基于集群的Web服务器负载平衡算法研究

随着互联网访问量的激增,基于集群的Web服务器成为发展趋势,有效、公平的负载平衡直接决定系统的整体性能。建立BCMP排队预测模型,通过控制回路反馈动态调整预测因子,最终利用基于概率预测的PPLB算法平衡分配请求。仿真结果表明,该算法较传统负载平衡算法具有更高的效率和实用性。

Research on Load Balancing Algorithm of Web Server Based on Cluster

It is a crucial task to increase utilization of cluster web servers through effective and fair load balancing specifically when it comes to advent of dynamic contents and database-driven applications on the internet.First,established BCMP line forecasting model,then use a control loop feedback to adjust the predicted values periodically based on utility function.Last, this paper use PPLB algorithm to balance allocation request.The algorithm significantly improves both the throughput and mean response time in contrast to two existing load-balancing algorithms.