Research On The Load Balance Strategies Of Cloud Computing Federation Based On The Particle Swarm Optimization Algorithm

With the rapid development of cloud computing,the competition among Cloud Computing Service Providers(CCSP)is increasing.How to gain bigger business benefits for CCSP? Economically,the computing costs of CCSP will increase with the number of computing resources owned by CCSP,when the numbers of users and the numbers of user tasks are certain.However,CCSP's net income will decrease.During a certain period that users and tasks proliferated in the network,CCSP have to look for additional servers in order to satisfy users' demands.In order to avoid the waste of the servers in idler time,CCSP intends to form an Open Cloud Computing Federation(OCCF)to solve its own additional servers' requirements.But with the development of commercial cloud computing,in order to protect their own technological edge,CCSP set different standards for cloud computing.The cloud computing platforms are always heterogeneous in different CCSP.It's hardly impossible for applications to move and interoperate crossing platforms.So,it's difficult to accomplish OCCF.In order to solve these problems,this paper introduces the concept of complex network,compares the similarities between complex network and OCCF,indicates that OCCF has the characteristics of complex network,and proposes a new model of OCCF based the theory of regions,F-OCCF model.The F-OCCF model logically divides OCCF into four layers: application layer,middleware,virtual resource layer,physical resource layer.F-OCCF integrates the computing resources in the network,and provide computing service in the whole.Moreover,the F-OCCF model adopts the resource discovery mechanism,based on regions,which makes the resource discovery and sharing among resources more efficient.At the same time,the author did more researches on load balance of OCCF.This paper presents three classic load balance strategies: the strategy polling,the minimum number of connections,and the hash.It is found that these load balance strategies are not suitable for the OCCF because of its distributed structure and huge computing resources.According to the emergent properties of OCCF,the author selected Particle Swarm Optimization(PSO)algorithm to solve the load balance problem of OCCF.The necessity of load balancing for OCCF is demonstrated through experiments.According to the experimental results,the author found when the inertia weight decreased with the iteration increasing,the population size is 5000,c1=0.5,c2=0.7,the PSO algorithm in the load balance of OCCF got the good performance.After the load balance based on PSO algorithm,the average load rate of OCCF model is reduced,the global computing efficiency is significantly improved.