To search, Click below search items.

 

All Published Papers Search Service

Title

On Load Balancing Model for Cluster Computers

Author

Huajie Zhang

Citation

Vol. 8  No. 10  pp. 263-269

Abstract

Recently, cluster computers become a viable and less expensive alternative to multiprocessor systems. However, a serious difficulty in concurrent programming of a cluster computer system is how to deal with scheduling and load balancing of such a system which may consist of heterogeneous computers. Based on the four-tuple (dynamic load balancing model) which is proposed by other scholars, this paper introduces a five-tuple to better solve the above problems. The five-tuple involves balancing environment, task allotment, load estimate, scheduling strategy and scheduling evaluation. The five-tuple interprets the full aspects of load balancing and makes the logical relationship between the five tuples more clear by using formalized language. According to the five-tuple model and the corresponding solution of each tuple, this paper realizes the model under certain experimental conditions in the Linux cluster system. Experiments are given testing the five-tuple based on tasks that are independent, having a little or no relativity, uncertain executing time. The result shows that the model proposed in this paper can make the load balancing better. What¡¯s more, scheduling evaluation shows that the strategies are reasonable.

Keywords

Parallel, Cluster computers, Load balancing model, Load estimate

URL

http://paper.ijcsns.org/07_book/200810/20081040.pdf