基于随机高级Petri网模型的服务器均衡集群  

An Approach to Servers Balanced Cluster Based on Stochastic High-Level Petri Net Model

作　　者：黄光球[1] 刘兆明[2] 

机构地区：[1]西安西安建筑科技大学管理学院,华为技术有限公司710055 [2]华为技术有限公司

出　　处：《微计算机信息》2006年第05X期134-136,共3页Control & Automation

基　　金：国家863高科技发展计划资助项目(2004aa1z2460)

摘　　要：在均衡集群中请求的分配和选择服务是影响服务器集群性能的关键。本文基于随机高级Petri网(SHLPN)模型提出了动态反馈请求负载分配算法和加权队列选择的综合均衡调度方案。根据集群中各服务器实体的实时负载状况动态分配请求并结合请求权值实施服务,提高了系统的负载均衡能力。It is key to influence servers clustering performance for allocation of requests and selection of services in duster of oad balancing. This paper proposes a dynamic feedback requesting load assignment algorithm and a synthetically balanced dispatch policy from weighting queues selecting based on the stochastic high level Petri net model （SHLPN）. Requests are dynamically allocated among servers according to real-time load condition of servers; implementation of services is unified according to weighted values of requests. The techniques have greatly improved load balancing ability of system.

关 键 词：随机高级Petri网(SHLPN) 动态反馈 服务器集群 请求分配 服务选择 

分 类 号：TP368.5[自动化与计算机技术—计算机系统结构]