The Strategy Research On The Load Balancing Based On Web Service

Web service is an independent programming language, computing platforms and networking protocols development standards. It shields the distributed complexities for application developers by using advanced program abstract. The main function of Load balancing is to balance the system resources among servers, in order to optimize the over all system performance. Load balancing technology can be integrated into middleware, which can be used in heterogeneous systems with access transparency and extensible capability. In the middle of the module can achieve a variety of load balancing algorithms flexibly, in order to meet different load balancing demands of applications, and make the systems become more scalable, configurable, and robust.In this paper, we designs a load balancer: WSBLB (Web Service Based Load Balancer), and actualizes dynamical load balancing scheduling policy according to the models of distributed systems. In order to describe system load more accurately, we choose the CPU utilization, memory utilization, system response time, input and output streams and the process number as the parameters to measure the system load.In order to avoid increasing pressure on WSBLB and causing single weak point in system resulting from monitoring of server status, this strategy will make the load information collection program run on the server nodes which are in cluster. when the value of the load difference from two interfacing acquisition cycle exceeds the standard, the collection program will call the Web Services on the balancer, reporting its own load status. WSBLB run the synchronized operation according to the information collected from all the loads, and replace the service sequence. This approach has considerably reduced the additional connection and waste of resources because of the measure changes of load.Performance parameters of servers in different time period is made as the measure of tracking system load and a parameter setting for each factor describes the sighificant factors in different time periods that affect the system load.Thus avoid the wrong effect on the date caused by single factor,so that a system performance is optimized. On the whole, the system load changes more smoothly.In order to verify the effect of the improved strategy, we designed a test program. And three components, client, load balancer, server, are included. Deploy a simple Web services on the server, including page views, upload files and download files and simple calculation. Run test tools on the client to simulate a large number of requests, and the balancer serves for the middle of the task scheduling according to different load-balancing strategy. According to the data, analyse the impact of the load-balancing strategy on system performance.Through the experiments and comparison among some widely used load balancing strategies, such as LARD, it proves'that this improved'method can effectively improve the load balancing capabilities, and it is feasible in heterogeneous and different regions of the cluster system.