作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2006, Vol. 32 ›› Issue (19): 143-145,. doi: 10.3969/j.issn.1000-3428.2006.19.052

• 网络与通信 • 上一篇    下一篇

基于时间序列的服务器负载预测

杨 伟1,2,朱巧明1,2,李培峰1,2,钱培德1,2   

  1. (1. 苏州大学计算机科学与技术学院,苏州 215006;2. 江苏省计算机信息处理技术重点实验室,苏州215006)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2006-10-05 发布日期:2006-10-05

Server Load Prediction Based on Time Series

YANG Wei1,2, ZHU Qiaoming1,2, LI Peifeng1,2, QIAN Peide1,2   

  1. (1. School of Computer Science & Technology, Soochow University, Suzhou 215006; 2. Key Lab of Computer Information Processing of Jiangsu Province, Suzhou 215006)
  • Received:1900-01-01 Revised:1900-01-01 Online:2006-10-05 Published:2006-10-05

摘要: 在负载均衡集群中,为了实现高效的动态负载均衡,需要对服务器的负载进行有效的预测。在介绍服务器负载特性和时间序列的3种模型的基础上,根据服务器负载动态变化的规律,提出了基于时间序列的服务器负载预测方法,并利用这一方法对IP网关服务器的负载进行了预测,实验结果证明该方法具有较好的实用效果。

关键词: 负载特性, 时间序列, 负载预测, ARMA

Abstract: In order to realize an effectively dynamic load balance in load balance cluster, it is necessary to predict server load accurately. This article introduces the properties of server load and three models of time series, then proposes the method of server load prediction, time series based on the regularity of load’s dynamic varieties, and applies this method to predict load of IP control gateway server. It is proved by the experiment that the predicting method has highly effectiveness in practice.

Key words: Load properties, Time series, Load prediction, ARMA