Research And Design The Load Balancing Algorithm In Cloud Computing

Cloud computing is a new model of business and computing. It changes the network service model, as well as the way people using network and computers. From the interests of customers, service provider of cloud computing has to provide services to meet the customers’needs. As to cloud technology, cloud computing is a kind of computing model which supplies the storage resources, computing resources and software services to customers by Internet. So that we can separate the management of users and computing resources.Based on the analysis of load balancing technology and algorithms on cloud computing and cluster systems, this paper proposes two novel algorithms respectively for private clouds and public clouds, improving the performance of both centralized and distributed load balancing algorithms. The details include the following two aspects:On the one hand, this paper first compares the performance of the existing centralized algorithms, including swarm algorithm, polling algorithm and minimum-priority strategies, and then proposes a novel algorithm for the private cloud scenario, which based on the combination of the rotation algorithm and minimum priority strategies, to improve the performance of the current algorithm. Experiments are designed to evaluate the effectiveness of the novel algorithm. There are two innovations:one is the introduction of the concept of sequencing and update in the polling cycle, and the other is the two load balancer load tables proposed to avoid inconsistencies.On the other hand, based on the analysis of current distributed load balancing algorithms, the paper proposes an improved random walk algorithm for public cloud environment, which is validated for its efficiency compared with the current algorithm. Innovation includes two aspects:the first is the introduction of the concept of average node number of free resources in the random walk algorithm, which speed the location of the transfer node and balancing action. And the second is the geographical factors proposed in the random walk algorithm for cluster grouping, which bring balance within groups and the interaction between groups, with priority for inter group.At the end, the paper concludes with summarization of the contribution of the work and the on-going work to be carried out.