A Multi-agent Based Load Balancing Framework in Cloud Environment

摘要

Cloud computing is a commercial application of distributed system. With the rapid development of cloud computing, resource management problem has been widely studied and discussed. Most of the existing researches concern resource allocation by launching virtual machines, however, instant monitoring of the hosts, load balancing of the cloud platform is much more important for cloud providers. In this paper, a multi-Agent based distributed load balancing framework is proposed for cloud platform, it uses workload prediction technology and threshold buffering strategies to reduce the migration of VMs. Experiments show that our method can significantly reduce the workload data transmission and solve the frequent migration problem, which can effectively achieve load balancing, promote the utilization of the entire data center resources.