一种新的基于预约的拥塞避免机制

摘要/Abstract

摘要：

由于高速互连网络上的负载不均衡，一些网络结点成为了热点，可能导致部分结点或是链路拥塞，这会极大地降低互连网络的性能。现有的基于预约的拥塞避免技术SRP可以进行主动的拥塞避免，极大地缓解了由于热点问题所带来的负面效应。但是，在热点模式下，其它非热点结点的路由器资源绝大多数处于空闲状态，为了进一步充分利用互连网络的资源，提升互连网络性能，提出了一种基于SRP改进的中间结点缓存技术IRP。IRP可以根据不同的拓扑，例如胖树，有效地利用热点的邻居结点的路由器资源，先利用胖树拓扑的多路径将报文发送给空闲路由器,一旦目的结点路由器可利用,则将缓存报文发送给目的结点，降低互连网络的延迟。

关键词: 热点, 拥塞, 路由器, SRP, IRP

Abstract:

Because of the unbalanced traffic in high speed connected networks, some nodes in the networks become hotspots, resulting in congestion in these nodes and related channels , which can harm the performance of high speed connected networks greatly. There is an existing congestion avoidance scheme named Speculative Reservation Protocol (SRP) which can avoid congestion actively, and the SRP eliminates the negative effect induced by hotspots problem tremendously. But in the hotspots model, the resources in most of routers which don't connect any hotspot are idle. In order to fully use the resources and enhance the performance of high speed connected networks, we put forward a novel congestion avoidance scheme called Intermediate Reservation Protocol (IRP) based on reservation. The IRP can effectively use the resources of the routers of neighbor nodes judged by different topologies, such as Fattree. The IRP first sends the packets to the idle routers via the multipath of Fattree. When the destination's router is available, the IRP packets can be sent to the destination nodes, and the latency of high speed connected networks is thus reduced.

Key words: hot-spots;congestion;router;SRP;IRP

朱成阳，柴燕涛，董德尊，张鹤颖，庞征斌. 一种新的基于预约的拥塞避免机制[J]. J4, 2016, 38(02): 240-248.

ZHU Chengyang，CHAI Yantao，DONG Dezun，ZHANG Heying,PANG Zhengbin. A novel congestion avoidance scheme based on reservation [J]. J4, 2016, 38(02): 240-248.

参考文献［31］

［1］	Quiles F J, Johnson I, Naven F. Efficient, scalable congestion management for interconnection networks［J］. IEEE Micro, 2006 (5):5266.
［2］	Pfister G F, Norton V A. “Hot spot” contention and combining in multistage interconnection networks［J］. IEEE Transactions on Computers, 1985, 100(10):943948.
［3］	Brakmo L S, Peterson L L. TCP Vegas:End to end congestion avoidance on a global Internet［J］. IEEE Journal on Selected Areas in Communications, 1995, 13(8):14651480.
［4］	Jain R. A delaybased approach for congestion avoidance in interconnected heterogeneous computer networks［J］. ACM SIGCOMM Computer Communication Review, 1989, 19(5):5671.
［5］	Scott S L, Thorson G. The Cray T3E network:Adaptive routing in a high performance 3D torus［J］. Hot Interconnects IV, 1996(8):157160.
［6］	Adiga N R, Almasi G S, Aridor Y, et al. An overview of the BlueGene/L supercomputer［C］∥Proc of ACM/IEEE 2002 Conference on Supercomputing, 2002:60.
［7］	Myrinet 2000 Series Networking［EB/OL］. ［20000205］.http:∥www.cspi.com/multicomputer/products/2000_series_networking/2000_networking.htm.
［8］	Quadrics QsNet［EB/OL］. ［20150309］.http:∥doc.quadrics.com.
［9］	InfiniBand Trade Association. InfiniBand Architecture［EB/OL］. Specification Volume 1. Release 1.0.［20130402］.http:∥www.infinibandta.com/.
［10］	Yang C Q, Reddy A V S. A taxonomy for congestion control algorithms in packet switching networks［J］.IEEE Network, 1995, 9(4):3445.
［11］	Ramakrishnan K, Floyd S, Black D. The addition of explicit congestion notification (ECN) to IP:IETF 2001:RFC 3168［S］.2001.
［12］	Ferrer J L L, Baydal E, Robles A, et al. Congestion management in mins through marked and validated packets［C］∥Proc of the 15th Euromicro International Conference on Parallel, Distributed and NetworkBased Processing, 2007:254261.
［13］	Ferrer J L, Baydal E, Robles A, et al. A scalable and early congestion management mechanism formins［C］∥Proc of 2010 18th Euromicro International Conference on Parallel, Distributed and NetworkBased Processing (PDP), 2010:4350.
［14］	Santos J R, Turner Y, Janakiraman G. Endtoend congestion control for Infiniband［C］∥Proc of IEEE 22nd Annual Joint Conference of Computer and Communications, 2003:11231133.
［15］	Jiang N, Becker D U, Michelogiannakis G, et al. Network congestion avoidance through speculative reservation［C］∥Proc of 2012 IEEE 18th International Symposium on High Performance Computer Architecture (HPCA), 2012:112.
［16］	Zarifis K, Miao R, Calder M, et al. DIBS:Justintime congestion mitigation for data centers［C］∥Proc of the ACM 9th European Conference on Computer Systems, 2014:6.
［17］	García P J, Flich J, Duato J, et al. Dynamic evolution of congestion trees:Analysis and impact on switch architecture［M］∥High Performance Embedded Architectures and Compilers. Berlin Heidelberg:Springer, 2005:266285.
［18］	Gran E G, Eimot M, Reinemo S A, et al. First experiences with congestion control in InfiniBand hardware［C］∥Proc of 2010 IEEE International Symposium on Parallel & Distributed Processing (IPDPS), 2010:112.
［19］	Floyd S. TCP and explicit congestion notification［J］. ACM SIGCOMM Computer Communication Review, 1994, 24(5):823.
［20］	Jacobson V. Congestion avoidance and control［C］∥Proc of ACM SIGCOMM Computer Communication Review, 1988, 18(4):314329.
［21］	InfiniBand Trade Association. Infiniband architecture specification ［Z］. 1st ed. 2007.
［22］	Turner Y, Santos J R, Janakiraman G J. An approach for congestion control in InfiniBand［Z］. Palo Alto:HP Laboratories, Internet Systems and Storage Laboratory, 2001:1.
［23］	Pfister G, Gusat M, Denzel W, et al. Solving hot spot contention using Infiniband architecture congestion control［C］∥Proc of HPIDC 2005, 2005:1.
［24］	Gusat M, Craddock D, Denzel W, et al. Congestion control in InfiniBand networks［C］∥Proc of the 13th IEEE Symposium on High Performance Interconnects, 2005:158159.
［25］	Besta M, Hoefler T. Slim fly:A cost effective lowdiameter network topology［C］∥Proc of IEEE International Conference on High Performance Computing, Networking, Storage and Analysis(SC’14), 2014:348359.
［26］	Michelogiannakis G,Jiang N,Becker D,et al.Channel reservation protocol for oversubscribed channels and destinations［C］∥Proc of International Conference on High Performance Computing, Networking, Storage and Analysis, 2013:52.
［27］	Dally W J, Towles B P. Principles and practices of interconnection networks［M］. New York:Elsevier, 2004.
［28］	Scott S, Abts D, Kim J, et al. The blackwidow highradix clos network［J］. ACM SIGARCH Computer Architecture News, 2006, 34(2):1628.
［29］	Alizadeh M, Greenberg A, Maltz D A, et al. Data center tcp (dctcp)［J］. ACM SIGCOMM Computer Communication Review, 2011, 41(4):6374.
［30］	Abts D, Marty M R, Wells P M, et al. Energy proportional datacenter networks［J］. ACM SIGARCH Computer Architecture News, 2010, 38(3):338347.
［31］	Yang X J, Liao X K, Lu K, et al. The TianHe1A supercomputer:Its hardware and software［J］. Journal of Computer Science and Technology, 2011, 26(3):344351.

[1]	柴燕涛，董德尊，张鹤颖，朱成阳，廖湘科. 基于SDN架构的高性能网络拥塞避免策略[J]. J4, 20160101, 38(01): 1-10.
[2]	朱绪全, 江逸茗, 马海龙, 包婉宁, 张进. 拟态防御体系OSPF协议研究及分析[J]. 计算机工程与科学, 2023, 45(02): 204-214.
[3]	康子扬, 彭凌辉, 周干, 林博, 王蕾. 一种用于片上网络的拥塞感知哈密尔顿最短路径路由算法[J]. 计算机工程与科学, 2022, 44(06): 986-993.
[4]	段晨, 彭伟, 王宝生. 数据中心网络路由研究进展[J]. 计算机工程与科学, 2022, 44(04): 631-644.
[5]	梁崇山, 戴艺, 徐炜遐. 一种基于Chiplet集成技术的超高阶路由器设计[J]. 计算机工程与科学, 2022, 44(02): 207-213.
[6]	李树楠1,2,3，詹男杰1,2,3，章玥1,2. 面向数据中心租户带宽特征的虚拟拥塞控制方法性能分析[J]. 计算机工程与科学, 2018, 40(09): 1562-1571.
[7]	任秀江1，斯添浩1，周建毅1，谢向辉2. 基于网络包延迟偏差的硬件动态拥塞控制机制[J]. 计算机工程与科学, 2018, 40(02): 200-209.
[8]	刘培磊，唐晋韬，王挺，谢松县，岳大鹏，刘海池. 基于词向量语义聚类的微博热点挖掘方法 [J]. 计算机工程与科学, 2018, 40(02): 313-319.
[9]	路文斌，宋新亮，卢宏生，丁亚军. 一种高阶混合互连网络拓扑结构[J]. 计算机工程与科学, 2017, 39(10): 1781-1787.
[10]	杨文祥，董德尊，李存禄，雷斐，孙凯旋，吴际. 一种多级无缓存高阶路由器的设计与实现[J]. 计算机工程与科学, 2017, 39(02): 245-251.
[11]	曹杨，王宝生，张晓哲. 一种虚拟路由器资源映射算法研究[J]. 计算机工程与科学, 2016, 38(10): 1994-2000.
[12]	陈文广，牛玉刚，邹媛媛. 基于改进AOMDV路由协议的WSNs拥塞控制和能耗均衡策略[J]. 计算机工程与科学, 2016, 38(09): 1776-1783.
[13]	杨文祥，董德尊，雷斐，李存禄，吴际，孙凯旋. 高阶路由器结构研究综述[J]. 计算机工程与科学, 2016, 38(08): 1517-1523.
[14]	陈庭平，虞万荣，吴纯青. TCP传输中往返时延偏移智能响应机制研究[J]. 计算机工程与科学, 2016, 38(08): 1647-1653.
[15]	李勇1，2,蔡梦思1,李黎1,邹凯1. 不同信息共享程度下交通拥塞控制研究[J]. 计算机工程与科学, 2016, 38(07): 1391-.