中间应用服务器动态负载均衡的物理模型  

Physical Models of Dynamic Load Balancing among Application Servers

作　　者：陈志刚[1] 曾志文[1] 

机构地区：[1]中南工业大学信息工程学院,长沙410083

出　　处：《计算机工程》2001年第1期44-45,64,共3页Computer Engineering

基　　金：教育部"科学技术研究重点计划项目";高等学校骨干教师资格计划资助项目

摘　　要：由于客户机访问服务器的随机性，使得各中间应用服务器所接收到的任务出现差异，在某些时刻，可能有的中间应用服务器负载过重，而同时有的中间应用服务器却很空闲或完全闲置．此时，为充分利用网络硬件资源及缩短客户等待时间，可将负载过重的中间应用服务器的一部分任务转移给较空闲的中间应用服务器，从而使各中间应用服务器的负载大致均衡．为此，该文通过引入两个物理模型及相应算法解决这一问题，并为解决类似问题开辟了一条新的途径．Because of the randomicity when client calls server, there is difference between tasks to be received by every middle application server. sometime, some application servers will load to much while some ones else will be very idle or even no task. In order to fully utilizing the network hardware resources and shortening the clients' waiting time, we can transfer a part of tasks of application server whose loads are too heavy to those which are idler it is a good means that every application server's load will be approximate balance. This article uses two physical models and homologous arithmetic to solve this problem and inaugurates a new approach which can be used to solve similar problems.

关 键 词：中间应用服务器 动态负载均衡 物理模型 水库调度 

分 类 号：TP368.5[自动化与计算机技术—计算机系统结构] TV697.11[自动化与计算机技术—计算机科学与技术]