Research On Load Balancing Based On Cloud Computing

Cloud computing provides scalable and cheap distributed ability through the networkwhich changes the traditional delivery mode.Cloud computing is not a new product,it isdeveloped by traditional computer and network technology,just like distributed computing、parallel computing、utility computing、network storage、virtualization and load balancing.Virtualization is the core of cloud computing,through virtualization,server and storage in thenetwork was virtualized into a resource pool which is provided to customes.Google、amazon、Microsoft、IBM、China Telecom and other large enterprises are spending large sum of moneyfor the cloud computing research to promote the development of the cloud computing.Entering the21st century,the study of cloud computing has strong practical significance,at thesame time, it is the trend of The Times.In the cloud environment, large amount of tasks are requested by customers whice willcause some servers overload and some servers idle.Load balancing as one of the three majortechnical difficulties of cloud computing has an impact on the performance of cloudcomputing and the feeling of the service quality to the customers.The development time ofCloud load balancing is short, although hardware and intelligent algorithm are proposed tosolve this problem,there are still some rooms need to be improved.So how to improve theutilization rate of resources to achieve load balancing has become the research hot spot.Based on the detailed analysis and summary of the research background,significance ofcloud computing and introduces some key technologies of cloud computing,the followingworks about load balancing of cloud computing are done:(1)This paper analyzes and sums up the existing load balancing technologies and theirsadvantages and disadvantages. Finally choosing dynamic load balancing algorithm toimplement the load balancing.Based on this, this paper explains three prcesses to realize theload balance in the cloud environment:information acquisition and monitoring,data analysis,flexible resource scheduling.(2)This paper analysis the advantages and disadvantages of traditional genetic algorithm,improves the algorithm so that it will not get a local solution, then states the basic principlesof cellular automata,as cellular automaton is suitable for parallel computing,this papercombines genetic algorithm with cellular automata, proposes an improved genetic algorithm based on cellular automata(CGA) and applies the algorithm to the load balancing in cloudenvironment.(3)This paper uses Cloudsim as the simulation platform to simulate datacenter、virtualresources and user tasks under cloud environment,compares CGA to the improve geneticalgorithm in adapter four,the results prove that the proposed improved algorithm is better inperformance than genetic algorithm, it is a feasible load balancing algorithm for a cloudenvironment.