摘要
针对数据库集群负载问题,提出了一种动态负载均衡方法,并进一步设计、实现了包括CPU使用率、磁盘存储量、磁盘响应效率、网络延时、内存使用率等在内的多指标的节点负载测量和实时监控。该算法根据各节点的负载反馈信息进行任务分配,实现了负载均衡。性能分析和实验表明,该算法具有较高的负载均衡度和较低的系统开销。
For the database cluster load problem,a dynamic load balancing algorithm is given. Multi-node load balancing measurement and real-time monitoring is designed and implemented,such as CPU utilization,storage capacity,disk response efficiency,network delay,memory utilization and so on. Through load information of the nodes,the algorithm can get load mission groups,and attains load balancing. Performance analysis and simulation results show the algorithm has lower system cost and better load balancing factor.
出处
《微型机与应用》
2011年第2期68-71,78,共5页
Microcomputer & Its Applications
基金
国家自然科学基金(40801160)
国家自然科学基金(60902036)
国家863项目(2007AA120402)
