基于缓存层级结构的多核Web服务器动态请求调度算法  

Cache hierarchy-based scheduling algorithm for dynamic requests in multi-core Web servers

作　　者：尤国华[1] 谭杨[1] 赵英[1] 

机构地区：[1]北京化工大学信息科学与技术学院信息中心,北京100029

出　　处：《计算机应用研究》2016年第1期87-91,共5页Application Research of Computers

基　　金：中央高校基本科研业务费资助项目(YZ1319)

摘　　要：针对Web服务器中传统的动态请求调度算法,如先到先服务算法等,不能充分利用多核处理器并行性的问题,根据多核Web服务器缓存结构的特点提出了基于缓存层级结构的多核Web服务器动态请求调度算法。该算法从处理器核心间的负载均衡和多核处理器的缓存层级结构特点两个方面对动态请求进行调度。基于此算法进行了仿真实验,实验结果表明该算法可保持多核处理器核心间的负载均衡,并能有效缓解多核Web服务器的乒乓效应。In order to solve the problem that traditional scheduling algorithms for dynamic requests, such as first come first served algorithm, couldn＇ t fully exploit the parallelism of multi-core processors, this paper proposed a cache hierarchy-based scheduling algorithm for dynamic requests in multi-core Web servers according to the characteristics of multi-core processors＇ cache hierarchy. The algorithm scheduled the dynamic requests on the basis of load balance between processing cores and characteristics of multi-core processors＇ cache hierarchy. The simulation experiment based on the algorithm shows that the pro- posed algorithm can keep load balance between processing cores and alleviate the ping-pong effect in multi-core Web server ef- fectively.

关 键 词：WEB服务器 多核处理器 调度 动态请求 缓存结构 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TP393[自动化与计算机技术—计算机科学与技术]