A Load Balancing and Overload Controlling Architecture in Clouding Computing

摘要

The cloud computing provides better management and utilization of underlying infrastructures with resource virtualization. However, most of existing load balance models are unable to meet the special requirements of business such as the imbalance of peak period, the impact of nonlinear cross between varieties of business. Thus the reliability, scalability and self-organization of service becomes a significant challenge in cloud computing. The better load balancing can help to achieve more effective resource utilization by fairly allocating computing resources. In this paper, the load balancing method based on naive Bayesian model is proposed to maximize using system resources. A new business system architecture is designed and the system prototype is implemented with message oriented middleware, distributed coordinator technology and web services. The experiment results show that this architecture for load balancer performed well in overload controlling and fault-tolerance in cloud environments.