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 ›› 2007, Vol. 33 ›› Issue (04): 97-99. doi: 10.3969/j.issn.1000-3428.2007.04.033

• 网络与通信 • 上一篇    下一篇

增强型动态反馈随机分发负载均衡算法

张 昊,廖建新,朱晓民   

  1. (北京邮电大学网络与交换技术国家重点实验室,北京 100876)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-02-20 发布日期:2007-02-20

Advanced Dynamic Feedback and Random Dispatch Load-balance Algorithm

ZHANG Hao, LIAO Jianxin, ZHU Xiaomin   

  1. (State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications, Beijing 100876)
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-02-20 Published:2007-02-20

摘要: 针对移动定位平台集群服务器的系统特征,该文提出了一种新的负载均衡算法。该算法采用周期动态反馈方法,并将节点接纳新请求所带来的增量引入到对节点负载的预测中去,根据节点剩余处理能力动态调整节点分配的概率。在实验室搭建的移动定位平台上,对本算法以及Pick-KX和DFB (Dynamic Feedback)两种负载均衡算法在异构环境中的系统吞吐量、系统应答时间进行了对比测试,结果显示在高负载情况下该算法优于另外两种算法。

关键词: 移动定位平台, 负载均衡算法, 异构集群

Abstract: Based on the system architecture and software structure of MPP (mobile position platform), a new dynamic load-balance algorithm is proposed. It bases on dynamic feedback and imports the increment for admitting new request into the load forecast. It dynamically adjusts the dispatching probability according to the remainder process capability of each node. Experiments on the performance of algorithm are taken in MPP in laboratory and the algorithm is compared with Pick-KX algorithm and DFB (Dynamic FeedBack) algorithm in average throughput and average response time. Results of experiments show that the algorithm is better than the other two algorithms under high system loading condition.

Key words: Mobile position platform, Load-balance algorithm, Heterogeneous cluster