摘要
负载均衡是提高分布式系统性能的重要技术,同时也是系统高可用性、可扩展性、冗余性的必然要求。针对分布式系统任务调度不均衡问题,在分析和建立系统仿真和任务调度模型的基础上,提出了一种基于公平指标的任务调度负载均衡算法,推导出在多节点条件下的任务分配方法,并在此模型下改进了基于公平指标的负载均衡算法。最后,在Linux平台下,进行了仿真实验和性能比较。实验结果表明,该算法是有效的,它可以有效地提高分布式系统的性能和效率。
Load-balancing is an important technology for improving performance of distributed system.It is also requirement of high-availability,expansibility and redundancy in distributed system.In order to solve the problem of unbalancing task scheduling in distributed system,a model of simulation system and task scheduling is given,the fair scheduling algorithm（FSA） based on fair index is also drawn.The method of task distribute in multi-nodes is deduced and the algorithm of load-balancing based on fair index is improved.From the simulation experiment in Linux platform,the conclusions are drawn that it is reasonable and feasible of FSA can improve the performance and efficiency of distributed system.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第6期1362-1363,1400,共3页
Computer Engineering and Design
基金
广东省自然科学基金项目(05011896)
茂名学院科学研究资金项目(203282)
关键词
任务调度
公平指标
动态分配
负载均衡
分布式系统
task scheduling
fair standard
dynamic distribute
load balancing
distributed system
