基于ATCA流媒体网络服务器研究
摘要
流媒体系统随着Internet的发展而不断壮大,对于底层的计算平台也提出了更高的要求——更为强大的计算性能、更为强大的数据传输能力、与强大计算性能相配套的散热性能和更为强大的系统管理能力,现存的计算架构难以应对。而先进电信计算架构(Advanced Telecom Computing Architecture,ATCA)在计算性能、网络传输能力、管理性能以及专属特性设计方面和原有的计算架构相比都有着明显的优势。ATCA是由PICMG组织(PCI Industrial Computer Manufacturers Group)提出的PICMG3.X一系列规范总称,是高性能的电信和工业计算应用开发开放的标准,非常适合构建流媒体系统。
本文阐述了基于ATCA的流媒体系统架构,分析了系统组成,并且设计了各个模块的功能.系统具有高度可伸缩性,统一的管理平台,灵活的内容分发子系统,广泛的硬件、应用和内容支持,丰富的接口等特点。
在流媒体系统中网络服务器设计是关键之一。在网络服务器设计过程中,本文总结了一些关键设计问题,如数据拷贝、上下文转换、内存分配、锁竞争等,同时提出了针对上述问题所要遵循的原则。
在网络服务器模型中,有线程驱动模型和事情驱动模型。虽然线程是表示并发的主要方式,但是在线程驱动模型服务器中,随着线程数目不断增加,系统的开销不断增长,到一定程度后会导致系统总体性能的下降;而事件驱动模型网络服务器用一个线程处理所有任务,避免了同步引起的很多问题,但是事件驱动模型不能利用SMP来提高性能,并且事件处理线程有可能阻塞。
基于线程驱动模型和事件驱动模型的特点,本文提出了网络服务器设计框架混合模型,充分利用了两种模型的优点。最后归纳了如何运用Wrap、Pipeline、Replicate、Combine等设计模式构建基于混合模型的流媒体网络服务器。最后阐述了CDN系统的结构以及其中的关键技术。
With the rapid development of Internet, streaming media system expands it's use continuously, so it requires the rock-bottom computing platform higher performance—more powerful computing performance 、 more high speed data transmission、 more formidable heat dissipation ability and more strong system management ability. The extant computing platform cannot meet these requirements. But ATCA (Advanced Telecom Computing Architecture) has the obvious superiority in all of computing performance, the ability of network transmission, management performance as well as the private characteristic design aspect etc. ATCA is a series of standards PICMG3.X proposed by PICMG (PCI Industrial Computer Manufacturers Group) ,is open standard for high performance telecommunication and industry system development, so it is very suitable using in steaming media server system.This paper elaborates streaming media system architecture based on ATCA, analyzes the system composition, and defines each module function. The system is highly expandable. Also, it has uniform management platform, flexible content distribution system, widespread support of hardware, application and content, rich interface and so on.Network server design is one of the most important key in implementing a streaming media system. In design the network server, some essential questions are summarized, like data copy, context switches, the memory Allocation, lock contention and so on simultaneously proposes some principles in view of above questions.Threaded server and event-driver server is the two most general network servers. Threads have become the dominant form of expressing concurrency, but in threaded server, as the number of threads in a system grows, operating
system overhead increases, leading to a decrease in the overall performance of the system. Event-driven systems are typically structured as a program thread that loops continuously, processing events of different types from a queue. Event-driven programming avoids many of the bugs associated with synchronization. However, Events generally cannot take advantage of SMP systems for performance. Also, event processing threads can block regardless of the I/O mechanisms used.According to Threaded Servers and Event-drive Servers' characteristics, a hybrid thread and event server is proposed to obtain both benefit of threads and events. The paper also induces how to use design rules such as Wrap, Pipeline, Combine and Replicate to design streaming media network server. At last, it describes the CDN system architecture and looks into the key technologies.
本文阐述了基于ATCA的流媒体系统架构,分析了系统组成,并且设计了各个模块的功能.系统具有高度可伸缩性,统一的管理平台,灵活的内容分发子系统,广泛的硬件、应用和内容支持,丰富的接口等特点。
在流媒体系统中网络服务器设计是关键之一。在网络服务器设计过程中,本文总结了一些关键设计问题,如数据拷贝、上下文转换、内存分配、锁竞争等,同时提出了针对上述问题所要遵循的原则。
在网络服务器模型中,有线程驱动模型和事情驱动模型。虽然线程是表示并发的主要方式,但是在线程驱动模型服务器中,随着线程数目不断增加,系统的开销不断增长,到一定程度后会导致系统总体性能的下降;而事件驱动模型网络服务器用一个线程处理所有任务,避免了同步引起的很多问题,但是事件驱动模型不能利用SMP来提高性能,并且事件处理线程有可能阻塞。
基于线程驱动模型和事件驱动模型的特点,本文提出了网络服务器设计框架混合模型,充分利用了两种模型的优点。最后归纳了如何运用Wrap、Pipeline、Replicate、Combine等设计模式构建基于混合模型的流媒体网络服务器。最后阐述了CDN系统的结构以及其中的关键技术。
With the rapid development of Internet, streaming media system expands it's use continuously, so it requires the rock-bottom computing platform higher performance—more powerful computing performance 、 more high speed data transmission、 more formidable heat dissipation ability and more strong system management ability. The extant computing platform cannot meet these requirements. But ATCA (Advanced Telecom Computing Architecture) has the obvious superiority in all of computing performance, the ability of network transmission, management performance as well as the private characteristic design aspect etc. ATCA is a series of standards PICMG3.X proposed by PICMG (PCI Industrial Computer Manufacturers Group) ,is open standard for high performance telecommunication and industry system development, so it is very suitable using in steaming media server system.This paper elaborates streaming media system architecture based on ATCA, analyzes the system composition, and defines each module function. The system is highly expandable. Also, it has uniform management platform, flexible content distribution system, widespread support of hardware, application and content, rich interface and so on.Network server design is one of the most important key in implementing a streaming media system. In design the network server, some essential questions are summarized, like data copy, context switches, the memory Allocation, lock contention and so on simultaneously proposes some principles in view of above questions.Threaded server and event-driver server is the two most general network servers. Threads have become the dominant form of expressing concurrency, but in threaded server, as the number of threads in a system grows, operating
system overhead increases, leading to a decrease in the overall performance of the system. Event-driven systems are typically structured as a program thread that loops continuously, processing events of different types from a queue. Event-driven programming avoids many of the bugs associated with synchronization. However, Events generally cannot take advantage of SMP systems for performance. Also, event processing threads can block regardless of the I/O mechanisms used.According to Threaded Servers and Event-drive Servers' characteristics, a hybrid thread and event server is proposed to obtain both benefit of threads and events. The paper also induces how to use design rules such as Wrap, Pipeline, Combine and Replicate to design streaming media network server. At last, it describes the CDN system architecture and looks into the key technologies.
引文
[1] W. Richard Stevens, TCP/IP Illustrated, Volume 1: The Protocols. Addison-Wesley
[2] Jon C.Snader,高级TCP/IP编程.中国电力出版社
[3] DOUGLASE.COMER,DAVID.L.STEVENS,用TCP/IP进行网际互联第三卷:客户—服务器编程与应用(Linux/POSIX套接字版).电子工业出版社
[4] Eyal Menin, The Streaming Media Handbook. Pearson Education
[5] Bing Wang, Jim Kurose, Multimedia streaming via TCP: an analytic performance study. ACM Press
[6] R. Rejaie, M. Handley, and D. Estrin, "Quality adaptation for congestion controlled video playback over the Internet," in SIGCOMM, pp. 189-200, September 1999.
[7] J. vander Merwe, S. Sen, and C. Kalmanek, "Streaming video traffic: Characterization and network impact," in Proceedings of the Seventh International Web Content Caching and Distribution Workshop, August 2002.
[8] C. Krasic and J. Walpole, "Priority-progress streaming for quality-adaptive multimedia," in ACM Multimedia Doctoral Symposium 2001, (Ottawa, Canada), October 2001.
[9] M. Li, M. Claypool, R. Kinicki, and J. Nichols, "Characteristics of streaming media stored on the Internet," Tech. Rep. WPI-CS-TR-03-18, CS Department, Worcester Polytechnic Institute, May 2003.
[10] S. Bohacek, "A stochastic model of TOP and fair video transmission," in Proc. IEEE INFOCOM, 2003.
[11] O. Verscheure, P. Frossard, and M. Hamdi, "MPEG-2 video services over packet networks: Joint effect of encoding rate and data loss on user-oriented QoS," in Proc. of NOSSDAV, July 1998.
[12] W. Richard Stevens, UNIX Network Programming Volume 1 Networking APIs: sockets and XTI(second edition). 清华大学出版社, 2000
[13] W.Richard Stevens,UNIX网络编程[M].杨继张译.北京:清华大学出版社,2000.
[14] 华为公司,以运营商为核心打造IPTV价值链,中国电信业,2005 No.4
[15] 张江、吴南山、杨士强,IPTV及其DRM需求,现代电视技术,2005 No.4
[16] 汪卫国,国外电信运营商IPTVA务发展分析,世界电信,2005 Vol.18 No.3。
[17] 王荣生,RTSP协议在视频点播系统中的应用,计算机应用与软件,2004 Vol.21 No.11
[18] Zhuang Guo, Schaeffer, J. Szafron, D. Earl, Using Generative Design Patterns to Develop Network Server Applications. Parallel and Distributed Processing Symposium, 2005. Proceedings. 19th IEEE International.
[19] Carrera, E. V. Bianchini, R. Evaluating, cluster-based network servers. High-Performance Distributed Computing, 2000. Proceedings.
[20] Lui Sha, Xue Liu, Ying Lu, Abdelzaher, Queueing model based network server performance control. Real-Time Systems Symposium, 2002. RTSS 2002. 23rd IEEE.
[21] Fineberg, k practical architecture for implementing end-to-end QoS in an IP network. Communications Magazine, IEEE. Volume 40, Issue 1, Jan. 2002 Page(s): 122-130.
[22] Liabotis, I. Garton, D. Prnjat, O. Sacks, Resource monitoring on an application layer active network server. Network Operations and Management Symposium, 2002. NOMS 2002. 2002 IEEE/IFIP 15-19 April 2002 Page(s): 911-913
[23] Cranor, C. D. Green, M. Kalmanek, C. Shur, D. Sibal, S. Vander Merwe, J. E. Sreenan, Enhanced streaming services in a content distribution network. Internet Computing, IEEE, Volume 5, Issue 4, July-Aug. 2001 Page(s): 66-75.
[24] Zhiwen Xu, Xiaoxin Guo, Yunjie Pang, Zhengxuan Wang, The powered method of proxy cache for streaming media. Communications, Circuits and Systems, 2004. ICCCAS 2004. 2004 International Conference on Volume 1, 27-29 June 2004 Page(s):507-511 Vol. 1.
[25] Shudong Jin, Bestavros, A. Iyengar, Accelerating Internet streaming media delivery using network-aware partial caching. Distributed Computing Systems, 2002. Proceedings. 22nd International Conference on 2-5 July 2002 Page(s): 153-160
[26] Hong, D. P. Albuquerque, C. Oliveira, C. Suda, T. Evaluating the impact of emerging streaming media applications on TCP/IP performance Communications Magazine, IEEE. Volume 39, Issue 4, April 2001 Page(s): 76-82.
[27] Mosri, A. F. Reed, Effects of routing algorithms on the transmission of layered streaming media. Computer Science, 2003. ENC 2003. Proceedings of the Fourth Mexican International Conference on 8-12 Sept. 2003 Page(s): 284-290.
[28] Apostolopoulos, J. G. Wai-tian, Tan Wee. S. J, Performance of a multiple description streaming media content delivery network, Image Processing. 2002. Proceedings. 2002 International Conference on Volume 2, 22-25 Sept. 2002 Page(s): Ⅱ-189-Ⅱ-192 vol. 2.
[29] Pieper, J. Srinivasan, S. Dom, Streaming-media knowledge discovery. Computer. Volume 34, Issue 9, Sept. 2001 Page(s): 68-74.
[30] Hillestad, O. I. Libak, B. Perkis, Performance Evaluation of Multimedia Services Over IP Networks. Multimedia and Expo, 2005. ICME 2005. IEEE International Conference on 06-06 July 2005 Page(s): 1464-1467.
[31] 彭湃.计算机网络教程.机械工业出版社.2001
[32] 杨众,王钢.TCP/IP协议问题透析.计算机应用研究.2001
[33] 宽带内容平台技术基础.http://www.btxx.cn.net/chinatc/200203/20020302.htm, 2002. 6
[34] 张旭军.CDN:网络提速又一招. http://www.pcworld.com.cn/2001/back_issues/2132b.asp, 2001. 1
[35] Server Load Balancing. http://www.extremenetworks.com/Server_load_balancing.pdf.2004. 1
[36] Gloabal Server Load Balancing with Servering. http://www.foundrynet.com/solutions/gslb.pdf.2004. 2
[37] 尤晋.在宽带网中应用高速缓存.网络世界,2001.3.19.
[2] Jon C.Snader,高级TCP/IP编程.中国电力出版社
[3] DOUGLASE.COMER,DAVID.L.STEVENS,用TCP/IP进行网际互联第三卷:客户—服务器编程与应用(Linux/POSIX套接字版).电子工业出版社
[4] Eyal Menin, The Streaming Media Handbook. Pearson Education
[5] Bing Wang, Jim Kurose, Multimedia streaming via TCP: an analytic performance study. ACM Press
[6] R. Rejaie, M. Handley, and D. Estrin, "Quality adaptation for congestion controlled video playback over the Internet," in SIGCOMM, pp. 189-200, September 1999.
[7] J. vander Merwe, S. Sen, and C. Kalmanek, "Streaming video traffic: Characterization and network impact," in Proceedings of the Seventh International Web Content Caching and Distribution Workshop, August 2002.
[8] C. Krasic and J. Walpole, "Priority-progress streaming for quality-adaptive multimedia," in ACM Multimedia Doctoral Symposium 2001, (Ottawa, Canada), October 2001.
[9] M. Li, M. Claypool, R. Kinicki, and J. Nichols, "Characteristics of streaming media stored on the Internet," Tech. Rep. WPI-CS-TR-03-18, CS Department, Worcester Polytechnic Institute, May 2003.
[10] S. Bohacek, "A stochastic model of TOP and fair video transmission," in Proc. IEEE INFOCOM, 2003.
[11] O. Verscheure, P. Frossard, and M. Hamdi, "MPEG-2 video services over packet networks: Joint effect of encoding rate and data loss on user-oriented QoS," in Proc. of NOSSDAV, July 1998.
[12] W. Richard Stevens, UNIX Network Programming Volume 1 Networking APIs: sockets and XTI(second edition). 清华大学出版社, 2000
[13] W.Richard Stevens,UNIX网络编程[M].杨继张译.北京:清华大学出版社,2000.
[14] 华为公司,以运营商为核心打造IPTV价值链,中国电信业,2005 No.4
[15] 张江、吴南山、杨士强,IPTV及其DRM需求,现代电视技术,2005 No.4
[16] 汪卫国,国外电信运营商IPTVA务发展分析,世界电信,2005 Vol.18 No.3。
[17] 王荣生,RTSP协议在视频点播系统中的应用,计算机应用与软件,2004 Vol.21 No.11
[18] Zhuang Guo, Schaeffer, J. Szafron, D. Earl, Using Generative Design Patterns to Develop Network Server Applications. Parallel and Distributed Processing Symposium, 2005. Proceedings. 19th IEEE International.
[19] Carrera, E. V. Bianchini, R. Evaluating, cluster-based network servers. High-Performance Distributed Computing, 2000. Proceedings.
[20] Lui Sha, Xue Liu, Ying Lu, Abdelzaher, Queueing model based network server performance control. Real-Time Systems Symposium, 2002. RTSS 2002. 23rd IEEE.
[21] Fineberg, k practical architecture for implementing end-to-end QoS in an IP network. Communications Magazine, IEEE. Volume 40, Issue 1, Jan. 2002 Page(s): 122-130.
[22] Liabotis, I. Garton, D. Prnjat, O. Sacks, Resource monitoring on an application layer active network server. Network Operations and Management Symposium, 2002. NOMS 2002. 2002 IEEE/IFIP 15-19 April 2002 Page(s): 911-913
[23] Cranor, C. D. Green, M. Kalmanek, C. Shur, D. Sibal, S. Vander Merwe, J. E. Sreenan, Enhanced streaming services in a content distribution network. Internet Computing, IEEE, Volume 5, Issue 4, July-Aug. 2001 Page(s): 66-75.
[24] Zhiwen Xu, Xiaoxin Guo, Yunjie Pang, Zhengxuan Wang, The powered method of proxy cache for streaming media. Communications, Circuits and Systems, 2004. ICCCAS 2004. 2004 International Conference on Volume 1, 27-29 June 2004 Page(s):507-511 Vol. 1.
[25] Shudong Jin, Bestavros, A. Iyengar, Accelerating Internet streaming media delivery using network-aware partial caching. Distributed Computing Systems, 2002. Proceedings. 22nd International Conference on 2-5 July 2002 Page(s): 153-160
[26] Hong, D. P. Albuquerque, C. Oliveira, C. Suda, T. Evaluating the impact of emerging streaming media applications on TCP/IP performance Communications Magazine, IEEE. Volume 39, Issue 4, April 2001 Page(s): 76-82.
[27] Mosri, A. F. Reed, Effects of routing algorithms on the transmission of layered streaming media. Computer Science, 2003. ENC 2003. Proceedings of the Fourth Mexican International Conference on 8-12 Sept. 2003 Page(s): 284-290.
[28] Apostolopoulos, J. G. Wai-tian, Tan Wee. S. J, Performance of a multiple description streaming media content delivery network, Image Processing. 2002. Proceedings. 2002 International Conference on Volume 2, 22-25 Sept. 2002 Page(s): Ⅱ-189-Ⅱ-192 vol. 2.
[29] Pieper, J. Srinivasan, S. Dom, Streaming-media knowledge discovery. Computer. Volume 34, Issue 9, Sept. 2001 Page(s): 68-74.
[30] Hillestad, O. I. Libak, B. Perkis, Performance Evaluation of Multimedia Services Over IP Networks. Multimedia and Expo, 2005. ICME 2005. IEEE International Conference on 06-06 July 2005 Page(s): 1464-1467.
[31] 彭湃.计算机网络教程.机械工业出版社.2001
[32] 杨众,王钢.TCP/IP协议问题透析.计算机应用研究.2001
[33] 宽带内容平台技术基础.http://www.btxx.cn.net/chinatc/200203/20020302.htm, 2002. 6
[34] 张旭军.CDN:网络提速又一招. http://www.pcworld.com.cn/2001/back_issues/2132b.asp, 2001. 1
[35] Server Load Balancing. http://www.extremenetworks.com/Server_load_balancing.pdf.2004. 1
[36] Gloabal Server Load Balancing with Servering. http://www.foundrynet.com/solutions/gslb.pdf.2004. 2
[37] 尤晋.在宽带网中应用高速缓存.网络世界,2001.3.19.