软交换设备高可用性的研究与实现
摘要
通信系统对于可用性有着很高要求,传统电信系统的四五类交换机已经达到了很高的稳定性及高可用性,作为下一代网络的核心控制设备的软交换也必须满足高可用性(High Availability)的要求。但是由于高可用性技术的复杂性,并且目前在技术实现上较少有公开的描述,其设计仍然是研究的重点。
本文首先通过对高可用性的理论及技术加以研究,掌握可用性的基本概念、度量手段和分析方法,从系统结构、软硬件方面研究了目前通用的高可用性技术。在此基础上,对高可用性软交换设备进行研究设计,提出了一系列实现软交换设备高可用性解决方法。
本文按照开放式高可用性系统架构的要求,对高可用性软交换设备给出了总体设计,并对于硬件平台、高可用性管理中间件、应用层同步等组件的设计与实现进行详细研究分析。在设计过程中,充分地分析了通用承载平台适用于高可用性控制的热插拔基础、通信基础、故障检测与控制基础,以硬件冗余和软件控制作为高可用性的实现基础,以有效地实现故障处理和提高应用的可靠性为系统高可用性地的实现途径。对高可用性管理中间件的设计实现加以重点研究分析,设计了相关的高可用性系统状态机,实现了系统的初始化和切换功能。
本文对于系统的故障检测及冗余结构设计等方面的关键技术进行了深入的研究与探讨,其中论及的技术包括:看门狗、心跳检测、N+1冗余备份。最后,通过连续时间马尔可夫模型对本系统采用的N+1冗余结构的可靠性进行分析,对于系统的高可用性进行测试,取得了较为理想的结果。
The telecom system always pays much attention on High availability. The Class 4 and 5 switch in traditional telecom system has already achieve High Availability.As the core equipment of NGN,the Softswitch must run with High Availability. However,because of the complication of HA technology and the lack of open description, the methods to implement a HA system in Softswitch equipment becomes our research focus.
Through the study of HA technology,we can master the basic conception and implementation of HA system.So that we can design and research the implementation of a HA system and conclude a set of solutions.
The overall system framework which is built on the open architecture HA solution provided by HA forum is discussed,and also the HA implementation of application layer,OS layer,hardware layer and most important,the HA manage middleware.In the course of design, characteristic of the universal hardware platform have been completely analyzed from the point of view of high availability, such as the ability in hot swapping, inter-communicating, fault detection and control. With the fact that high availability is composed by hardware and software reliability and maintainability, one way to realize high availability is to improve software reliability and combine hardware redundancy and software maintenance, so as to get the ability to deal with hardware fault and improve the availability of application layer. And the design of the HA manage middleware and Fault detect mechanism are discussed detailedly.
And some critical technology,including heartbeat,watchdog,and redundancy of N+1module is deeply researched. Finally, the result of the analysis of high availability in logic and performance shows that the whole solution has been proven to be able to achieve the goal of high availability.
本文首先通过对高可用性的理论及技术加以研究,掌握可用性的基本概念、度量手段和分析方法,从系统结构、软硬件方面研究了目前通用的高可用性技术。在此基础上,对高可用性软交换设备进行研究设计,提出了一系列实现软交换设备高可用性解决方法。
本文按照开放式高可用性系统架构的要求,对高可用性软交换设备给出了总体设计,并对于硬件平台、高可用性管理中间件、应用层同步等组件的设计与实现进行详细研究分析。在设计过程中,充分地分析了通用承载平台适用于高可用性控制的热插拔基础、通信基础、故障检测与控制基础,以硬件冗余和软件控制作为高可用性的实现基础,以有效地实现故障处理和提高应用的可靠性为系统高可用性地的实现途径。对高可用性管理中间件的设计实现加以重点研究分析,设计了相关的高可用性系统状态机,实现了系统的初始化和切换功能。
本文对于系统的故障检测及冗余结构设计等方面的关键技术进行了深入的研究与探讨,其中论及的技术包括:看门狗、心跳检测、N+1冗余备份。最后,通过连续时间马尔可夫模型对本系统采用的N+1冗余结构的可靠性进行分析,对于系统的高可用性进行测试,取得了较为理想的结果。
The telecom system always pays much attention on High availability. The Class 4 and 5 switch in traditional telecom system has already achieve High Availability.As the core equipment of NGN,the Softswitch must run with High Availability. However,because of the complication of HA technology and the lack of open description, the methods to implement a HA system in Softswitch equipment becomes our research focus.
Through the study of HA technology,we can master the basic conception and implementation of HA system.So that we can design and research the implementation of a HA system and conclude a set of solutions.
The overall system framework which is built on the open architecture HA solution provided by HA forum is discussed,and also the HA implementation of application layer,OS layer,hardware layer and most important,the HA manage middleware.In the course of design, characteristic of the universal hardware platform have been completely analyzed from the point of view of high availability, such as the ability in hot swapping, inter-communicating, fault detection and control. With the fact that high availability is composed by hardware and software reliability and maintainability, one way to realize high availability is to improve software reliability and combine hardware redundancy and software maintenance, so as to get the ability to deal with hardware fault and improve the availability of application layer. And the design of the HA manage middleware and Fault detect mechanism are discussed detailedly.
And some critical technology,including heartbeat,watchdog,and redundancy of N+1module is deeply researched. Finally, the result of the analysis of high availability in logic and performance shows that the whole solution has been proven to be able to achieve the goal of high availability.
引文
[1] PCI Industrial Computer Manufactures Group.CompactPCI Hot Swap Specification 2.1 ,R2.0.2001:4~18.
[2] Alexey Polonsky.CompactPCI hot swapping. CompactPCI and AdvancedTCA Systems,2004,4:12~15.
[3] Intel, Enclosure Management .The Intelligent Platform Management Interface and enclosure management. Electronic Engineering, 2000:19.
[4] David Greenfield. Server Management Systems. Network Magazine, 2004,19(4):68~69.
[5] Don Harrrison,Honey J . System management and compactPCI systems[J].CompactPCl Systems,1999,68(1):58~ 62.
[6]陈进,申峰.CompactPCI技术特点与发展.广东通信技术,2003,23:37~40.
[7] Chuck Hill.CompactTCA update.CompactPCI and AdvancedTCA Systems,2005,1:22~23.
[8]王洛欣,申卫昌,朱莉明等.先进的通信计算机架构ATCA及应用研究.科学技术与工程, 2006 ,6:2132~2134.
[9] PCI Industrial Computer Manufactures Group.PICMG 3 . 0 short form specification[z].2003.25~31.
[10] Performance Technologies Inc.CompactPCI IP backplane standard for next generation networks. Electronic Engineering. 2001,4:44.
[11] Legato? Co-StandbyServer? 2000 White Paper .
[12] ROSE Datasystems.ROSE HA高可用性白皮书.
[13]卢美莲,程时端.软交换技术.中兴通讯技术,2002, 49(4):49.
[14]郎为民,靳焰,王逢东等.软交换体系结构研究.数字通信世界, 2007, 2:41~43.
[15]易清明,陈燕燕.软交换技术发展及其应用.移动通信, 2005,10:130~131.
[16] Paul Drew ,Chris Gallon. Next-Generation VoIP Network Architecture. MSF Technical Report,2003:1~19.
[17]糜正琨,王文鼎.软交换技术与协议.第一版.北京:北京邮电大学出版社. 2002: 35~75.
[18]魏航,杨学良.下一代网络(NGN)的体系结构与软交换协议.计算机应用,2003,23(12):75~78.
[19]信息产业部电信研究院;中国联合通信有限公司;中国电信集团公司. YD/T 1434-2006.软交换设备总体技术要求.2006.
[20] 5020移动软交换——安全可靠的电信级移动软交换系统.通信世界,2006,30:19~20.
[21] H.Cristian,Understanding fault-tolerant distributed systems,Communications of the ACM, 1999, 34(2): 56~78.
[22]卢德平.计算机系统的可靠性浅析.高性能计算技术,2005,172:1~4.
[23]傅佩琛,赵霖.计算机系统硬件软件可靠性理论及其应用.国防工业出版社,1990:11~77.
[24]刘耀,胡越明,金利峰.高可靠计算机系统的容错技术.计算机工程,2004,12(4):590~592.
[25]蒋乐天,徐国治,应忍冬等.系统可靠性和可用性分析技术.电讯技术,2002,4:124~126.
[26] Tim Anderson, Todd Grabbe, Julian Hammersley, et al. Providing Open Architecture High Availability Solutions. Revision 1.0. 2001:13~57.
[27]唐爱龙,蒋华,黄秋勇.软件可靠性技术的研究.柳州师专学报,2006,21(2):114~115.
[28] A. Avizienis.The Method of N-Version Programming in Software Fault Tolerance. M.R.Lyued, Wiley, 1995:23~46.
[29]刘睿涛.并行计算机高可用性分析与设计.计算机工程与科学,2005,27(3):104~110.
[30]徐鹏,苏森,陈俊亮.基于软交换的下一代网络的高可用性研究.高技术通讯,2003,12:12~14.
[31]赵京玺.软交换系统可靠性分析.现代通信,2005,12:33~34.
[32]朱木成,晏宪华,袁鑫昌.软交换设备之间互通的研究和实现.计算机工程与应用,2004,15:154~156.
[33] Williams S. The softswitch advantage. IEEE Review, Issue 4, 2002, 48:25~29.
[34] Ki-Young Jung,Mi-Jung Hong.Design considerations for NGN softswitch-element management system,Network Operations and Management Symposium. IEEE/IFIP Vol 1,2004: 19~23.
[35]史宏达,CompactPCI总线规范及其热插拔性能简介.军事通信技术,2002,23(1):76~77.
[36] PCI Industrial Computer Manufactures Group. CompactPCI Packet Switching Backplane. PICMG 2.16 Draft 0.6.5,2001:12~35.
[37]陈艳峰,张光昭,胡乐明.CompactPCI热插拔技术研究现状评述.测控技术,2002,21(10):1~5.
[38]陈磊,CompactPCI热切换技术的应用.测控技术,2004,23(8):44~46.
[39] YANG.M.H, ZHU.X.M, LAO.J.X, et al. Design and reliability analysis for parlay gateway. Proceedings of International Conference on Computer Communication(ICCC 2004)[C], 2004:195~200.
[40] Smart Xie. High Level Design of HA manager. UTStarcom technology Document. 2001:17~35.
[41]陈义俊.watchdog技术与应用.抚顺石油学院学报.1995,15(3):51~54.
[42]李观文,衣平,邓英华.看门狗技术在改善系统可靠性中的应用.机床电器2005,1:26~28.
[43] Redunant System Slot Software for the Intel NetStructure ZT5550 High Availability Processor Board Software Manual. Intel Corporation. 2002:43~54.
[44]董恩强,沈雁,赵合计.强实时环境下的高可靠性技术研究.计算机工程与科学,2006,28(3):102~104.
[45] Zonghao Hou, Yongxiang Huang, Shouqi Zheng. Design and implementation of heartbeat in multi-machine environment. Advanced Information Networking and Applications, 2003. AINA 2003. 17th International Conference on 27-29 March 2003:583~586.
[46]刘延夫,基于马尔可夫分析方法的软件系统可靠性研究:[硕士论文],长春:长春理工大学计算机软件与理论,2005.
[2] Alexey Polonsky.CompactPCI hot swapping. CompactPCI and AdvancedTCA Systems,2004,4:12~15.
[3] Intel, Enclosure Management .The Intelligent Platform Management Interface and enclosure management. Electronic Engineering, 2000:19.
[4] David Greenfield. Server Management Systems. Network Magazine, 2004,19(4):68~69.
[5] Don Harrrison,Honey J . System management and compactPCI systems[J].CompactPCl Systems,1999,68(1):58~ 62.
[6]陈进,申峰.CompactPCI技术特点与发展.广东通信技术,2003,23:37~40.
[7] Chuck Hill.CompactTCA update.CompactPCI and AdvancedTCA Systems,2005,1:22~23.
[8]王洛欣,申卫昌,朱莉明等.先进的通信计算机架构ATCA及应用研究.科学技术与工程, 2006 ,6:2132~2134.
[9] PCI Industrial Computer Manufactures Group.PICMG 3 . 0 short form specification[z].2003.25~31.
[10] Performance Technologies Inc.CompactPCI IP backplane standard for next generation networks. Electronic Engineering. 2001,4:44.
[11] Legato? Co-StandbyServer? 2000 White Paper .
[12] ROSE Datasystems.ROSE HA高可用性白皮书.
[13]卢美莲,程时端.软交换技术.中兴通讯技术,2002, 49(4):49.
[14]郎为民,靳焰,王逢东等.软交换体系结构研究.数字通信世界, 2007, 2:41~43.
[15]易清明,陈燕燕.软交换技术发展及其应用.移动通信, 2005,10:130~131.
[16] Paul Drew ,Chris Gallon. Next-Generation VoIP Network Architecture. MSF Technical Report,2003:1~19.
[17]糜正琨,王文鼎.软交换技术与协议.第一版.北京:北京邮电大学出版社. 2002: 35~75.
[18]魏航,杨学良.下一代网络(NGN)的体系结构与软交换协议.计算机应用,2003,23(12):75~78.
[19]信息产业部电信研究院;中国联合通信有限公司;中国电信集团公司. YD/T 1434-2006.软交换设备总体技术要求.2006.
[20] 5020移动软交换——安全可靠的电信级移动软交换系统.通信世界,2006,30:19~20.
[21] H.Cristian,Understanding fault-tolerant distributed systems,Communications of the ACM, 1999, 34(2): 56~78.
[22]卢德平.计算机系统的可靠性浅析.高性能计算技术,2005,172:1~4.
[23]傅佩琛,赵霖.计算机系统硬件软件可靠性理论及其应用.国防工业出版社,1990:11~77.
[24]刘耀,胡越明,金利峰.高可靠计算机系统的容错技术.计算机工程,2004,12(4):590~592.
[25]蒋乐天,徐国治,应忍冬等.系统可靠性和可用性分析技术.电讯技术,2002,4:124~126.
[26] Tim Anderson, Todd Grabbe, Julian Hammersley, et al. Providing Open Architecture High Availability Solutions. Revision 1.0. 2001:13~57.
[27]唐爱龙,蒋华,黄秋勇.软件可靠性技术的研究.柳州师专学报,2006,21(2):114~115.
[28] A. Avizienis.The Method of N-Version Programming in Software Fault Tolerance. M.R.Lyued, Wiley, 1995:23~46.
[29]刘睿涛.并行计算机高可用性分析与设计.计算机工程与科学,2005,27(3):104~110.
[30]徐鹏,苏森,陈俊亮.基于软交换的下一代网络的高可用性研究.高技术通讯,2003,12:12~14.
[31]赵京玺.软交换系统可靠性分析.现代通信,2005,12:33~34.
[32]朱木成,晏宪华,袁鑫昌.软交换设备之间互通的研究和实现.计算机工程与应用,2004,15:154~156.
[33] Williams S. The softswitch advantage. IEEE Review, Issue 4, 2002, 48:25~29.
[34] Ki-Young Jung,Mi-Jung Hong.Design considerations for NGN softswitch-element management system,Network Operations and Management Symposium. IEEE/IFIP Vol 1,2004: 19~23.
[35]史宏达,CompactPCI总线规范及其热插拔性能简介.军事通信技术,2002,23(1):76~77.
[36] PCI Industrial Computer Manufactures Group. CompactPCI Packet Switching Backplane. PICMG 2.16 Draft 0.6.5,2001:12~35.
[37]陈艳峰,张光昭,胡乐明.CompactPCI热插拔技术研究现状评述.测控技术,2002,21(10):1~5.
[38]陈磊,CompactPCI热切换技术的应用.测控技术,2004,23(8):44~46.
[39] YANG.M.H, ZHU.X.M, LAO.J.X, et al. Design and reliability analysis for parlay gateway. Proceedings of International Conference on Computer Communication(ICCC 2004)[C], 2004:195~200.
[40] Smart Xie. High Level Design of HA manager. UTStarcom technology Document. 2001:17~35.
[41]陈义俊.watchdog技术与应用.抚顺石油学院学报.1995,15(3):51~54.
[42]李观文,衣平,邓英华.看门狗技术在改善系统可靠性中的应用.机床电器2005,1:26~28.
[43] Redunant System Slot Software for the Intel NetStructure ZT5550 High Availability Processor Board Software Manual. Intel Corporation. 2002:43~54.
[44]董恩强,沈雁,赵合计.强实时环境下的高可靠性技术研究.计算机工程与科学,2006,28(3):102~104.
[45] Zonghao Hou, Yongxiang Huang, Shouqi Zheng. Design and implementation of heartbeat in multi-machine environment. Advanced Information Networking and Applications, 2003. AINA 2003. 17th International Conference on 27-29 March 2003:583~586.
[46]刘延夫,基于马尔可夫分析方法的软件系统可靠性研究:[硕士论文],长春:长春理工大学计算机软件与理论,2005.