摘要
针对目前web集群的研究现状,从数据挖掘的角度提出一种web集群的负载均衡方法,它首先基于事务对web集群的访问进行文档聚类,形成一些聚类数据。将这些聚类按时间序列预测其负载强度,然后,用NSPP算法动态分配到集群系统各真实服务器上,使整个系统负载均衡。实验与分析结果表明了该方法的有效性。
This paper aimes at the present research about web cluster,an approach of load-balance of web cluster are introduced.The approach can be divided into two steps.Firstly,based on the access to web cluster,we can aggregate the document and get some collective data.According to these data,we can predict the intensity of load balancing based on time series models?Secondly,the collective data can be assigned to individual server in the web cluster by NSPP Algorithm to ensure the load balance of the whole system.Experiment and analysis prove its effective.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第25期59-61,共3页
Computer Engineering and Applications
基金
湖南省自然科学基金资助(编号:02JJY2097)
关键词
聚类
时间序列
负载均衡
cluster,time series model,load balance
