摘要
随着Internet的发展,路由器成为了网络性能的瓶颈。路由算法的效率和鲁棒性成为网络研究领域的热点之一。分布式系统采用并行运行,避免了单点故障。本文提出的分布式路由器使用IP做为任务分配粒度,利用Hash算法实现负载均衡。改进了基于心跳和检测点的故障检测机制,在较低的系统开销下缩短了系统检测的响应时间。仿真试验的结果表明,该机制可行且高效。
With the development of Internet, routers are becoming the bottleneck of networks. The performance and robustness of routers are important research areas in network. Distributed system can provide the parallel execution and avoid the single point of failure. In this paper, Distributed Routers （DR） uses IP as the task assignment granularity and Hash algorithm to implement the load balance. The fault detection based on the heartbeat mechanism and checkpoint is improved to decrease the detection response time under the low system expense. The results of simulation show that DR is an effective parallel router scheme in distributed environment.
出处
《计算机科学》
CSCD
北大核心
2006年第9期40-41,44,共3页
Computer Science
基金
四川省青年软件创新工程资助项目
编号:2004AA0354。
关键词
分布式
路由
负载均衡
容错
Distributed system, Router, Load balance, Fault tolerance
