作 者: ;
机构地区: 广东商学院信息学院
出 处: 《现代计算机》 2005年第10期20-22,共3页
摘 要: 传统的请求分配算法作为一种集中式的静态分配负载均衡算法,一直被认为存在分配不准确的缺陷[1],主要有两个原因:一是集中调度容易成为系统的连接瓶颈,二是静态分配时对各处理机的动态性能把握不够准确。本文针对这两个问题采用自适应的集中调度,使之具有可扩展性,并且严格根据各处理机的实时性能参数实施静态分配,使之适用于异构的服务器集群。 Traditional request dispatching algorithm as a static centralized load-balancing algorithm has been considered existing defects of inaccurate dispatching. There are two main reasons: one is that centralized scheduling is trend to be a bottleneck and the other is that static dispatching cannot predefine accurately those dynamic performances of the Web server clusters. The proposed scheduling strategy concentrates Oll how to solve these two problems. In this paper, the centralized scheduling can be extendable and adaptive and the static dispatching will work strictly according with the real-time performance parameters of the Web-Server clusters so that this algorithm can fit for heterogeneous systems.