基于广域网的细粒度对等动态漂移技术研究
|
摘要
随着网络规模和应用的不断扩展，网络环境下的服务可用性问题越来越受到关注。人们发现高可用性的获得仅仅依靠单个服务器的性能提高是不够的，还必须从系统结构方面进行改进。动态漂移技术是在多备份服务器系统的基础上实现的一种通过改变系统的整体结构而提高系统性能的技术。通过服务迁移的方式，该技术可以实现服务器在局部故障下的不间断传输。
针对目前广泛应用的动态漂移技术实现方法的不足，本文紧密围绕跨广域网、细粒度、无前端调度器、对客户方透明的对等动态漂移实现的关键技术，开展了研究工作，所做的工作主要集中在：
1) 在对现有动态漂移技术实现方法进行分析和比较的基础上，提出了跨广域网、细粒度、无前端调度器、对客户方透明的对等动态漂移(WGCTEM-WAN fine-Grained no-Controller Transparent Equal dynamic Migration)思路和模型；
2) 对于所使用的相关协议(包括TCP协议和SCTP协议)进行了研究；
3) 对基于WAN的细粒度TCP连接迁移技术做了深入研究，并对迁移系统中的辅助技术做了研究和实现；
4) 设计和实现了一个基于HTTP协议的WGCTEM基本原型系统；
5) 对WGCTEM基本原型在应急响应和事故恢复系统中的应用进行了总体设计。
本文设计并实现了一个基于HTTP应用的WGCTEM基本原型，该原型对于提高系统的生存性有较高的指导意义。WGCTEM原型已经在应急响应和事故恢复系统中得到了应用，为项目验收打下了坚实的基础。
全文共分六章，第一章为绪论部分，主要介绍了课题的研究背景、动态漂移技术的应用背景和本文的主要研究内容；第二章为原理阐述部分，阐述了动态漂移技术的发展现状，并着重对现有实现方法的优劣做了比较；第三章对细粒度连接迁移技术及其各种实现方式做了深入论述；第四章提出了一种新的动态漂移思路，即跨广域网、细粒度、无前端调度器、对客户方透明的对等动态漂移，并对以TCP连接迁移为核心的动态漂移体系结构做了研究；第五章阐述了WGCTEM基本原型的实现技术，并对该原型在应急响应和事故恢复系统中的应用进行了总体设计；第六章为结论部分，总结了本文所做的工作和今后的研究方向。
With network scale expanding and network applications developing, people pay more and more attentions to service availability in network environments. In order to get high availability, We not only improve a single server's performance, but try to change the system architecture. For a system with the same servers, Dynamic migration technique can enhance its performance by altering the system architecture. By means of service migration, the technique may implement continuous data transmission under server breakdown.
After analyzing the disadvantages of the traditional realizations of dynamic migration technique, we focus on our work on the pivotal implementation technology of WGCTEM (based-on-WAN, fine-Grained, no-Controller, Transparent-to-client and Equal-among-servers dynamic Migration). The following contents are discussed in the thesis.
1) Based on the analysis and comparison of traditional implementations of dynamic migration technique, An idea of WGCTEM is presented;
2) The related protocols such as TCP and SCTP are mainly discussed;
3) Fine-grained TCP connection migration technique is discussed deeply, and assistant technologies in migration system are implemented;
4) WGCTEM prototype based on HTTP are designed and implemented;
5) The overall designs for combining WGCTEM prototype with the emergency-response and fault-recovery system are put forward.
The WGCTEM prototype based on HTTP is significant to the improvement of the system survivability and has been applied into the emergency-response and fault-recovery system, which lays the foundations for the final examination of 863 experts.
There are totally six chapters in this paper. In chapter one, the background of our project, application background of dynamic migration technique and the main research areas are introduced. In chapter two, various realizations of dynamic migration technique are discussed. In chapter three, fine-grained TCP connection migration technique based on WAN and its various implementations are discussed. In chapter four, a new idea, namely WGCTEM, is presented, and dynamic migration architecture based on TCP connection migration is discussed. In chapter five, the implementation technology of WGCTEM prototype is discussed, and the overall designs for combining WGCTEM prototype with the responding and recovering system are presented. At last, the work and contributions of this paper are summarized, and the future work is put forward.
针对目前广泛应用的动态漂移技术实现方法的不足，本文紧密围绕跨广域网、细粒度、无前端调度器、对客户方透明的对等动态漂移实现的关键技术，开展了研究工作，所做的工作主要集中在：
1) 在对现有动态漂移技术实现方法进行分析和比较的基础上，提出了跨广域网、细粒度、无前端调度器、对客户方透明的对等动态漂移(WGCTEM-WAN fine-Grained no-Controller Transparent Equal dynamic Migration)思路和模型；
2) 对于所使用的相关协议(包括TCP协议和SCTP协议)进行了研究；
3) 对基于WAN的细粒度TCP连接迁移技术做了深入研究，并对迁移系统中的辅助技术做了研究和实现；
4) 设计和实现了一个基于HTTP协议的WGCTEM基本原型系统；
5) 对WGCTEM基本原型在应急响应和事故恢复系统中的应用进行了总体设计。
本文设计并实现了一个基于HTTP应用的WGCTEM基本原型，该原型对于提高系统的生存性有较高的指导意义。WGCTEM原型已经在应急响应和事故恢复系统中得到了应用，为项目验收打下了坚实的基础。
全文共分六章，第一章为绪论部分，主要介绍了课题的研究背景、动态漂移技术的应用背景和本文的主要研究内容；第二章为原理阐述部分，阐述了动态漂移技术的发展现状，并着重对现有实现方法的优劣做了比较；第三章对细粒度连接迁移技术及其各种实现方式做了深入论述；第四章提出了一种新的动态漂移思路，即跨广域网、细粒度、无前端调度器、对客户方透明的对等动态漂移，并对以TCP连接迁移为核心的动态漂移体系结构做了研究；第五章阐述了WGCTEM基本原型的实现技术，并对该原型在应急响应和事故恢复系统中的应用进行了总体设计；第六章为结论部分，总结了本文所做的工作和今后的研究方向。
With network scale expanding and network applications developing, people pay more and more attentions to service availability in network environments. In order to get high availability, We not only improve a single server's performance, but try to change the system architecture. For a system with the same servers, Dynamic migration technique can enhance its performance by altering the system architecture. By means of service migration, the technique may implement continuous data transmission under server breakdown.
After analyzing the disadvantages of the traditional realizations of dynamic migration technique, we focus on our work on the pivotal implementation technology of WGCTEM (based-on-WAN, fine-Grained, no-Controller, Transparent-to-client and Equal-among-servers dynamic Migration). The following contents are discussed in the thesis.
1) Based on the analysis and comparison of traditional implementations of dynamic migration technique, An idea of WGCTEM is presented;
2) The related protocols such as TCP and SCTP are mainly discussed;
3) Fine-grained TCP connection migration technique is discussed deeply, and assistant technologies in migration system are implemented;
4) WGCTEM prototype based on HTTP are designed and implemented;
5) The overall designs for combining WGCTEM prototype with the emergency-response and fault-recovery system are put forward.
The WGCTEM prototype based on HTTP is significant to the improvement of the system survivability and has been applied into the emergency-response and fault-recovery system, which lays the foundations for the final examination of 863 experts.
There are totally six chapters in this paper. In chapter one, the background of our project, application background of dynamic migration technique and the main research areas are introduced. In chapter two, various realizations of dynamic migration technique are discussed. In chapter three, fine-grained TCP connection migration technique based on WAN and its various implementations are discussed. In chapter four, a new idea, namely WGCTEM, is presented, and dynamic migration architecture based on TCP connection migration is discussed. In chapter five, the implementation technology of WGCTEM prototype is discussed, and the overall designs for combining WGCTEM prototype with the responding and recovering system are presented. At last, the work and contributions of this paper are summarized, and the future work is put forward.
引文
[1] Florin Sultan, Kiran Srinivasam, Deepa Iyer, and Liviu Iftode: Migratory TCP: Highly Available Internet Services Using Connection Migration, Department of Computer Science Rutgers University Piscataway, NJ 08854-8019.
[2] 杨兵，黄遵国，胡光明，胡华平：基于高可用性的动态漂移技术研究，计算机工程与科学(已录用)
[3] http://www.cnvod.net/.
[4] 庞毅林，蒋翠玲：进程迁移研究，计算机工程与科学，2001，23(5)．
[5] 黄遵国：多样化动态备份与生存技术研究，第九届C3I理论学术研讨会，2000，11．
[6] http://www.ctiforum.com/.
[7] http://www.bind.org.
[8] Alteon Web Systems. Layer 7 Web Switching. http://www.alteonwebsystems.com/products/whitepapers/layer7switching.
[9] http://www.linuxvirtualserver.org.
[10] Ralf.S.Engelschall, Load Balancing Your Web Site. http://www.newarchitectmag.com/archives/1998/05/engelschall/.
[11] DNS Round Robin High-Availability Load Sharing, http://.polyserve.com.
[12] Cisco Systems. Failover configuration for LocalDirector. http://www.cisco.com/warp/public/cc/pd/cssr.400/tech/locdf wp.htm.
[13] http://www.redhat.com/support/wpapers/piranha/.
[14] http://www.linuxvirtualserver.org.
[15] http://www.turbolinux.com.
[16] RFC2960 Stream Control Transmission Protocol, IETF Network Working Group Request for Com-ments: 2960, Category: Standards Track.
[17] Alex C. Snoeren, David G. Andersen, and Hari Balakrishnan: Fine-Grained Failover Using Connection Migration, MIT Laboratory for Computer Science Cambridge, MA 02139.
[18] Alex C. Snoeren and Hari Balakrishnan: An End-to-End Approach to Host Mobility, MIT Laboratory for Computer Science Cambridge, MA 02139.
[19] Florin Sultan, Kiran Srinivasan, Deepa Iyer, Liviu Iftode, Migratory TCP: Highly Available Internet Services Using Connection Migration, Rutgers University Technical Report DCS-TR-462, December 2001.
[20] Internet Draft, IETF, Nov. 2000. draft-snoeren-tcp-migrate-00.txt.
[21] S. Raman and S. McCanne:A model, analysis, and protocol framework for soft state-based communication. In Proc. ACM SIGCOMM '99, Sept. 1999.
[22] 黄遵国：一种信息安全事件的快速响应与恢复框架研究，计算机工程与科学，2001
[2] 杨兵，黄遵国，胡光明，胡华平：基于高可用性的动态漂移技术研究，计算机工程与科学(已录用)
[3] http://www.cnvod.net/.
[4] 庞毅林，蒋翠玲：进程迁移研究，计算机工程与科学，2001，23(5)．
[5] 黄遵国：多样化动态备份与生存技术研究，第九届C3I理论学术研讨会，2000，11．
[6] http://www.ctiforum.com/.
[7] http://www.bind.org.
[8] Alteon Web Systems. Layer 7 Web Switching. http://www.alteonwebsystems.com/products/whitepapers/layer7switching.
[9] http://www.linuxvirtualserver.org.
[10] Ralf.S.Engelschall, Load Balancing Your Web Site. http://www.newarchitectmag.com/archives/1998/05/engelschall/.
[11] DNS Round Robin High-Availability Load Sharing, http://.polyserve.com.
[12] Cisco Systems. Failover configuration for LocalDirector. http://www.cisco.com/warp/public/cc/pd/cssr.400/tech/locdf wp.htm.
[13] http://www.redhat.com/support/wpapers/piranha/.
[14] http://www.linuxvirtualserver.org.
[15] http://www.turbolinux.com.
[16] RFC2960 Stream Control Transmission Protocol, IETF Network Working Group Request for Com-ments: 2960, Category: Standards Track.
[17] Alex C. Snoeren, David G. Andersen, and Hari Balakrishnan: Fine-Grained Failover Using Connection Migration, MIT Laboratory for Computer Science Cambridge, MA 02139.
[18] Alex C. Snoeren and Hari Balakrishnan: An End-to-End Approach to Host Mobility, MIT Laboratory for Computer Science Cambridge, MA 02139.
[19] Florin Sultan, Kiran Srinivasan, Deepa Iyer, Liviu Iftode, Migratory TCP: Highly Available Internet Services Using Connection Migration, Rutgers University Technical Report DCS-TR-462, December 2001.
[20] Internet Draft, IETF, Nov. 2000. draft-snoeren-tcp-migrate-00.txt.
[21] S. Raman and S. McCanne:A model, analysis, and protocol framework for soft state-based communication. In Proc. ACM SIGCOMM '99, Sept. 1999.
[22] 黄遵国：一种信息安全事件的快速响应与恢复框架研究，计算机工程与科学，2001