基于排队理论的自适应负载均衡算法研究

负载均衡集群技术中的负载均衡算法是整个互联网络应用研究中一个重要问题。本文针对“需求快速变化”这一企业级管理软件开发中呈现的特点，以及各应用软件之间的互操作性差、应用集成水平较低，无法实现信息的共享与交流这一问题，提出一种基于排队理论的负载均衡调度算法。该算法根据当前的负载情况和服务器的剩余处理能力，计算出服务器在随后的固定时间间隔中所需要的平均到达率，进而得到任务分配策略，从而得到较为优秀的负载均衡效果。最后通过仿真实验和对比分析，证明了该算法具有较好性能。

Research for load balance algorithms based on queue theory

Load balance algorithm is important in the research of network applications. Aiming at characteristic of requirement changing rapidly in enterprise management software. Also, between software lack of interactive, and the level of application integration is low, those reason lead to sharing and communicating information difficult. To solve this problem, this paper introduces one load balance algorithm based on queuing theory. This load balance algorithm based on queuing theory, according to the present load situation and the rest processing ability, compute the need average arrival frequency of server in fixed time intervals, and then get the task allocation strategy. Also a simulative experiment system was developed.