IP网络的快速故障恢的复.pdf

IP 网络的快速故障恢复 张民贵, 刘 斌 （清华大学计算机科学与技术系, 北京, 100084 ） 摘 要: 随着互联网的迅速发展,人类通信对其依赖性日益增强,而 IP 层故障恢复能力低下,阻碍了互联网 性能的提高.近些年来,国际学术界对IP 网络快速故障恢复的方案研究异常活跃,提出了加快故障恢复速度的三 条途径:(1)加快 IP 路由收敛; (2)使用主动式故障恢复; (3)提高故障检测的速度与准确性.针对已有解决方案的 不足, 本文得出,要推动 IP 网络的快速故障恢复方案的实现,必须做好: (1)故障后的通信负载均衡;(2)互操作测 试及路由器体系结构的重新设计. 关键词: IP 网络;故障恢复;路由收敛;备份路径;负载均衡 中图分类号: TP393.05 文献标识码: A 文章编号: 0372-2112 （2008 ）071424 Fast Failure Recovery of IP Networks ZHANG Min-gui ，LIU Bin （Department of Computer Science and Technology, Tsinghua University, Beijing, 100084, China ） Abstract: Internet has become an essential infrastructure of communication. Limited survivability in the face of failure impairs the performance of Internet. Recent years, fast failure recovery of IP network has drawn much attention. As the state-of-art, there three ways to realize fast failure recovery of IP Networks: (1) to accelerate the convergence of IP routing; (2) to employ the proactive recovery strategy; (3) to increase the speed and accuracy of failure detection. This paper analyzes the defects of the existing solutions. In order to realize fast failure recovery of IP networks, we should: (1) balance the traffic load after recovery; (2) do interoperability test and redesign the routers’ architecture. Key words: IP network; failure recovery; routing convergence; backup path; load balance 1 引 言 [1] 1983 年,ARPANET 的所有主机更改成TCP/IP 协议栈,并在网络中加入路由器,从而形成了互联网的雏形 . 最初,互联网主要用于传输非实时业务,如收发电子邮件,浏览网页等.互联网的路由收敛需要数秒的时间[2],对于 非实时业务来说,这个时间是可以忍受的.而在接下来的 20