An Improved LVS Dynamic Load Balancing Algorithm

The main goal of this thesis is to improve the performance of Linux Virtual Server (LVS) clusters, increase system's response speed and Enhance its stability. The main result and contribution are as follows:First, the cluster and load balancing technologies are introduced, the theory and implementation mechanism of LVS are explained, the structure of nuclear software IPVS in LVS and the implementation process of each module are analyzed, its weakness is also pointed out.Second, a J-Load Balancing algorithm is presented. The real-time parameters of the performance of each real server is timely gathered by a cross-platform proxy procedure, and then is analyzed by load-balancer. The load-balancer chooses the most suited server and then transmits the request to it. At last, the required result is returned to client.Experiments show that the algorithm can balance the load of Web server clusters effectively. Results suggest that this algorithm is better than traditional algorithms in performance.