摘要
分析了多种负载均衡算法的优缺点及其适用场景,提出一种改进的周期自适应动态负载均衡算法.该算法将反馈负载信息的周期长度与网络中请求数量相关联,即当请求数量增多时,各服务器节点较多处于忙碌状态,系统会自动缩短反馈周期,保证了反馈信息的实时性;当请求数量减少时,各服务器节点较多处于闲置状态,系统会自动增加周期,有效地降低了计算各节点负载量的资源消耗.本文算法和动态加权轮询算法、静态加权轮询算法在JCF(Java Component Framework)中间件平台上的实验结果表明,该算法在增加集群系统的吞吐量和降低服务请求响应时间两方面有着显著效果,并在现实应用中有较大的价值.
This paper analyzed applicable scenarios,advantages and disadvantages of multiple load balancing algorithms,and proposed an improved cycle adaptive dynamic load balancing algorithm. The algorithm makes the cycle length of feedback load information vary with the number of requests through network,when requested quantity increases,each server node is almost in the busy state,and the system will automatically shorten the feedback cycle to ensure the real-time information feedback; conversely,each server node is almost in the idle state,and the system will automatically increase the feedback cycle to reduce resource consumption for calculating load of these nodes effectively. The results of experiments which this algorithm,dynamic weighted round-robin algorithm and static weighted round-robin algorithm are done in the JCF（ Java Component Framework） middleware platform showthat the algorithm has a significant effect in the respect of both increasing throughput and reducing service request response time about cluster system,and have greater value in real applications.
出处
《小型微型计算机系统》
CSCD
北大核心
2015年第7期1476-1480,共5页
Journal of Chinese Computer Systems
基金
国家科技支撑计划项目(2012BAH21F02)资助
2013年民航科技创新引导资金重大专项项目(MHRD 20130106)资助
中国民航大学预研重大项目(3122014P004)资助
中国民航大学中央高校基金项目(3122014C016)资助
