计算机与现代化

• 算法设计与分析 • 上一篇    下一篇

一种基于预测模型的负载均衡算法

  

  1. (东华大学信息科学与技术学院,上海201620)
  • 收稿日期:2014-05-07 出版日期:2014-08-15 发布日期:2014-08-19
  • 作者简介:乔国娟(1989-),女,江苏扬州人,东华大学信息科学与技术学院硕士研究生,研究方向:计算机网络; 陈光(1957-), 男,广东汕头人,教授,硕士,研究方向:计算机网络,网络通信。
  • 基金资助:
    上海市教育科学重点基金资助项目(A1119)

A Load Balance Algorithm Based on Prediction Model

  1. (College of Information Science and Technology, Donghua University, Shanghai 201620, China)
  • Received:2014-05-07 Online:2014-08-15 Published:2014-08-19

摘要:

结合集群系统中服务节点负载变化的不确定性和随机性的特点,提出WPLS-SVM组合的Linux服务器集群系统负载预测模型。该模型
能够依据采样值对集群中节点的负载情况做出预测,根据节点的负载预测结果结合随机区间调度算法可以有效地实现对请求任务的分配
与调度。仿真结果表明,较传统的负载均衡算法该算法克服了响应速度慢、预测精度差等缺点,提高了集群系统的利用率。

关键词: Linux虚拟服务器, 小波包分解, 最小二乘支持向量机, 负载均衡, 预测模型

Abstract:

 A load balancing model based on WPLS-SVM (Wavelet Packet and LS-SVM) of Linux server cluster system is
presented, using the uncertainty and randomness characteristics of load changing in cluster system. First, WPLS-
SVM prediction model is constructed and training sample of load is known. Then, the arriving requests are
distributed according to the results of load prediction by random interval scheduling algorithm. The simulation
results show that this model overcomes the slow response and poor prediction accuracy shortcomings and improves
the overall performance of the cluster system.

Key words: Linux virtual server, wavelet packet, LS-SVM, load balancing, prediction model

中图分类号: