基于QoS的区分服务模型及管理机制研究
摘要
随着Internet规模的迅速扩大,以及多媒体技术的蓬勃发展,QoS已经成为下一代互联网正常运行的关键问题。随之引发了IntServ、DiffServ、CBR以及MPLS-TE等一系列新技术的出现。作为QoS的实现模型,DiffServ凭借其简单性和可扩展的特点,成为目前实现IP QoS的首选方案和研究热点。其“边缘分类,内部转发”的思想使人们走出了IntServ/RSVP模型的困境,并为解决IP QoS问题提供了新的思路。
DiffServ模型的实现是由一系列具体机制来保障的。其中作用于系统核心路由器的队列管理机制和队列调度机制表面上是按照不同的PHB类别对数据分组进行入队管理与调度转发的,但从本质上体现的是对网络资源的管理与配置,这两种机制被认为是实现IP QoS的核心问题。因此,上述两种机制的研究对于IP QoS具有重要的理论意义和实践价值。
本文的主要工作如下:
1、对DiffServ模型进行了深入的研究,从技术角度分析了DiffServ与IntServ模型以及MPLS模型相结合的实施方案。
2、分析了以DropTail为代表的PQM(Passive Queue Management)类算法的缺点以及当前几种典型的AQM(Active Queue Management)类算法如RED、RIO-C的不足之处,并在此基础之上提出了一种新的AQM算法——PFRIO。该算法旨在解决同一PHB组中,具有不同丢弃优先级的各类数据分组之间由于共享缓冲队列而引发的带宽分配的公平性问题。理论分析和仿真实验验证了该算法的可行性和有效性。
3、通过对多种队列调度算法的分析和比较,选出WRR算法作为DiffServ模型中保证QoS方案的队列调度算法。
4、在考虑QoS保证的基础上,本文提出能同时实现EF、AF、BE多种PHB的一种DiffServ的架构模型,仿真实验证明该架构模型实现QoS保证的有效性。
Along with the rapid expansion of the Internet, as well as the vigorous development of multimedia technical, QoS already became the key question to the normal operation of next generation Internet. Along with it has initiated a series of new technical appearance, like IntServ, DiffServ, CBR as well as MPLS-TE and so on. As the QoS realization model, DiffServ relies on its simplicity and the expandable characteristic, becomes to first choice plan and the research hot spot to realize the IP QoS at present.It’s thought "the edge classification, the interior retransmits" helps people go out the IntServ/RSVP model’s difficult position, and provide a new mentality for solved IP QoS question .
The realization of DiffServ model is safeguard by a series of concrete mechanisms. Queue management and queue scheduling mechanism that function to the system core router make scheduling management and retransmits on data packets according to Different PHB category, but essentially manifests the management and the disposition to the network resources, is the core question for realizes the IP QoS. Therefore, made research to the above two mechanisms have the important theory significance and the practice value for IP QoS.
The main work of the thesis as follows:
1.Made a thorough research to the DiffServ model, and analyzed the implementation plan about the unify of the DiffServ model and the IntServ model as well as the unify of the DiffServ model and the MPLS model from the technical angle.
2.Made an analysis to the shortcomings of the PQM algorithm represent by DropTail as well as the deficiency of several typical AQM algorithms currently like RED, RIO-C, and proposed a new AQM algorithm—PFRIO. This algorithm is for the purpose of solving band width assignment fairness question which initiates by date packets in an identical PHB group but has different priority that sharing buffer queue. And confirmed the feasibility and the validity of the algorithm through the theoretical analysis and the simulation experiment.
3. Through analysis and the comparison to many kinds of queue scheduling algorithm, selects the WRR as the scheduling algorithm that guaranteed the QoS plan in the DiffServ model .
4.Considering the QoS guarantee, this thesis article proposed a DiffServ overhead construction model that can simultaneously realize EF, AF, BE PHB and proved this overhead construction model realizes the QoS guarantee validity through the simulation experiment.
DiffServ模型的实现是由一系列具体机制来保障的。其中作用于系统核心路由器的队列管理机制和队列调度机制表面上是按照不同的PHB类别对数据分组进行入队管理与调度转发的,但从本质上体现的是对网络资源的管理与配置,这两种机制被认为是实现IP QoS的核心问题。因此,上述两种机制的研究对于IP QoS具有重要的理论意义和实践价值。
本文的主要工作如下:
1、对DiffServ模型进行了深入的研究,从技术角度分析了DiffServ与IntServ模型以及MPLS模型相结合的实施方案。
2、分析了以DropTail为代表的PQM(Passive Queue Management)类算法的缺点以及当前几种典型的AQM(Active Queue Management)类算法如RED、RIO-C的不足之处,并在此基础之上提出了一种新的AQM算法——PFRIO。该算法旨在解决同一PHB组中,具有不同丢弃优先级的各类数据分组之间由于共享缓冲队列而引发的带宽分配的公平性问题。理论分析和仿真实验验证了该算法的可行性和有效性。
3、通过对多种队列调度算法的分析和比较,选出WRR算法作为DiffServ模型中保证QoS方案的队列调度算法。
4、在考虑QoS保证的基础上,本文提出能同时实现EF、AF、BE多种PHB的一种DiffServ的架构模型,仿真实验证明该架构模型实现QoS保证的有效性。
Along with the rapid expansion of the Internet, as well as the vigorous development of multimedia technical, QoS already became the key question to the normal operation of next generation Internet. Along with it has initiated a series of new technical appearance, like IntServ, DiffServ, CBR as well as MPLS-TE and so on. As the QoS realization model, DiffServ relies on its simplicity and the expandable characteristic, becomes to first choice plan and the research hot spot to realize the IP QoS at present.It’s thought "the edge classification, the interior retransmits" helps people go out the IntServ/RSVP model’s difficult position, and provide a new mentality for solved IP QoS question .
The realization of DiffServ model is safeguard by a series of concrete mechanisms. Queue management and queue scheduling mechanism that function to the system core router make scheduling management and retransmits on data packets according to Different PHB category, but essentially manifests the management and the disposition to the network resources, is the core question for realizes the IP QoS. Therefore, made research to the above two mechanisms have the important theory significance and the practice value for IP QoS.
The main work of the thesis as follows:
1.Made a thorough research to the DiffServ model, and analyzed the implementation plan about the unify of the DiffServ model and the IntServ model as well as the unify of the DiffServ model and the MPLS model from the technical angle.
2.Made an analysis to the shortcomings of the PQM algorithm represent by DropTail as well as the deficiency of several typical AQM algorithms currently like RED, RIO-C, and proposed a new AQM algorithm—PFRIO. This algorithm is for the purpose of solving band width assignment fairness question which initiates by date packets in an identical PHB group but has different priority that sharing buffer queue. And confirmed the feasibility and the validity of the algorithm through the theoretical analysis and the simulation experiment.
3. Through analysis and the comparison to many kinds of queue scheduling algorithm, selects the WRR as the scheduling algorithm that guaranteed the QoS plan in the DiffServ model .
4.Considering the QoS guarantee, this thesis article proposed a DiffServ overhead construction model that can simultaneously realize EF, AF, BE PHB and proved this overhead construction model realizes the QoS guarantee validity through the simulation experiment.
引文
[1] Crawley E, Nair R,Rajagopalan B,et al.A framework for QoS-based routing in the internet[J].IETF Internet RFC 2386,1998.
[2] Shenker S,Partridge C, R.Guerin.Specification of Guaranteed Quality of Service[J] RFC2212,September 1997.
[3] 林闯,单志广,盛立杰等.区分服务及其几个热点问题的研究[J].计算机学报,2000,23(4):419–433.
[4] Braden B,Clark D,Crowcroft J.Recommendation on queue management and congestion avoidance in the Internet [J].IETF Internet RFC 2309,1998.
[5] Sally Floyd . TCP and explicit congestion notification [J] . ACM Computer Communication Review,1994,24(5):10–23.
[6] Floyd S,Jacobson V.Random early detection gateways for congestion avoidance [J].IEEE/ACM Transactions on Networking,1993,1(4):397–413
[7] Bitorika A,Robin M,Huggard M.A survey of active queue management schemes [J].Tech.Rep,Sept,2003.
[8] S.Shenker,C.Partridge,R.Guerin.Specification of Guaranteed Quality of Service [J],IETF RFC 2212,Sept,1997.
[9] J.Wroclawski.Specification of the Controlled - Load Network Element Service [J],IETF Internet RFC 2211,Sept,1997.
[10] 林闯,单志广.计算网络的服务质量,第一版[M].清华大学出版社,2004.4:50–133.
[11] 谢臻,李翔.IP 分组网中基于 Difsery 的 QOS 技术策略及部署 [J].计算机与现代化,2006.5,第 5 期:97–100.
[12] S Brim,B Carpenter , F Le Faucheur.Per hop behavior identification codes [S].IETF Internet RFC 2836,2000.
[13] Ramakrishan K,Floyd S. A proposal to add explicit congestion notification(ECN) to IP [S].IETF Internet RFC 2481, January 1999.
[14] Jacobson V,Nichols K and K Poduri.An Expedited Forwarding PHB [S].IETF Internet RFC 2598,1999.
[15] J Heinanen,F Baker,W Weiss,J Wroclawski.Assured Forwarding PHB Group [S].IETF Internet RFC 2597,1999.
[16] Nichols K, Blake S, Baker F, et al. Definition of the differentiated services field (DS field) in the IPv4 and IPv6 header [S].IETF Internet RFC 2474, 1998.
[17] Braden R,Clark D,Shenker S.Integrated Services in the Internet Architecture:An Overview [J].IETF Internet RFC 1633,1994.
[18] Bernet Y, Ford P,et al. A framework for Integrated Services Operation over DiffServ Networks [J].IETF Internet RFC 2998, November 2000.
[19] E.Rosenetal. Multiprotocol Label Switehing Architecture [J].IETF Internet RFC 3031,January 2001.
[20] F.Le Faucheur,W.Lai. Requirements for support of Differentiated Services–aware MPLS Traffic Engineering [J].IETF RFC3564,July 2003.
[21] 曲延光,刘云超.Internet 主动队列管理算法研究 [J].计算机应用,2003,28(10):36–38.
[22] D.Lin,R.Morris.Dynamics of random early detection [C].ACM SIGCOMM’97,1997:127–137.
[23] J.O.Teunis,T.V.Lakshman,H.W.Larry.SRED:Stabilized RED [C].In Proc.IEEE INFOCOM’99.New York,USA,1999:1346–1355.
[24] S.Floyd,R.Gummadi,S.Shenker.Adaptive RED:An algorithm for increasing the robustness of RED’s active queue management [J].ACIRI Technical Report,2001.
[25] R.Pan ,B.Probhakar,K.Psounis.CHOKe:A stateless active queue management scheme for approximating fair bandwidth allocation [C].IEEE INFOCOM’00,vol.2,Tel Aviv,Israel,March 2000:942-951.
[26] D.Clark, W. Fang. Explicit allocation of best-effort packet delivery service.IEEE/ACM Trans on Networking [C], Vol.6,No.4, Aug 1998:362-373.
[27] C.Hollot,V.Misra,D.Towsley,W.Gong.Analysis and design of controllers for AQM routers supporting TCP flows [C].IEEE Transactions on automatic control,2002 47(6): 945-959.
[28] 章淼,吴建平,林闯.P2I:一种新的主动队列管理算法 [J].计算机学报,2008,26(10):1288-1294.
[29] H.Zhang,B.Liu,W.Dou.Design of a robust active queue management algorithm based on feedback compensation [C].ACM SIGCOMM 2003,2003:277-285.
[30] Hashem E.Analysis of random drop for gateway congestion control [C].Report LCS TR-465,Laboratory for Computer Science,MIT,Cambridge,MA,1989:103.
[31] S.Floyd.NS network simulator [EB].http://www.mash.cs.berkeley.edu/ns/.
[32] 徐雷鸣 ,庞博等.NS 与网络模拟 [M].北京:人民邮电出版社,2003.
[33] 高文宇. 网络 QoS 控制中的队列管理和接纳控制研究 [D].中南大学博士学位论文,2005 :13-15.
[34] A.K.Parekh,R.G.Gallager.A Generalized Processor Sharing approach to flow control in integrated services networks:The single-node case[C].IEEE/ACM Transactions on networking,1993,1(3):344-357.
[35] A.K.Parekh,R.G.Gallager.A Generalized Processor Sharing approach to flow control in integrated services networks:The multiple-node case[C].IEEE/ACM Transactions on networking,1994,2(2):137-150.
[36] A.Demers,S.Keshav,S.Shenker.Analysis and simulation of a fair queuing algorithm [C].ACM SIGCOMM’89,1989.
[37] L.Zhang.Virtual clock: A new traffic control algorithm for packet switching networks [C].ACM SIGCOMM’90,1990.
[38] M.Katevenis,S.Sidiropoulos, C.Courcoubetis.Weighted round-robin cell multiplexing in a general-poupose ATM switch chip [C].IEEE Journal on Selected Areas in Communications,Oct.1991,9(8):1265-1279.
[39] M.Shreedhar and Gvarghese.Efficient fair queueing using deficit round robin [C].IEEE/ACM Transactions on networking,June 1996,4(3):375-385.
[40] S.Golestani.A self-clocked fair queueing scheme for broadband applications [C].IEEE INFOCOM’94,1994:636-645.
[41] R.Bennett ,H.Zhang. WF2Q:worst-case fair weighted fair queueing [C].IEEEINFOCOM’96,Mar 1996 : 120-128.
[42] Y.Jia ng,C.K.Tham ,C.C.Ko.A probabilistic priority sheduling discipline for high-speed networks [C]. IEEE HPSR’01,2001:1-5.
[43] Y.Jiang,C.K.Tham, C.C.Ko.A probabilistic priority sheduling discipline for multi-service networks [C]. IEEE ISCC’01,July 2001:450-455.
[44] T.Anker,et al.Probabilistic Fair Queuing[C]. IEEE Workshop on HPSR’2001,2001:397-401.
[45] M-Zhang,et al.Probabilistic packet scheduling:achieving proportional share bandwidth allocation [C]. IEEE INFOCOM’02,New York,June 2002.
[46] C.R.Beonett,H.Zhang.Hierarchical packet fair queueing algorithms [C].IEEE/ACM Transactions on networking,Oct 1997,5(5):675-689.
[47] O.Altinatas,dal.Urgency-based round robin:A new scheduling discipline for multiservice packet switching networks [C]. IEICE Transactions on Communication,Nov 2001 ,E81-B(11):2013-2022.
[2] Shenker S,Partridge C, R.Guerin.Specification of Guaranteed Quality of Service[J] RFC2212,September 1997.
[3] 林闯,单志广,盛立杰等.区分服务及其几个热点问题的研究[J].计算机学报,2000,23(4):419–433.
[4] Braden B,Clark D,Crowcroft J.Recommendation on queue management and congestion avoidance in the Internet [J].IETF Internet RFC 2309,1998.
[5] Sally Floyd . TCP and explicit congestion notification [J] . ACM Computer Communication Review,1994,24(5):10–23.
[6] Floyd S,Jacobson V.Random early detection gateways for congestion avoidance [J].IEEE/ACM Transactions on Networking,1993,1(4):397–413
[7] Bitorika A,Robin M,Huggard M.A survey of active queue management schemes [J].Tech.Rep,Sept,2003.
[8] S.Shenker,C.Partridge,R.Guerin.Specification of Guaranteed Quality of Service [J],IETF RFC 2212,Sept,1997.
[9] J.Wroclawski.Specification of the Controlled - Load Network Element Service [J],IETF Internet RFC 2211,Sept,1997.
[10] 林闯,单志广.计算网络的服务质量,第一版[M].清华大学出版社,2004.4:50–133.
[11] 谢臻,李翔.IP 分组网中基于 Difsery 的 QOS 技术策略及部署 [J].计算机与现代化,2006.5,第 5 期:97–100.
[12] S Brim,B Carpenter , F Le Faucheur.Per hop behavior identification codes [S].IETF Internet RFC 2836,2000.
[13] Ramakrishan K,Floyd S. A proposal to add explicit congestion notification(ECN) to IP [S].IETF Internet RFC 2481, January 1999.
[14] Jacobson V,Nichols K and K Poduri.An Expedited Forwarding PHB [S].IETF Internet RFC 2598,1999.
[15] J Heinanen,F Baker,W Weiss,J Wroclawski.Assured Forwarding PHB Group [S].IETF Internet RFC 2597,1999.
[16] Nichols K, Blake S, Baker F, et al. Definition of the differentiated services field (DS field) in the IPv4 and IPv6 header [S].IETF Internet RFC 2474, 1998.
[17] Braden R,Clark D,Shenker S.Integrated Services in the Internet Architecture:An Overview [J].IETF Internet RFC 1633,1994.
[18] Bernet Y, Ford P,et al. A framework for Integrated Services Operation over DiffServ Networks [J].IETF Internet RFC 2998, November 2000.
[19] E.Rosenetal. Multiprotocol Label Switehing Architecture [J].IETF Internet RFC 3031,January 2001.
[20] F.Le Faucheur,W.Lai. Requirements for support of Differentiated Services–aware MPLS Traffic Engineering [J].IETF RFC3564,July 2003.
[21] 曲延光,刘云超.Internet 主动队列管理算法研究 [J].计算机应用,2003,28(10):36–38.
[22] D.Lin,R.Morris.Dynamics of random early detection [C].ACM SIGCOMM’97,1997:127–137.
[23] J.O.Teunis,T.V.Lakshman,H.W.Larry.SRED:Stabilized RED [C].In Proc.IEEE INFOCOM’99.New York,USA,1999:1346–1355.
[24] S.Floyd,R.Gummadi,S.Shenker.Adaptive RED:An algorithm for increasing the robustness of RED’s active queue management [J].ACIRI Technical Report,2001.
[25] R.Pan ,B.Probhakar,K.Psounis.CHOKe:A stateless active queue management scheme for approximating fair bandwidth allocation [C].IEEE INFOCOM’00,vol.2,Tel Aviv,Israel,March 2000:942-951.
[26] D.Clark, W. Fang. Explicit allocation of best-effort packet delivery service.IEEE/ACM Trans on Networking [C], Vol.6,No.4, Aug 1998:362-373.
[27] C.Hollot,V.Misra,D.Towsley,W.Gong.Analysis and design of controllers for AQM routers supporting TCP flows [C].IEEE Transactions on automatic control,2002 47(6): 945-959.
[28] 章淼,吴建平,林闯.P2I:一种新的主动队列管理算法 [J].计算机学报,2008,26(10):1288-1294.
[29] H.Zhang,B.Liu,W.Dou.Design of a robust active queue management algorithm based on feedback compensation [C].ACM SIGCOMM 2003,2003:277-285.
[30] Hashem E.Analysis of random drop for gateway congestion control [C].Report LCS TR-465,Laboratory for Computer Science,MIT,Cambridge,MA,1989:103.
[31] S.Floyd.NS network simulator [EB].http://www.mash.cs.berkeley.edu/ns/.
[32] 徐雷鸣 ,庞博等.NS 与网络模拟 [M].北京:人民邮电出版社,2003.
[33] 高文宇. 网络 QoS 控制中的队列管理和接纳控制研究 [D].中南大学博士学位论文,2005 :13-15.
[34] A.K.Parekh,R.G.Gallager.A Generalized Processor Sharing approach to flow control in integrated services networks:The single-node case[C].IEEE/ACM Transactions on networking,1993,1(3):344-357.
[35] A.K.Parekh,R.G.Gallager.A Generalized Processor Sharing approach to flow control in integrated services networks:The multiple-node case[C].IEEE/ACM Transactions on networking,1994,2(2):137-150.
[36] A.Demers,S.Keshav,S.Shenker.Analysis and simulation of a fair queuing algorithm [C].ACM SIGCOMM’89,1989.
[37] L.Zhang.Virtual clock: A new traffic control algorithm for packet switching networks [C].ACM SIGCOMM’90,1990.
[38] M.Katevenis,S.Sidiropoulos, C.Courcoubetis.Weighted round-robin cell multiplexing in a general-poupose ATM switch chip [C].IEEE Journal on Selected Areas in Communications,Oct.1991,9(8):1265-1279.
[39] M.Shreedhar and Gvarghese.Efficient fair queueing using deficit round robin [C].IEEE/ACM Transactions on networking,June 1996,4(3):375-385.
[40] S.Golestani.A self-clocked fair queueing scheme for broadband applications [C].IEEE INFOCOM’94,1994:636-645.
[41] R.Bennett ,H.Zhang. WF2Q:worst-case fair weighted fair queueing [C].IEEEINFOCOM’96,Mar 1996 : 120-128.
[42] Y.Jia ng,C.K.Tham ,C.C.Ko.A probabilistic priority sheduling discipline for high-speed networks [C]. IEEE HPSR’01,2001:1-5.
[43] Y.Jiang,C.K.Tham, C.C.Ko.A probabilistic priority sheduling discipline for multi-service networks [C]. IEEE ISCC’01,July 2001:450-455.
[44] T.Anker,et al.Probabilistic Fair Queuing[C]. IEEE Workshop on HPSR’2001,2001:397-401.
[45] M-Zhang,et al.Probabilistic packet scheduling:achieving proportional share bandwidth allocation [C]. IEEE INFOCOM’02,New York,June 2002.
[46] C.R.Beonett,H.Zhang.Hierarchical packet fair queueing algorithms [C].IEEE/ACM Transactions on networking,Oct 1997,5(5):675-689.
[47] O.Altinatas,dal.Urgency-based round robin:A new scheduling discipline for multiservice packet switching networks [C]. IEICE Transactions on Communication,Nov 2001 ,E81-B(11):2013-2022.