving from overloaded node to underloaded node, ant will update TP. Equation for updating TP pheromone is

TP( t+1 ) = ( 1 - βeva )TP(t) + ∆TP

Where,

βeva = Pheromone evaporation rate

TP = Trailing pheromone of the edge before the move

TP( t+1 ) = Trailing pheromone of the edge after the move

∆TP = Change in TP

Since pheromones evaporate and diffuse away, the strength of the trail when it is encountered by another ant is a function of the original strength, and the time since the trail was laid. 1

7

Con…. βeva is pheromone evaporation factor which represents the pheromone volatilization

degree during unit time. Correspondingly, (1- βeva) represents the degree of residual pheromone. The value of βeva is between [0, 1].

The greater the value of βeva , pheromone evaporate faster, and past searches has a small influence on next step searching

18

OverloadedUnderloaded After originating from head nodes, ant move to node called as nextnode.. The

encountered nextnode status can either be overloaded or under loaded.Now this nextnode is became current node.Suppose, the load on

currentnode is greater than threshold i.e. status of currentnode is overloaded.

Now, ant will search for underloaded node among the neighboring nodes of the currentnode.

Here, ant can either get all overloaded neighbors or one underloaded node with minimum load. If ant get underloaded node then it will move to that node and update TP otherwise, ant will select the node which has minimum TP among neighbor nodes of currentnode and then move to that node and then update TP.

Now, the node on which ant is moved, is became a currentnode. Now, redistribution of load is done if and only if currentnode is

underloaded. Otherwise, it will again search for the underloaded node among neighbor nodes of currentnode.

19

Underloaded Overloaded Now suppose, the load on currentnode is less than threshold i.e. status of

currentnode is underloaded. Now, ant will search for overloaded node among the neighboringnodes of the currentnode.

Here, ant can either get all underloaded neighbors or one overloaded node with

maximum load. if ant get overloaded node then it will move to that node and update FP otherwise,

ant will select the node which has maximum FP among neighbor nodes of currentnode and then move to that node and then update FP.

Now, the node on which ant is moved, is became a currentnode. Now, redistribution of load is done if and only if currentnode is overloaded.

Otherwise, it will again search for the overloaded node among neighbor nodes of currentnode.

20

Flowchart

FP:Foraging PheromonesTP:Trailing Pheromones

Flowchart

Literature surveyPaper Description Pros ConsLoad Balancing of Nodes in Cloud Using Ant Colony OptimizationIEEE 2012 14th International Conference on Modeling and Simulation

Antscontinuously update a single result set rather than updatingtheir own result set

•System can continue functioning properly even at peak usage hours• It gives optimumsolution of load.• it is centralized no single point of failure

•.Network overhead because of the large number of ants•Points of initiation of ants and number of ants are not clea

A Technique Based on Ant Colony Optimization for Load Balancing in Cloud Data Center IEEE 2014 13th International Conference on Information Technolog

It uses redistribution policy and considers number of requests.

•high availability of resources•increasing the throughput•maximum resource utilization.

•It doesn’t consider server’s CPU power, memory etc

22

Con….Paper Description Pros ConsAn Ant Colony Based Load Balancing Strategy in Cloud Computing Springer 2014 Advanced Computing, Networking and Informatics - Volume 2

soft computing based algorithm on ant colony optimization has been proposed to initiate the load balancing under cloud computing architecture

•guarantees the QoS requirement

• Fault tolerance issues does not consider• All jobs are predicted with same priority here, which may not be the actual scenario.

Ant colony Optimization: A Solution of Load balancing in Cloud International Journal of Web & Semantic Technology (IJWesT) 2012

A heuristic algorithm based on ant colony optimization has been proposed

• The pheromone update mechanism has been proved as a efficient and effective tool to balance the load. This modification supports to minimize the make span of the cloud computing based services.

• This technique does not consider the fault tolerance issues.

23

ConclusionThis load balancing technique based on Ant Colony Optimization gives

optimal resource utilization. The performance of the system is enhanced with high availability of

resources, thereby increasing the throughput. This increase in throughput is due to the optimal utilization of resources

Future work is to implement this technique with the consideration of server’s CPU power, memory etc while redistributing load and also to do implementation of algorithm for finding out neighbors of a node with particular Data Center Network Architecture.

Future work

24

References: Nitin and Ravi Rastogi, Kumar Nishant, Pratik Sharma, Vishal Krishna,Chhavi

Gupta and Kunwar Pratap Singh,Load Balancing of Nodes in Cloud Using Ant Colony Optimization, IEEE 2012 14th International Conference on Modelling and Simulation

Ekta Gupta , Vidya Deshpande . A Technique Based on Ant Colony Optimization for Load Balancing in Cloud Data Center ,IEEE 14 13th International Conference on Information Technology.

Shagufta khan and Niresh Sharma, Ant Colony Optimization for Effective Load Balancing In Cloud Computing, ISSN 2278-6856 IJETTCS

Santanu Dam, Gopa Mandal,Kousik Dasgupta,and Paramartha Dutta.An Ant Colony Based Load Balancing Strategy in Cloud Computing. Springer International Publishing Switzerland 2014

Klaithem Al Nuaimi, Nader Mohamed, Mariam Al Nuaimi and Jameela Al-Jaroodi .A Survey of Load Balancing in Cloud Computing: Challenges and Algorithms. 2012 IEEE Second Symposium on Network Cloud Computing and Applications

Ratan Mishra and Anant Jaiswal Ant colony Optimization: A Solution of Load balancing in Cloud. International Journal of Web & Semantic Technology (IJWesT) Vol.3, No.2, April 2012.

https://en.wikipedia.org/wiki/Ant_colony_optimization_algorithms

25

THANK YOU!!!