Research On Load Balancing Strategy Based On SLA Optimization In Cloud Computing

Huge commercial value and the market value is the real driving force of its development of cloud computing. Cloud computing service providers signed SLA (Service Level Agreement) with customer to provide high QoS (quality of service) and ensure their own interests. With the development of cloud computing technology, SLA has become the standard of QoS. As the large scale of cloud data center, the heterogeneous between servers and different task requirements, it will cause load imbalance between servers and the overall decline in performance of the whole system. The user’s QoS can’t be guaranteed, and the service providers also need to pay the penalty because of against SLA. Therefore, the research of optimizational load balancing strategy based on SLA on how to provide reliable service and ensure the rational use of resources and economic becomes the main content of this paper.Firstly, through the research on the load characteristics of data center and virtual machine migration characteristics in cloud environment, we found that the traditional load balancing strategies only reduce the single node and ignore the importance of evaluating the virtual machine migration based on security of SLA. According to the characteristics of the load in the cloud environment and the way of service provided, fees and costs, in this paper we propose a virtual machine selection strategy based on SLA profit. According to the benefit of virtual machine migration, to select the appropriate virtual machine to migrate, meet the needs of customers and increase the revenues of the service providers at the same time.Secondly, we introduce the AR model for the prediction of the load which can efficiently avoid the unnecessary migration caused by instantaneous peak and reduce the extra cost of system. We propose a new selection strategy of destination host which is suitable for heterogeneous servers, it can reduce energy consumption and operating costs of data center. And we present the mechanism of virtual machine migration limit based on probability forwarding thought, which can solve the critical mass and get better performance in energy consumption and load balancing.Finally, the experiment results show that our load balancing strategy based on SLA can improve the SLA violation rate and energy consumption of data center. Through the reasonable scheduling of virtual machine migration, it can ensure the users’ QoS and reduce the penalty efficiently, the proposed strategy can maximize the interests of all parties in the SLA.