ReDAL: An Efficient and Practical Request Distribution Technique for Application Server Clusters
IEEE Transactions on Parallel and Distributed Systems
To balance or unbalance load in size-interval task allocation
Probability in the Engineering and Informational Sciences
| Hi-index | 0.00 |
Web server is recently composed of a server clusterfor keeping good quality of services in spite ofsuddenly rushed connections. LVS (Linux VirtualServer) is one of methods to build server cluster inLinux and it has its own load balancing mechanism[1]. The existing load balancing mechanism, however,reaches the limit often and fails to keep load inbalance, in particularly electric commerce sites thatuse typically dynamic script files more than staticHTML pages. In this paper, we suggest a newmechanism that overcome the limit and discuss howto define the load of each real server node in servercluster and determine the limit of the load. In thesuggested mechanism, the weight table of a loadbalancer is updated according to the variation of theload and the distribution of service requests iscontrolled according to the table. The new mechanismshows better results to distribute the load evenlyamong the real servers than the existing mechanism,when many scripts are especially included.