摘要
计算机网络作为现代社会重要的基础设施，其可靠性、稳定性至关重要。而作为网络系统关键设备的核心路由器，其可靠性的高低将直接决定整个网络是否能够正常工作。
本文首先介绍了高性能核心路由器GF8516的体系结构，分析了计算机系统可靠性理论的基本原理。基于可靠性研究的一个重要结论——冗余技术可以建立起整体性能超过其部件性能的系统，我们结合网络设备的特点，总结出核心路由器可靠性设计的原则：关键部件的冗余结构设计和高效的故障管理技术相结合。
在GF8516核心路由器中，我们对其关键部件—主控处理机和交换处理机采用了1+1冗余结构设计，并采用高效的故障管理技术实现了它们的动态冗余。本文对冗余设计和故障管理的关键技术进行了深入的探讨，系统分析了在进行冗余设计和实现故障管理过程中涉及到的技术难点，给出了若干可行的方案，并结合GF85616的体系结构和技术特点进行了比较与论证。
本文第五章对GF8516核心路由器的可用度进行了计算和评估，上述所讨论到的技术应用到GF8516的可靠性设计中，取得了令人满意的效果。文章最后指出下一步的研究方向和应该进行的工作。
As an essential part of social infrastructure, the computer network is highly signified as to its reliability and stability. However, its operation is directly determined by the reliability of core router, the key equipment to the network system.
This paper begins with an introduction of the architecture of high performance core router GF8561 and an analysis of the basic principle of the reliability theory concerning computer systems. Then, a system whose overall capability is superior to that of its components is established on the basis of redundancy technique drawn from reliability theory. Combined with features of the network equipments, the integrity of key components' redundancy configuration design and high-efficient technique for fault management as the principle of core router's reliability design is deduced.
We have applied 1+1 redundancy configuration design to control processor and switch processor, the key components of core router GF8516, whose dynamic redundancy is further realized through a combination with techniques for fault management. This paper reports on technical problems in the process of redundancy design and fault management and proposes feasible solutions.
The fifth chapter deals with prediction and estimation about the usability of core router GF8516. So far, the application of the aforementioned technique to GF8516 reliability design produces satisfying results. The last part of the paper offers suggestions for further research.
引文
[1] Chris Oggerino, "SSL and TLS Designing and Building Secure Systems", Addison Wesley Longman, Inc, June,2002
[2] Bill Krause, "Use Processor Redundancy for Maximum Reliability", http://www.CommsDesign.com, February 1, 2002
[3] BellCore TR-332 技术文档，http://telecom-info.telcodia.com/site-cgi/ido/index.html
[4] Chuck Hill, "High-availability Systems Made Easy", http://www.Comms- Design.com, January 3, 2001
[5] T. Li, B. Cole, P. Morton "Cisco Hot Standby Router Protocol (HSRP)", RFC2281, March 1998
[6] The Torrent Networking Technologies Paper, "Routing and Layer3 Switching: Understanding the Critical Difference", 1997
[7] David C.Lee, Daniel L. Lough,and Phillip E. Benchoff, "The Next Generation of the Internet: Aspects of the Internet Protocol Version 6", in IEEE Network Magazine, Jan/Feb 1998
[8] Peter Newman, Greg Minshall,Tom Lyon,and Lzrry Huston,Ipsilon Networks Inc. "IP Switching and Gigbit Routers", in IEEE Communication Magazine, Jan 1997
[9] Kevin Thompson, Gregory J.Miller, and Rick Wilder "Wide-Area Internet Traffic Patterns and Characteristics" in IEEE Network Magazine, Nov/Dec 1997.
[10] Andrew S.Tanenbaum "Computer Networks (third edition)", Prentice Hall PTR
[11] 刘石丹，林晓东，“高可用性系统结构的研究与实现”，1998年3月
[12] 高文 祝明发，“基于生灭过程的机群系统高可用性分析与设计”，2000年5月
[13] 吕建伟，刘中华，“可修复系统可用性模型的建立与分析”，1999年1月
[14] 张志群，孔令山等，“高速数据交换系统双机热备份的设计实现”，2001年6月
[15] 张宇，洪炳熔，“软件容错技术的研究现状与发展”，1999年9月
[16] 季会嫒，孟亚等，“一种容错系统可靠性分析方法”，2001年6月
[17] 王先培，“计算机网络可靠性概念体系及分析方法的研究”，1999年1月
[18] 周明天，汪文勇，《TCP/IP网络原理与技术》，清华大学出版社，1993年12月
[19] Terry Slattery著，苏金树等译，《Cisco网络高级IP路由技术》，机械工业出版社，1999年6月
[20] 何宁，张云龙，吴晨，“核心路由器发展趋势”，2001年11月
[21] Gary R. Wright, W. Richard Stevens著，陆雪莹等译，《TCP/IP详解 卷2：实现》，机械工业出版社，2000年7月
[22] IBM, "IBM Packet Routing Switch PRS64G 64 Gbps Switch Core Switch Control Microcode for Reference design", 2001
[23] IBM, "Hardware Reference Platform Technical Description", 2001
[24] IBM, "SWICC General Description", 2001
[25] IBM, "SINGLE IBM PowerPRS64G Reference design SWICC Installation / Operation
User's Manual", 2001
[26] IBM, "PowerPRSTM 64G(PRIZMA-EP) Switch Core 64 Gbps Switch Board-Advanced Functional Specifications Version 1． 1", 2001
[27] InfineOn, "SEROCCO-H 2 Channel Serial Optimized Communication Controller for HDLC/PPP", Sep, 2000
[28] IBM, " mM Network Processor Software Overview ", December 11,2001
[29] WindRiver, "Network Protocol Toolkit User's Guide (5． 4 Edition 2) "
[30] WindRiver, "SENS for Tornado Release Notes (1． 0 Edition 1)