摘要
近年来,服务器集群(ServerCluster)技术以其扩展性好、处理能力强、易于管理等优点受到了国内外研究机构的广泛关注和应用,其中的负载均衡算法更是研究的热点,但目前普遍缺乏算法的理论分析。该文对Pick-K、Pick-KX算法的机理进行了深入的分析,揭示了它们的内在本质,并提出了动态负载均衡算法的理论模型,为以后设计更好的均衡算法提供了理论上的支持,同时也可以从理论上定性地分析已有算法的效果。文章根据建立的理论模型,提出了一种新的负载均衡算法Pick-T,能够根据参数在更新周期内自动确定较好的分配向量,从而从整体上达到较好的负载均衡。
Making use of the server cluster to handle the requests which are increasing quickly is becoming more and more popular.Especially the load-balancing algorithm becomes the hotspot of the research.For the absence of theoretical analysis to algorithm,this paper suggests a theoretical model to provide support for designing better load-balancing algorithm on the base of Pick-K and Pick-KX.Further more,a theoretical model is made use of to design a new load-balancing algorithm---Pick-T.In the end,this algorithm is proved to reach better effects by theoretical analysis.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第25期78-80,99,共4页
Computer Engineering and Applications
