期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2019
卷号：19
期号：8
页码：8-13
出版社：International Journal of Computer Science and Network Security
摘要：Load balancing is the efficient distribution of incoming client requests in a server pool or virtual machines. For example, it allows high-audience website servers not to be overloaded. One of the challenging scheduling problems in Cloud load balancing is the choice of Virtual Machines (VM) when assigning tasks or migrating VM. This paper proposes a load balancing strategy for distributed cloud data center to be applied in two levels control: Physicals Machines and Clusters. Load balancing is done by two managers, they ensure exchange information and decide afterwards the level concerned with load balancing. Measurement of load is based on load information, including CPU utilization and memory utilization. Virtual Machines are allowed to be migrated between different federations to distribute loads while the communication costs are also incurred. Therefore, the objectives of this strategy are twofold: reducing the load of the overloaded physicals machines and decreasing the communication costs among different federations. The proposed method attempts not only giving a good Cloud balancing but also ensures reducing response time and communication cost and enhancing performance of the whole system.
