首页> 外文会议>Informatics, 2009. BCI '09 >Load Balancing in a Cluster-Based P2P System
【24h】

Load Balancing in a Cluster-Based P2P System

机译:基于群集的P2P系统中的负载平衡

获取原文

摘要

In this paper we study load balancing in a cluster based p2p system. The cluster based architecture is chosen after the study of various real systems. We then define the notion of load for that system and describe coefficients that measure the different parameters of load. We use these coefficients to define load factors that can be used to measure each nodeȁ9;s load. These factors are also used in two algorithms that balance the load fairly among all peers in a cluster. We improve the algorithmsȁ9; balancing performance using replication of popular objects. A set of metrics is used to evaluate the cost of each algorithm and the effect on systemȁ9;s performance. Our simulation results show that it is possible to improve the fairness on load distribution without replication but with a small reduce on systemȁ9;s throughput. And under the cost of replicating a small amount of objects we improve fairness without reducing systemȁ9;s throughput.
机译:在本文中,我们研究了基于集群的p2p系统中的负载平衡。在研究了各种实际系统之后,才选择基于集群的体系结构。然后,我们定义该系统的负载概念,并描述测量不同负载参数的系数。我们使用这些系数来定义可用于测量每个节点9的负载的负载因子。在两个算法中也使用了这些因素,它们在群集中的所有对等方之间公平地平衡了负载。我们改进算法ȁ9;通过复制流行对象来平衡性能。一组度量标准用于评估每种算法的成本以及对系统性能的影响。我们的仿真结果表明,无需复制就可以提高负载分配的公平性,而对系统的吞吐量则有少量降低。并且以复制少量对象为代价,我们提高了公平性,而又不降低系统的吞吐量; 9。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号