网络高性能交换结构及调度算法研究
|
摘要
近十几年来，互联网络的业务流量呈现持续爆炸性的增长，网络节点（路由器、交换机等）正在逐渐成为网络基础设施发展的关键制约因素，而交换结构（SwitchFabric）是网络节点的物理“核心”，研究高线速、多端口、可扩展性强并且能够提供更好的QoS服务支持的新型交换结构及相关的调度算法是目前的重要课题。在此背景之下，本文针对新型高性能交换结构、低抖动调度算法、多级交换结构的冲突丢包以及全光交换网络的输出波长冲突调度这四个方面的关键问题进行研究。本文的研究获得国家发改委下一代互联网CNGI项目“支持平滑演进的IPv6路由交换设备研制及产业化”以及武汉邮电科学研究院光纤通信技术和网络国家重点实验室开放基金资助项目“下一代光接入关键技术和系统研究”(No.2010OCTN-03)的支持。
输出缓冲排队交换结构（OQ）是目前为止最能够适应各种服务质量需求的交换结构，但是由于其N倍加速的特点，在今天的高速交换网络中难以得到广泛应用。基于交叉矩阵的交换结构可以避免N倍加速问题，但是受到交叉矩阵调度机制的限制，难以支持复杂的服务质量需求。交叉点缓冲交换结构是目前性能最好的交换结构之一，但是由于交换结构队列的限制，非均衡流量下的性能很不理想。为了解决这一问题，本文提出了负载均衡-交叉点缓冲交换结构体系方法，这种新型的交换结构通过引入负载均衡机制将输入端口的非均衡流量变换为中间交换端口的近似均衡流量，从而有效的解决了单纯交叉点缓冲在非均衡流量下吞吐量性能不足的缺陷并且保留了交叉点缓冲交换结构的高服务质量特性，通过实验和理论分析证明了负载均衡-交叉点缓冲交换结构的稳定性和模拟输出缓冲排队的有效性。同时，为了解决引入负载均衡机制可能带来的交换数据流失序问题，进一步改进了负载均衡-交叉点缓冲交换结构，提出了虚拟队列负载均衡-交叉点缓冲交换结构，通过在交叉点缓冲中引入虚拟的输入队列和相应的分级调度算法，有效的保证了交换过程中数据流的各个数据包有序输出。
部分连接多级交换网络具有结构简单，易于扩展等等优点，对于构建高速光交换网络系统具有重要意义。但是多级交换网络也存在着固有的内部链路冲突问题，可能导致数据包的丢失从而影响了网络应用的服务质量。为了能够在保持部分连接多级交换网络的优点的同时，有效的解决内部链路冲突的问题，提出用信息技术中的删除信道模型模拟多级交换网络内部固有冲突造成的数据包丢失，采用编码技术从交换数据流中恢复出丢失的数据包而不改动内部结构和选路算法。新的算法在最大限度的保持部分连接多级交换网络的特点的同时，有效解决了由于内部链路冲突而造成的丢包的问题。
传统研究集中于吞吐量、延时、丢包率等常见的服务质量参数，但是传输抖动也是影响音视频业务主要电信级参数。在目前的混合输入排队-交叉点缓冲交换结构中，低抖动调度的研究仍是一个电信级业务的热点。针对传统方法中动态调度算法抖动抑制性能较差、帧调度算法复杂度高等问题提出基于混合输入排队-交叉点缓冲交换结构（CICQ）的混合动态-帧调度算法和基于帧的低抖动调度算法，将混合输入排队-交叉点缓冲交换结构中基于流量的低抖动帧调度转化为典型的优化问题，通过解优化问题获得了最佳的帧调度。
能够提供巨大的交换容量和业务带宽的波分复用光交换系统是目前构建未来骨干传输交换网络的基础，而在光交换系统中，为了解决竞争输出波长而产生的冲突问题，必须高效率的调度交换结构中的输出光波长和有限波长转换器资源。针对目前的波长冲突调度算法在充分利用波长和转换器资源上不理想的问题，提出波长可调整的冲突调度算法，将可调整的波长带入波长转换器分配阶段与空闲波长一起调度，从而高效的利用了波长和转换器资源，提高了全光交换结构的性能。
With Internet traffic explosively increasing in the recent decades, the internet nodes,routers or switches for instance, are becoming the main restraining factors of developinghigh speed internet. As the physical core of a network node, Switch Fabric is required to bemore flexible, better Quality of Service supporting and larger capacity. Therefore theresearch on new switch fabric structure and its scheduling algorithm with higher speed andbetter performance turn out to be the hotspots of network infrastructure. In this work, fourkey aspects of the research, which are "new type of high performance switch structure","low jitter scheduling algorithm","network coding for multistage switch" and "efficientcontention scheduling scheme on optical switch ", have been studied and elaboratelyillustrated in the dissertation.
So far output-queued switch is the fabric switch that meets most various QoSdemands, although it is suffered from N times internal speedup issue, which can be avoidedby using crossbar and buffered queue based switches. However, in such switches, only onepacket can pass through the crossbar from input port to output port, which constrains theswitch performance and further more limits the complexity of QoS demands. On the otherhand, crosspoint-queued switch is the fabric switch that has the best performance. Howeversince its crosspoint buffer is not as large as input buffer, the performance ofcrosspoint-queued switch under non-uniform traffic is much lower than that under uniformtraffic. To solve this problem, a new load-balance crosspoint-queued switch is proposed.By adding a load-balance stage in front of crosspoint queued stage, it turns non-uniformtraffic at input ports into uniform traffic at internal ports of crosspoint-queued stage. Bythis means the issue of non-uniform traffic is well settled while the high performance ofcrosspoint-queued switch is still kept. Nevertheless, adopting this new switch structuremay cause another problem on out of order. Therefore an improved load-balancecrosspoint-queued switch with virtual input queues is also proposed. It can keep thesequence of input packet using virtual queues at the crosspoint buffers.
The most remarkable features of partial connected multistage switch structure are flexiblility and internal simple routing, which make it very suitable for construction ofmodern optical switching network. However, due to the inherent internal switch pathcollision, its performance is suffered from packet loss. By the modern network codingmethod, this internal switch path collision problem can be emulated as a binary erasurechannel on coding theory. Meanwhile bit errors on binary erasure channel can be correctedby a channel coding method. We proposed a redundant correcting coding scheme toimproving the packet loss without disturbing the internal structure of partial connectedmultistage switch. Analysis and simulation both show that coding method can reduce thepacket loss of partial connected multistage switch structure.
Rapidly growing multimedia applications on Internet require more QoSs than thetraditional Internet applications. Thus in the research work, not only quality of throughput,delay and packet loss but also transmitting jitter, which has import influence on networkvoice and video application, have to be concerned. The low jitter scheduling algorithmfor combined input and crosspoint-queued switch needs to be improved as well as thenowaday frame and matrix decomposition based low jitter scheduling algorithms are fartoo complex for online scehduling. In this work, a hybrid lower jitter scheduling algorithmthat can achieve low jitter performance with low complexity is proposed for combinedinput and crosspoint-queued switch. Then another optimization based algorithm with goodjitter performance is also proposed for frame based scheduling.
Based on Wavelength Division Multiplex (WDM), Modern Optical PacketSwitch(OPS) system provides large switch capacity as well as saves costs. Thus it is verysuitable for building the infrastructure of modern backbone switch network. Nevertheless,without optical buffers inside, traditional all-optical switch sufferes from packet loss due tothe output wave length blocking, which can be avoided by applying a more efficientcontention resolving algorithm. Current contention scheduling algorithms only giveprecedence to output wavelength distribution phrase, while it is unable to efficientlyexploit the available wavelength as well as limited range converter. To improve this, a newwavelength-adjustable contention scheduling algorithm with better packet lossperformance and throughput is proposed in this dissertation. It adjusts the pre-allocatedoutput wavelengthes during wavelength converter scheduling phrase as well as effeciently consumes switch resources, output wavelengthes or wavelength converters for instance.
输出缓冲排队交换结构（OQ）是目前为止最能够适应各种服务质量需求的交换结构，但是由于其N倍加速的特点，在今天的高速交换网络中难以得到广泛应用。基于交叉矩阵的交换结构可以避免N倍加速问题，但是受到交叉矩阵调度机制的限制，难以支持复杂的服务质量需求。交叉点缓冲交换结构是目前性能最好的交换结构之一，但是由于交换结构队列的限制，非均衡流量下的性能很不理想。为了解决这一问题，本文提出了负载均衡-交叉点缓冲交换结构体系方法，这种新型的交换结构通过引入负载均衡机制将输入端口的非均衡流量变换为中间交换端口的近似均衡流量，从而有效的解决了单纯交叉点缓冲在非均衡流量下吞吐量性能不足的缺陷并且保留了交叉点缓冲交换结构的高服务质量特性，通过实验和理论分析证明了负载均衡-交叉点缓冲交换结构的稳定性和模拟输出缓冲排队的有效性。同时，为了解决引入负载均衡机制可能带来的交换数据流失序问题，进一步改进了负载均衡-交叉点缓冲交换结构，提出了虚拟队列负载均衡-交叉点缓冲交换结构，通过在交叉点缓冲中引入虚拟的输入队列和相应的分级调度算法，有效的保证了交换过程中数据流的各个数据包有序输出。
部分连接多级交换网络具有结构简单，易于扩展等等优点，对于构建高速光交换网络系统具有重要意义。但是多级交换网络也存在着固有的内部链路冲突问题，可能导致数据包的丢失从而影响了网络应用的服务质量。为了能够在保持部分连接多级交换网络的优点的同时，有效的解决内部链路冲突的问题，提出用信息技术中的删除信道模型模拟多级交换网络内部固有冲突造成的数据包丢失，采用编码技术从交换数据流中恢复出丢失的数据包而不改动内部结构和选路算法。新的算法在最大限度的保持部分连接多级交换网络的特点的同时，有效解决了由于内部链路冲突而造成的丢包的问题。
传统研究集中于吞吐量、延时、丢包率等常见的服务质量参数，但是传输抖动也是影响音视频业务主要电信级参数。在目前的混合输入排队-交叉点缓冲交换结构中，低抖动调度的研究仍是一个电信级业务的热点。针对传统方法中动态调度算法抖动抑制性能较差、帧调度算法复杂度高等问题提出基于混合输入排队-交叉点缓冲交换结构（CICQ）的混合动态-帧调度算法和基于帧的低抖动调度算法，将混合输入排队-交叉点缓冲交换结构中基于流量的低抖动帧调度转化为典型的优化问题，通过解优化问题获得了最佳的帧调度。
能够提供巨大的交换容量和业务带宽的波分复用光交换系统是目前构建未来骨干传输交换网络的基础，而在光交换系统中，为了解决竞争输出波长而产生的冲突问题，必须高效率的调度交换结构中的输出光波长和有限波长转换器资源。针对目前的波长冲突调度算法在充分利用波长和转换器资源上不理想的问题，提出波长可调整的冲突调度算法，将可调整的波长带入波长转换器分配阶段与空闲波长一起调度，从而高效的利用了波长和转换器资源，提高了全光交换结构的性能。
With Internet traffic explosively increasing in the recent decades, the internet nodes,routers or switches for instance, are becoming the main restraining factors of developinghigh speed internet. As the physical core of a network node, Switch Fabric is required to bemore flexible, better Quality of Service supporting and larger capacity. Therefore theresearch on new switch fabric structure and its scheduling algorithm with higher speed andbetter performance turn out to be the hotspots of network infrastructure. In this work, fourkey aspects of the research, which are "new type of high performance switch structure","low jitter scheduling algorithm","network coding for multistage switch" and "efficientcontention scheduling scheme on optical switch ", have been studied and elaboratelyillustrated in the dissertation.
So far output-queued switch is the fabric switch that meets most various QoSdemands, although it is suffered from N times internal speedup issue, which can be avoidedby using crossbar and buffered queue based switches. However, in such switches, only onepacket can pass through the crossbar from input port to output port, which constrains theswitch performance and further more limits the complexity of QoS demands. On the otherhand, crosspoint-queued switch is the fabric switch that has the best performance. Howeversince its crosspoint buffer is not as large as input buffer, the performance ofcrosspoint-queued switch under non-uniform traffic is much lower than that under uniformtraffic. To solve this problem, a new load-balance crosspoint-queued switch is proposed.By adding a load-balance stage in front of crosspoint queued stage, it turns non-uniformtraffic at input ports into uniform traffic at internal ports of crosspoint-queued stage. Bythis means the issue of non-uniform traffic is well settled while the high performance ofcrosspoint-queued switch is still kept. Nevertheless, adopting this new switch structuremay cause another problem on out of order. Therefore an improved load-balancecrosspoint-queued switch with virtual input queues is also proposed. It can keep thesequence of input packet using virtual queues at the crosspoint buffers.
The most remarkable features of partial connected multistage switch structure are flexiblility and internal simple routing, which make it very suitable for construction ofmodern optical switching network. However, due to the inherent internal switch pathcollision, its performance is suffered from packet loss. By the modern network codingmethod, this internal switch path collision problem can be emulated as a binary erasurechannel on coding theory. Meanwhile bit errors on binary erasure channel can be correctedby a channel coding method. We proposed a redundant correcting coding scheme toimproving the packet loss without disturbing the internal structure of partial connectedmultistage switch. Analysis and simulation both show that coding method can reduce thepacket loss of partial connected multistage switch structure.
Rapidly growing multimedia applications on Internet require more QoSs than thetraditional Internet applications. Thus in the research work, not only quality of throughput,delay and packet loss but also transmitting jitter, which has import influence on networkvoice and video application, have to be concerned. The low jitter scheduling algorithmfor combined input and crosspoint-queued switch needs to be improved as well as thenowaday frame and matrix decomposition based low jitter scheduling algorithms are fartoo complex for online scehduling. In this work, a hybrid lower jitter scheduling algorithmthat can achieve low jitter performance with low complexity is proposed for combinedinput and crosspoint-queued switch. Then another optimization based algorithm with goodjitter performance is also proposed for frame based scheduling.
Based on Wavelength Division Multiplex (WDM), Modern Optical PacketSwitch(OPS) system provides large switch capacity as well as saves costs. Thus it is verysuitable for building the infrastructure of modern backbone switch network. Nevertheless,without optical buffers inside, traditional all-optical switch sufferes from packet loss due tothe output wave length blocking, which can be avoided by applying a more efficientcontention resolving algorithm. Current contention scheduling algorithms only giveprecedence to output wavelength distribution phrase, while it is unable to efficientlyexploit the available wavelength as well as limited range converter. To improve this, a newwavelength-adjustable contention scheduling algorithm with better packet lossperformance and throughput is proposed in this dissertation. It adjusts the pre-allocatedoutput wavelengthes during wavelength converter scheduling phrase as well as effeciently consumes switch resources, output wavelengthes or wavelength converters for instance.
引文
[1] Andersen T, Owicki S, Saxe J, Thacher C. High Speed Switch Scheduling for LocalArea Networks[J].ACM Transactions on Computer Systems,1993,11(4):319-352
[2] McKeown N. The iSLIP Scheduling Algorithm for Input-Queued Switches[J].IEEE/ACM Transactions on Networking,1999,7(2):188–201
[3] Serpanos D N, Antoniadis P I, FIRM:a Class of Distributed Scheduling Algorithms forHigh-speed ATM Switches with Multiple Input Queues[C].IEEE INFOCOM'00,july,2000:548-555
[4] Li Y, Panwar S, Chao H J.The dual round robin matching switch with exhaustiveservice[C]. IEEE Workshop on High Performance Switching andRouting(HPSR'02),Nov,2002:58-63
[5] Giaccone P, Shah D Prabhakar B. An implementable parallel scheduler forinput-queued switches[J].IEEE Microwave,2002,22(1):19-25
[6] Giaccone P, Prabhakar B.,Shah D.Towards simple,high-performance schedulers forhigh-aggregate bandwidth switches[C].IEEE INFOCOM'02,2002:1160-1169
[7] Wu H, Szymanski T H. Credit-based fair scheduling for Input-Queued switch[C].IEEEInternational Conference on Communications(ICC'04),2004:1973-1978
[8] Kumar N, Pan R, Shah D. Fair Scheduling in Input-Queued Switches underInadmissible Traffic[C]. IEEE Globecomm'04,2004:1713-1717
[9] He C,Yeung K L. D-LQF: An Efficient Distributed Scheduling Algorithm forInput-Queued Switches[C]. IEEE International Conference onCommunications(ICC'11),2011:899-907
[10]Bianco P, Andrea M, Giaccone G.Power Control for Crossbar-Based Input-QueuedSwitches[J]. IEEE Transaction on Computers,2013,62(1):74-82
[11]Saidi, H, Hashemi, M R. On the necessary and sufficient requirement of a CIOQswitch to emulate an Output Queued switch[J]. Optical Switching and Networking,2012,9(3):205-213
[12]伊鹏.基于带缓存交叉开关的交换结构研究:[博士学位论文].郑州:解放军信息工程大学图书馆,2006.
[13]郑燕峰.基于输入排队的可扩展交换结构调度算法的研究:[博士学位论文].北京：中国科学院计算技术研究所，2006.
[14]Rojas-Cessa R, Oki E, Jing Z, et al., CIXB-1: combined input-one-cell-crosspointbuffered switch [C]. Proc. IEEE High Performance Switching and Routing(HPSR’01), May2001:324–329
[15]Javidi T,Magil R,Hrabik T. A high-throuput scheduling algorithm for a bufferedcrossbar switch fabric[C]. IEEE International Conference onCommunications(ICC'01),June,2001:1581-1587
[16]Zhang X, Bhuyan L. An effient Algorthm for CombinedInput-Crosspoint-Queued(CICQ) switches. IEEE Globecom2004, November2004:1168-1173
[17]Mhamdi L, Mounir Hamdi. MCBF: A High-performance scheduling algorithm forbuffered crossbar switches. IEEE Communications Letters,2003,No.9:451-453
[18]Hua N, Wang P, JIN D,et al. Simple and Fair Scheduling Algorithm for CombinedInput-Crosspoint-Queued Switch[C]. IEEE International Conference onCommunications(ICC'07), July,2007:508-513
[19]Jin H, Pan D, Pissinou N. Achieving Flow Level Constant Performance Guarantees forCICQ Switches without Speedup[C]. IEEE Globecom2010,2010:384-388
[20]Yoshigoe K. The CICQ Switch with Virtual Crosspoint Queues for LargeRTT[C].IEEE International Conference on Communications(ICC'06),2006:299-303
[21]Giaccone P, Leonardi E. Asymetric Performance Limits of Switches with bufferedcrossbar Supporting Multicast Traffic[J]. IEEE Transaction on Information Theory,2008,54(2):595-607
[22]Jin H, Pan D, Liu J, Pissinou N. OpenFlow based Flow Level Bandwidth Provisioningfor CICQ Switches[C]. IEEE INFOCOM2011,June,2011:476-480
[23]Chao H J, Deng K L,Z Jing. PetaStar:a Petabit Photonic Packet Switch[J].IEEEJournal on Selected Areas in Communications,2003,21(7):1096-1112
[24]Oki E, Jing Z, Rojas-Cessa R, ChaoH J. Concurrent round-robin-based dispatchingschemes for Clos-network switches[J]. IEEE/ACM Transactions on Networking,2002,10(6):830-844
[25]Kleban J, Santos H. Packet Dispatching Algorithms with the Static ConnectionPatterns Scheme for Three-Stage Buffered Clos-Network Switches[C]. IEEEInternational Conference on Communications(ICC'07), July,2007:1104-1109
[26]Dong Z, Rojas-Cessaand R. Non-blocking Memory-Memory-Memory Clos-NetworkPacket Switch[C]. IEEE Sarnoff Symposium2011,May,2011:1-5
[27]Singh G, Bhattacharjee T P, Yadav R P,et al. Design of Non-Blocking andRearrangeable Modified Banyan Network with Electro-Optic MZI SwitchingElements[J]. World Academy of Science, Engineering and Technology,200843:225-230
[28]Danilewicz G, Kabacinski W, Rajewski R. The New Banyan-Based Switching FabricArchitecture Composed of Asymmetrical Optical Switching Elements[C]. IEEEGlobecom2009,2009:189-197
[29]Lin S, Yuang M. QoS Parallel Incremental Scheduling for Optical Pseudo-BanyanPacket Switching System[C]. IEEE12th International Conference on HighPerformance Switching and Routing(HPSR'11), Nov,2011:282-287
[30]Brenner M, Zimmermann A. Analysis of Delay Time Distributions in MultistageInterconnection Networks Considering Multicast Traffic[C]. IEEE7th Symposium onNetwork Computing and Applications(NCA'08),2008:236-239
[31]Zadedyurina O, Ofek Y. A Combined Space and Time Analysis of BanyanNetworks[C]. IEEE High Performance Switching and Routing(HSPR'09),2009:1-8
[32]Chen I Y, Tong L, Huang Y. Rearrangeable Nonblocking Optical InterconnectionNetwork Fabrics With Crosstalk Constraints[J]. IEEE/ACM Transactions onNetworking,2010,18(5):1413-1421
[33]Tucker R S, The Role of Optics and Electronics in high-capacity routers[J]. Journal ofLightwave Technology,2006,24(12):4655–4673
[34]Tuker R S,, Hinton K, Raskutti G, Energy consumption limits in high-speed opticaland electronic signal processing[J]. Electronics Letters,2007,43:906–908
[35]Danielsen S L, Hansen P B, and Stubkjaer K E, Wavelength conversion in opticalpacket switching[J]. Journal of Lightwave Technology,1998,16(6):2095–2108
[36]Eramo V, Listanti M, and Pacifici P.A comparison study on the number of wavelengthconverters needed in synchronous and asynchronous all-optical switchingarchitectures[J] Journal of Lightwave Technology,2003,21(2):340–355
[37]McKeown N, Mekkittikul A, Anantharam V. Achieving100%Throughput in anInput-Queued Switch[J]. IEEE Transactions on Communications,1999,47(8):1260-1267
[38]Chuang S T, Iyer S, McKeown N. Practical Algorithms for Performance Guarantees inBuffered Crossbars[C]. Proceedings of the IEEE24th Annual Joint Conference of theIEEE Computer and Communications Societies: March2005:981-991
[39]Zhang X, Mothanty S R, Bhuyan L N. Adaptive Max-Min Fair Scheduling in BufferedCrossbar Switches Without Speedup[C]. Proceedings of the26th IEEE InternationalConference on Computer Communications: May2007:454-462
[40]Hosaagrahara M, Sethu H. Max-Min Fair Scheduling in Input-Queued Switches[J].IEEE Transactions on Pararrel and Distribute Systems,2008,19(4):462-475
[41]Szymanski T H. A Low-Jitter Guaranteed-Rate Scheduling Algorithm forCrosspoint-Buffered Switches[C]. Proceedings of the IEEE Pacific Rim Conference onCommunications, Computers and Signal Processing: August2009:882-890
[42]Divakaran D M, Anhalt F, Altman E. Size-Based Flow Scheduling in a CICQSwitch[C]. Proceedings of the11th International Conference on High PerformanceSwitching and Routing: June2010:57-62
[43]Minkenberg A F, Iliadis C I, Engbersen A P. Design Issues in Next-GenerationMerchant Switch Fabrics[J]. IEEE/ACM Transactions on Networking,2007,15(6):1603–1615
[44]TSURITANI T, Liu L. Intelligent Control Plane Techniques for Next GenerationOptical Switching Networks[J]. China Communications,2012,9(8):29-39
[45]ITRS Committee. International Technology Roadmap for Semiconductors ExecutiveSummary[R]. New York: ITRS,2007:13-1
[46]Kanizo Y, Hay D, Keslassy I. The Crosspoint-Queued Switch[C]. Proceedings of theIEEE International Conference on Computer Communications: April2009:729-737
[47]Chang C S,Lee D S, Jou Y S. Load Balanced Birkhoff-von Neumann Switches, Part I:One-stage Buffering[J]. Computer Communications,2002,25(2):611–622
[48]Chang C S,Lee D S, Jou Y S. Load Balanced Birkhoff-von Neumann Switches, Part II:Multi-stage Buffering[J]. Computer Communications,2002,25(2):623–634
[49]Smiljanic A. Load Balancing Mechanisms in Clos Packet Switches[C]. IEEEInternational Conference on Communication, May,2004:304-308
[50]Liu X, Goldsmith A. Load Balancing and Switch Scheduling[C]. IEEE InternationalConference on Communication, March,2005:1010-1014
[51]Keslassy I. Israel H. Optimal Load-Balancing[C]. Proceeding of IEEE Infocomm’05,March,2005:1712-1722
[52]Hu Bing, Yeung K L, and Zhang Z Y. Load-balanced Three-stage switch[J]. Journal ofNetwork and Computer Applications,2012,35(1):502-509
[53]Rohas-Cessa R, Dong Z. Load-Balanced Combined Input-Crosspoint Buffered PacketSwitches[J]. IEEE Transactions on Communications,2011,59(5):1421-1433
[54]Dai J G, Prabhakar B. The Throughput of Data Switches With and Without Speedup[C].Proceedings of the IEEE19th Annual Joint Conference of the IEEE Computer andCommunications Societies: March2000:556-564
[55]Shen Yanming, Panswar S S, Chao H J. Providing100%Throughput in a BufferedCrossbar Switch[C]. IEEE High Performance Switching and Routing，May30-June1,2007, New York, USA,2007:1-8
[56]Magill R B, Rohrs C E. Output-Queued Switch Emulation by Fabrics With LimitedMemory[J]. IEEE Journal on Selected Areas in Communications,2003,21(4):606-615
[57]Shen Yanming, Panswar S S, Chao H J. Design and Performance Analysis of aPractical Load-Balanced Switch[J]. IEEE Transactions on Communications,2009,57(8):2420-2429
[58]McKeown N. SIM [EB/OL]. http://klamath.stanford.edu/tools/SIM/,2007-05-16/2009-07-11
[59]牛艳花,欧阳峰. IPTV服务质量关键技术研究[J].广播与电视技术,2009,3:71-74
[60]Keslassy I, Chuang S, Yu K, et al. Scaling Internet Routers using Optics[C].Proceedings of ACM SIGCOMM’03, August2003:1-16
[61]Chang C S, Lee D S, Shih Y J. Mailbox Switch: a Scalable Two-stage SwitchArchitecture for Conflict Resolution of Ordered Packets[J]. IEEE Transaction onCommunications,2008,56(1):136-149
[62]Yu C L, Chang C S, Lee D S. CR Switch: A Load-Balanced Switch with Contentionand Reservation[J]. IEEE Transation on Networking,2009,17(5):1659-1671
[63]Lin B. The Interleaved Matching Switch Architecture[J]. IEEE Transaction onCommunications,2009,57(2):1-10
[64]Clos C. A study of non-blocking switching networks[J]. Bell System TechnicalJournal,1953,32(5):406-424
[65]Pattavina A. Switching Theory–Architectures and Performance in Broadband ATMNetworks[M]. England: John Wiley&Sons,1998
[66]Zadedyurina O, Ofek O, Pattavina A. Space and Time Blocking versus Cost inAll-Optical Banyan Networks[C]. Proc of IEEE ICC'08.2008:5338-5343
[67]Shi Hongyuan, Sethu Harish. Virtual Circuit Blocking Probabilities in an ATM BanyanNetwork with bxb Switching Elements[C]. Advanced Simulation TechnologiesConference2001. SCS,2001:768-775
[68]Zhang Jing, Gu Huaxi. Partially adaptive routing algorithm for Benes network onchip[C]. Proc of IEEE ICCSIT'09.2009:614-618
[69]Lu E., Zheng S. Q. Parallel routing algorithms for non blocking electronic andphotonic switching networks[J]. IEEE Transactions on Parallel and DistributedSystems,2005,16(8):702-713
[70]Chan K., Yeung K., Chan S. A Refined Model for Performance Analysis of BufferedBanyan Networks with and without Priority Control[C]. Proc of IEEEGLOBECOM'97.1997:1745-1750
[71]Ahlswede R, Cai N, Yeung R. W., et al. Network information flow[J]. IEEE Trans.Inform. Theory,2000,46(4):1204–1216
[72]Li Shuoyen, Yeung R W, Cai Ning. Linear network coding[J]. IEEE Trans. Inf. Theory,2003,49(2):371–381
[73]Médard M., Kotter R., Effros M., et al. A random linear Network coding approach tomutlicast[J]. IEEE Trans. Inf. Theory,2006,52(10):4413–4430
[74]Kim M., Sundararajan J.K., Médard M., et al. Network Coding in a MulticastSwitch[J]. IEEE Trans. Inf. Theory,2011,57(1):436–460
[75]Rayanchu S., Sen S., Banerjee S., et al. Loss-Aware Network Coding for UnicastWireless Sessions: Design, Implementation, and Performance Evaluation[C]. Proc ofSIGMETRICS’08. New York, NY: ACM,2008:85-96
[76]Al-Shaikhi A.A., Ilow, J. Packet Loss Recovery Codes Based on VandermondeMatrices and Shift Operators[C]. IEEE ISIT'08,2008:1058-1062
[77]Wang Hui, Liang J., Kuo Jay. Overview of Robust Video Streaming with NetworkCoding[J]. Journal of Information Hiding and Multimedia Signal Processing,2010.1(1):36-50
[78]Yeung Raymond W. Information Theory and Network Coding[M]. Hong Kong:Springer,2008
[79]Morelos-Zaragoza Robert H. The Art of Error Correcting Coding[M]. England: JohnWiley&Sons,2006
[80]李挥,何伟,伊鹏等.排序集线器多级互连交换结构的多路径自路由模型[J].电子学报,2008,36(1):1-8
[81]李挥，王秉睿，黄佳庆等.负载均衡自路由交换结构[J].通信学报，2009,30(5):7-15
[82]Feng G. L., Deng R. H., Bao, F. Packet-loss resilient coding scheme with only XORoperations[J]. IEEE Communications,2004,151(4):322–328
[83]Claypool M, Tanner J. The effects of jitter on the perceptual quality of video[C]. ACMMultimedia’99(Part2),1999:115–118
[84]Steinmetz R. Human perception of jitter and media synchronization[J]. IEEE Journalon Selected Areas in Communications,1996,14(1),61–72
[85]Lombardo A, Schembra G. Applying Active Window Management for Jitter Controland Loss Avoidance in Video Streaming over TCP Connections[C]. IEEEGlobecom’10, June,2010:788-795
[86]Parekh A K, Gallager R G. A Generalized Processor Sharing Approach to Flow Controlin Integrated Service Networks: the Multiple Node Case[J]. IEEE/ACM Trans.Networking,1994,2(2),137-150
[87]Nabeshima N. Performance evaluation of a combined input and crosspoint queuedswitch[J]. IEICE Trans. Commun.,2000, E83-B(3):737-741
[88]Chen W J, Chang C S, and Huang H Y. Birkhoff-von Neumann Input BufferedCrossbar Switches[J]. IEEE Trans.Communications,2001,49,(7),:1145-1147
[89]Koksal C E, Gallager R G, Rohrs C E. Rate Quantization and Service Quality overSingle Crossbar Switches[C]. Proceeding of IEEE Infocom’04,2004:1962-1974
[90]Keslassy I, Kodialam M, Lakshamn T V. On Guaranteed Smooth Scheduling forInput-Queued Switches[J]. IEEE/ACM Trans. Networking,2005,13,(6):1384-1392
[91]Zhang X, Mohanty S R, and Bhuyan L N, Adaptive Max-Min Fair Scheduling inBuffered Crossbar Switches without Speedup[C]. Proceeding of IEEE Infocom2007,454-461
[92]Georgakopoulos G F. Buffered Cross-Bar Switches Revisited: Design Steps, Proofsand Simulations Towards Optimal Rate and Minimal Buffer Requirements[J].IEEE/ACM Trans. Networking,2008,16(6):1340-1351
[93]He S M, Sun S T, Guan H T, et al.. On Guaranteed Smooth Switching for BufferedCrossbar Switches[J]. IEEE/ACM Trans. Networking, Vol.16, No.3, June2008, pp.718-730
[94]Szymanski T H. A Low-Jitter Guaranteed-Rate Scheduling Algorithm forPacket-Switched IP Routers[J]. IEEE Transactions on Communications,2009,57(11):1910-1918
[95]Szymanski T H. Memory Requirements for Future Internet Routers withEssentially-Perfect QoS Guarantees[C]. IEEE Globecom’10Workshop CCNET,2010:398-407
[96]Olesinski W. Simple two-priority, low-jitter scheduler[C]. ACM/IEEE Symposium onArchitectures for Networking and Communications Systems (ANCS’10),2010:1-3
[97]Zhang J, Jin Y, Ansari N. Dichotomy Slot Allocation: A Low-Jitter Scheduling Schemefor Input-Queued Switches[C]. IEEE High Performance Switching andRouting(HPSR’07), April,2007:1-6
[98]John A. Stankovic, Marco S, et al. Deadline Scheduling for Real-Time Systems: EDFand Related Algorithms[M]. Kluwer Academic Publishers,1998
[99]Baruah S, Buttazzo G, Gorinsky S, et al. Scheduling periodic task systems to minimizeoutput jitter[C]. Real-Time Computing Systems and Applications(RTCSA '99),1999:62-69
[100] Mohanty S R, Bhuyan L N, Guaranteed Smooth Switch Scheduling with LowComplexity[C]. IEEE Globecom’05, June,2005:626-630
[101] Boyd S, and Vandenberghe L, Convex Optimization[M]. Cambridge University Press,2003
[102] Boyd S. CVX toolkit [EB/OL]. http://cvxr.com/cvx/,2008-03-01/2009-10-20
[103] Detti A, Parca G, Carrozzo V, Betti S, Optical Packet Network with Limited-rangeWavelength Conversion: A Novel Formalization of the Optimal SchedulingProblem[J]. Journal Lightwave Technology,2009,27(24):5607–5618
[104] Qin X D, Yang Y. Multicast Connection Capacity of WDM Switching Networks withLimited Wavelength Conversion[J]. IEEE/ACM Transactions on Networking,2004,12(3):526–538
[105] Eramo V, Listanti M, Spaziani M. Resources Sharing in Optical Packet Switches withLimited-range Wavelength Converters[J]. Journal of LightwaveTechnology,2005,23(2):671–687
[106] Li H L, Thing I J, Cost-saving Two-layer Wavelength Conversion in OpticalSwitching Network[j]. Journal of Lightwave Technology,2006,24(2):705–712
[107] Eramo V, Listanti M, Germoni A. Cost Evaluation of Optical Packet SwitchesEquipped with Limited-range and Full-range Converters for Contention Resolution[J].Journal of Lightwave Technology,2008,26(4):390–407
[108] Eramo V, Germoni A, Raffaelli C, et al. Multifiber Shared Per-wavelength All-opticalSwitching: Architectures, Control, and Performance[J]. Journal of LightwaveTechnology,2008,26(5):537–551
[109] Okonkwo C, Almeida R C, Martin R E, et al. Performance Analysis of an OpticalPacket Switch with Shared Parametric Wavelength Converters[J]. IEEECommununication Letters,2008,12(8):596–598
[110] Kitsuwan N, Rojas-Cessa R, Matsuura M, et al. Performance of Optical PacketWavelength Converters[J]. Journal of Optical Communication Networks,2010,2(8):558-569
[111] Raffaelli C, Savi M. Sharing Simple Wavelength Converters in Optical PacketSwitching: Performance and Power Consumption Evaluation[C]. IEEE Globecom’10,2010:287-296
[112] Liu H, Zhang B, Shi S A Novel Contention Resolution Scheme of Hybrid SharedWavelength Conversion for Optical Packet Switching[J]. Journal of LightwaveTechnology,2012,30(2):303-312
[113] Eramo V, Germoni A,Cianfrani A, et al. Evaluation of Power Consumption in LowSpatial Complexity Optical Switching Fabrics[J]. IEEE Journal of Selected Topics inQuantum Electronics,2011,17(2):356-368
[114] Raffaelli C, Savi M, Stavdas A. Multistage Shared-Per-Wavelength Optical PacketSwitch: Heuristic Scheduling Algorithm and Performance[J]. Journal of LightwaveTechnology,2009,27(5):1899-1910
[115] Rogiest W, Turck K D, Laevens K D et al. Optimized Channel and Delay Selectionfor Contention Resolution in Optical Networks[C]. IEEE Intenational Conference onCommunications(ICC’11), June,2011:1113-1119
[116] Kitsuwan N, Rojas-Cessa R, Matsuura M, et al. Performance of Optical PacketSwitches Based on Parametric Wavelength Converters[J]. Journal of OpticalCommunication and Networking,2010,2(8):558-569
[117]沈建华.光网络生存性及控制平面技术研究:[博士学位论文].南京:南京邮电大学图书馆:2012.
[118] Kitsuwan N, Oki E. Dynamic Pump-Wavelength Selection for Optical Packet Switchwith Recursive Parametric Wavelength Conversion[C].IEEE13th InternationalConference on High Performance Switching and Routing(HPSR’12),2012:174-178
[2] McKeown N. The iSLIP Scheduling Algorithm for Input-Queued Switches[J].IEEE/ACM Transactions on Networking,1999,7(2):188–201
[3] Serpanos D N, Antoniadis P I, FIRM:a Class of Distributed Scheduling Algorithms forHigh-speed ATM Switches with Multiple Input Queues[C].IEEE INFOCOM'00,july,2000:548-555
[4] Li Y, Panwar S, Chao H J.The dual round robin matching switch with exhaustiveservice[C]. IEEE Workshop on High Performance Switching andRouting(HPSR'02),Nov,2002:58-63
[5] Giaccone P, Shah D Prabhakar B. An implementable parallel scheduler forinput-queued switches[J].IEEE Microwave,2002,22(1):19-25
[6] Giaccone P, Prabhakar B.,Shah D.Towards simple,high-performance schedulers forhigh-aggregate bandwidth switches[C].IEEE INFOCOM'02,2002:1160-1169
[7] Wu H, Szymanski T H. Credit-based fair scheduling for Input-Queued switch[C].IEEEInternational Conference on Communications(ICC'04),2004:1973-1978
[8] Kumar N, Pan R, Shah D. Fair Scheduling in Input-Queued Switches underInadmissible Traffic[C]. IEEE Globecomm'04,2004:1713-1717
[9] He C,Yeung K L. D-LQF: An Efficient Distributed Scheduling Algorithm forInput-Queued Switches[C]. IEEE International Conference onCommunications(ICC'11),2011:899-907
[10]Bianco P, Andrea M, Giaccone G.Power Control for Crossbar-Based Input-QueuedSwitches[J]. IEEE Transaction on Computers,2013,62(1):74-82
[11]Saidi, H, Hashemi, M R. On the necessary and sufficient requirement of a CIOQswitch to emulate an Output Queued switch[J]. Optical Switching and Networking,2012,9(3):205-213
[12]伊鹏.基于带缓存交叉开关的交换结构研究:[博士学位论文].郑州:解放军信息工程大学图书馆,2006.
[13]郑燕峰.基于输入排队的可扩展交换结构调度算法的研究:[博士学位论文].北京：中国科学院计算技术研究所，2006.
[14]Rojas-Cessa R, Oki E, Jing Z, et al., CIXB-1: combined input-one-cell-crosspointbuffered switch [C]. Proc. IEEE High Performance Switching and Routing(HPSR’01), May2001:324–329
[15]Javidi T,Magil R,Hrabik T. A high-throuput scheduling algorithm for a bufferedcrossbar switch fabric[C]. IEEE International Conference onCommunications(ICC'01),June,2001:1581-1587
[16]Zhang X, Bhuyan L. An effient Algorthm for CombinedInput-Crosspoint-Queued(CICQ) switches. IEEE Globecom2004, November2004:1168-1173
[17]Mhamdi L, Mounir Hamdi. MCBF: A High-performance scheduling algorithm forbuffered crossbar switches. IEEE Communications Letters,2003,No.9:451-453
[18]Hua N, Wang P, JIN D,et al. Simple and Fair Scheduling Algorithm for CombinedInput-Crosspoint-Queued Switch[C]. IEEE International Conference onCommunications(ICC'07), July,2007:508-513
[19]Jin H, Pan D, Pissinou N. Achieving Flow Level Constant Performance Guarantees forCICQ Switches without Speedup[C]. IEEE Globecom2010,2010:384-388
[20]Yoshigoe K. The CICQ Switch with Virtual Crosspoint Queues for LargeRTT[C].IEEE International Conference on Communications(ICC'06),2006:299-303
[21]Giaccone P, Leonardi E. Asymetric Performance Limits of Switches with bufferedcrossbar Supporting Multicast Traffic[J]. IEEE Transaction on Information Theory,2008,54(2):595-607
[22]Jin H, Pan D, Liu J, Pissinou N. OpenFlow based Flow Level Bandwidth Provisioningfor CICQ Switches[C]. IEEE INFOCOM2011,June,2011:476-480
[23]Chao H J, Deng K L,Z Jing. PetaStar:a Petabit Photonic Packet Switch[J].IEEEJournal on Selected Areas in Communications,2003,21(7):1096-1112
[24]Oki E, Jing Z, Rojas-Cessa R, ChaoH J. Concurrent round-robin-based dispatchingschemes for Clos-network switches[J]. IEEE/ACM Transactions on Networking,2002,10(6):830-844
[25]Kleban J, Santos H. Packet Dispatching Algorithms with the Static ConnectionPatterns Scheme for Three-Stage Buffered Clos-Network Switches[C]. IEEEInternational Conference on Communications(ICC'07), July,2007:1104-1109
[26]Dong Z, Rojas-Cessaand R. Non-blocking Memory-Memory-Memory Clos-NetworkPacket Switch[C]. IEEE Sarnoff Symposium2011,May,2011:1-5
[27]Singh G, Bhattacharjee T P, Yadav R P,et al. Design of Non-Blocking andRearrangeable Modified Banyan Network with Electro-Optic MZI SwitchingElements[J]. World Academy of Science, Engineering and Technology,200843:225-230
[28]Danilewicz G, Kabacinski W, Rajewski R. The New Banyan-Based Switching FabricArchitecture Composed of Asymmetrical Optical Switching Elements[C]. IEEEGlobecom2009,2009:189-197
[29]Lin S, Yuang M. QoS Parallel Incremental Scheduling for Optical Pseudo-BanyanPacket Switching System[C]. IEEE12th International Conference on HighPerformance Switching and Routing(HPSR'11), Nov,2011:282-287
[30]Brenner M, Zimmermann A. Analysis of Delay Time Distributions in MultistageInterconnection Networks Considering Multicast Traffic[C]. IEEE7th Symposium onNetwork Computing and Applications(NCA'08),2008:236-239
[31]Zadedyurina O, Ofek Y. A Combined Space and Time Analysis of BanyanNetworks[C]. IEEE High Performance Switching and Routing(HSPR'09),2009:1-8
[32]Chen I Y, Tong L, Huang Y. Rearrangeable Nonblocking Optical InterconnectionNetwork Fabrics With Crosstalk Constraints[J]. IEEE/ACM Transactions onNetworking,2010,18(5):1413-1421
[33]Tucker R S, The Role of Optics and Electronics in high-capacity routers[J]. Journal ofLightwave Technology,2006,24(12):4655–4673
[34]Tuker R S,, Hinton K, Raskutti G, Energy consumption limits in high-speed opticaland electronic signal processing[J]. Electronics Letters,2007,43:906–908
[35]Danielsen S L, Hansen P B, and Stubkjaer K E, Wavelength conversion in opticalpacket switching[J]. Journal of Lightwave Technology,1998,16(6):2095–2108
[36]Eramo V, Listanti M, and Pacifici P.A comparison study on the number of wavelengthconverters needed in synchronous and asynchronous all-optical switchingarchitectures[J] Journal of Lightwave Technology,2003,21(2):340–355
[37]McKeown N, Mekkittikul A, Anantharam V. Achieving100%Throughput in anInput-Queued Switch[J]. IEEE Transactions on Communications,1999,47(8):1260-1267
[38]Chuang S T, Iyer S, McKeown N. Practical Algorithms for Performance Guarantees inBuffered Crossbars[C]. Proceedings of the IEEE24th Annual Joint Conference of theIEEE Computer and Communications Societies: March2005:981-991
[39]Zhang X, Mothanty S R, Bhuyan L N. Adaptive Max-Min Fair Scheduling in BufferedCrossbar Switches Without Speedup[C]. Proceedings of the26th IEEE InternationalConference on Computer Communications: May2007:454-462
[40]Hosaagrahara M, Sethu H. Max-Min Fair Scheduling in Input-Queued Switches[J].IEEE Transactions on Pararrel and Distribute Systems,2008,19(4):462-475
[41]Szymanski T H. A Low-Jitter Guaranteed-Rate Scheduling Algorithm forCrosspoint-Buffered Switches[C]. Proceedings of the IEEE Pacific Rim Conference onCommunications, Computers and Signal Processing: August2009:882-890
[42]Divakaran D M, Anhalt F, Altman E. Size-Based Flow Scheduling in a CICQSwitch[C]. Proceedings of the11th International Conference on High PerformanceSwitching and Routing: June2010:57-62
[43]Minkenberg A F, Iliadis C I, Engbersen A P. Design Issues in Next-GenerationMerchant Switch Fabrics[J]. IEEE/ACM Transactions on Networking,2007,15(6):1603–1615
[44]TSURITANI T, Liu L. Intelligent Control Plane Techniques for Next GenerationOptical Switching Networks[J]. China Communications,2012,9(8):29-39
[45]ITRS Committee. International Technology Roadmap for Semiconductors ExecutiveSummary[R]. New York: ITRS,2007:13-1
[46]Kanizo Y, Hay D, Keslassy I. The Crosspoint-Queued Switch[C]. Proceedings of theIEEE International Conference on Computer Communications: April2009:729-737
[47]Chang C S,Lee D S, Jou Y S. Load Balanced Birkhoff-von Neumann Switches, Part I:One-stage Buffering[J]. Computer Communications,2002,25(2):611–622
[48]Chang C S,Lee D S, Jou Y S. Load Balanced Birkhoff-von Neumann Switches, Part II:Multi-stage Buffering[J]. Computer Communications,2002,25(2):623–634
[49]Smiljanic A. Load Balancing Mechanisms in Clos Packet Switches[C]. IEEEInternational Conference on Communication, May,2004:304-308
[50]Liu X, Goldsmith A. Load Balancing and Switch Scheduling[C]. IEEE InternationalConference on Communication, March,2005:1010-1014
[51]Keslassy I. Israel H. Optimal Load-Balancing[C]. Proceeding of IEEE Infocomm’05,March,2005:1712-1722
[52]Hu Bing, Yeung K L, and Zhang Z Y. Load-balanced Three-stage switch[J]. Journal ofNetwork and Computer Applications,2012,35(1):502-509
[53]Rohas-Cessa R, Dong Z. Load-Balanced Combined Input-Crosspoint Buffered PacketSwitches[J]. IEEE Transactions on Communications,2011,59(5):1421-1433
[54]Dai J G, Prabhakar B. The Throughput of Data Switches With and Without Speedup[C].Proceedings of the IEEE19th Annual Joint Conference of the IEEE Computer andCommunications Societies: March2000:556-564
[55]Shen Yanming, Panswar S S, Chao H J. Providing100%Throughput in a BufferedCrossbar Switch[C]. IEEE High Performance Switching and Routing，May30-June1,2007, New York, USA,2007:1-8
[56]Magill R B, Rohrs C E. Output-Queued Switch Emulation by Fabrics With LimitedMemory[J]. IEEE Journal on Selected Areas in Communications,2003,21(4):606-615
[57]Shen Yanming, Panswar S S, Chao H J. Design and Performance Analysis of aPractical Load-Balanced Switch[J]. IEEE Transactions on Communications,2009,57(8):2420-2429
[58]McKeown N. SIM [EB/OL]. http://klamath.stanford.edu/tools/SIM/,2007-05-16/2009-07-11
[59]牛艳花,欧阳峰. IPTV服务质量关键技术研究[J].广播与电视技术,2009,3:71-74
[60]Keslassy I, Chuang S, Yu K, et al. Scaling Internet Routers using Optics[C].Proceedings of ACM SIGCOMM’03, August2003:1-16
[61]Chang C S, Lee D S, Shih Y J. Mailbox Switch: a Scalable Two-stage SwitchArchitecture for Conflict Resolution of Ordered Packets[J]. IEEE Transaction onCommunications,2008,56(1):136-149
[62]Yu C L, Chang C S, Lee D S. CR Switch: A Load-Balanced Switch with Contentionand Reservation[J]. IEEE Transation on Networking,2009,17(5):1659-1671
[63]Lin B. The Interleaved Matching Switch Architecture[J]. IEEE Transaction onCommunications,2009,57(2):1-10
[64]Clos C. A study of non-blocking switching networks[J]. Bell System TechnicalJournal,1953,32(5):406-424
[65]Pattavina A. Switching Theory–Architectures and Performance in Broadband ATMNetworks[M]. England: John Wiley&Sons,1998
[66]Zadedyurina O, Ofek O, Pattavina A. Space and Time Blocking versus Cost inAll-Optical Banyan Networks[C]. Proc of IEEE ICC'08.2008:5338-5343
[67]Shi Hongyuan, Sethu Harish. Virtual Circuit Blocking Probabilities in an ATM BanyanNetwork with bxb Switching Elements[C]. Advanced Simulation TechnologiesConference2001. SCS,2001:768-775
[68]Zhang Jing, Gu Huaxi. Partially adaptive routing algorithm for Benes network onchip[C]. Proc of IEEE ICCSIT'09.2009:614-618
[69]Lu E., Zheng S. Q. Parallel routing algorithms for non blocking electronic andphotonic switching networks[J]. IEEE Transactions on Parallel and DistributedSystems,2005,16(8):702-713
[70]Chan K., Yeung K., Chan S. A Refined Model for Performance Analysis of BufferedBanyan Networks with and without Priority Control[C]. Proc of IEEEGLOBECOM'97.1997:1745-1750
[71]Ahlswede R, Cai N, Yeung R. W., et al. Network information flow[J]. IEEE Trans.Inform. Theory,2000,46(4):1204–1216
[72]Li Shuoyen, Yeung R W, Cai Ning. Linear network coding[J]. IEEE Trans. Inf. Theory,2003,49(2):371–381
[73]Médard M., Kotter R., Effros M., et al. A random linear Network coding approach tomutlicast[J]. IEEE Trans. Inf. Theory,2006,52(10):4413–4430
[74]Kim M., Sundararajan J.K., Médard M., et al. Network Coding in a MulticastSwitch[J]. IEEE Trans. Inf. Theory,2011,57(1):436–460
[75]Rayanchu S., Sen S., Banerjee S., et al. Loss-Aware Network Coding for UnicastWireless Sessions: Design, Implementation, and Performance Evaluation[C]. Proc ofSIGMETRICS’08. New York, NY: ACM,2008:85-96
[76]Al-Shaikhi A.A., Ilow, J. Packet Loss Recovery Codes Based on VandermondeMatrices and Shift Operators[C]. IEEE ISIT'08,2008:1058-1062
[77]Wang Hui, Liang J., Kuo Jay. Overview of Robust Video Streaming with NetworkCoding[J]. Journal of Information Hiding and Multimedia Signal Processing,2010.1(1):36-50
[78]Yeung Raymond W. Information Theory and Network Coding[M]. Hong Kong:Springer,2008
[79]Morelos-Zaragoza Robert H. The Art of Error Correcting Coding[M]. England: JohnWiley&Sons,2006
[80]李挥,何伟,伊鹏等.排序集线器多级互连交换结构的多路径自路由模型[J].电子学报,2008,36(1):1-8
[81]李挥，王秉睿，黄佳庆等.负载均衡自路由交换结构[J].通信学报，2009,30(5):7-15
[82]Feng G. L., Deng R. H., Bao, F. Packet-loss resilient coding scheme with only XORoperations[J]. IEEE Communications,2004,151(4):322–328
[83]Claypool M, Tanner J. The effects of jitter on the perceptual quality of video[C]. ACMMultimedia’99(Part2),1999:115–118
[84]Steinmetz R. Human perception of jitter and media synchronization[J]. IEEE Journalon Selected Areas in Communications,1996,14(1),61–72
[85]Lombardo A, Schembra G. Applying Active Window Management for Jitter Controland Loss Avoidance in Video Streaming over TCP Connections[C]. IEEEGlobecom’10, June,2010:788-795
[86]Parekh A K, Gallager R G. A Generalized Processor Sharing Approach to Flow Controlin Integrated Service Networks: the Multiple Node Case[J]. IEEE/ACM Trans.Networking,1994,2(2),137-150
[87]Nabeshima N. Performance evaluation of a combined input and crosspoint queuedswitch[J]. IEICE Trans. Commun.,2000, E83-B(3):737-741
[88]Chen W J, Chang C S, and Huang H Y. Birkhoff-von Neumann Input BufferedCrossbar Switches[J]. IEEE Trans.Communications,2001,49,(7),:1145-1147
[89]Koksal C E, Gallager R G, Rohrs C E. Rate Quantization and Service Quality overSingle Crossbar Switches[C]. Proceeding of IEEE Infocom’04,2004:1962-1974
[90]Keslassy I, Kodialam M, Lakshamn T V. On Guaranteed Smooth Scheduling forInput-Queued Switches[J]. IEEE/ACM Trans. Networking,2005,13,(6):1384-1392
[91]Zhang X, Mohanty S R, and Bhuyan L N, Adaptive Max-Min Fair Scheduling inBuffered Crossbar Switches without Speedup[C]. Proceeding of IEEE Infocom2007,454-461
[92]Georgakopoulos G F. Buffered Cross-Bar Switches Revisited: Design Steps, Proofsand Simulations Towards Optimal Rate and Minimal Buffer Requirements[J].IEEE/ACM Trans. Networking,2008,16(6):1340-1351
[93]He S M, Sun S T, Guan H T, et al.. On Guaranteed Smooth Switching for BufferedCrossbar Switches[J]. IEEE/ACM Trans. Networking, Vol.16, No.3, June2008, pp.718-730
[94]Szymanski T H. A Low-Jitter Guaranteed-Rate Scheduling Algorithm forPacket-Switched IP Routers[J]. IEEE Transactions on Communications,2009,57(11):1910-1918
[95]Szymanski T H. Memory Requirements for Future Internet Routers withEssentially-Perfect QoS Guarantees[C]. IEEE Globecom’10Workshop CCNET,2010:398-407
[96]Olesinski W. Simple two-priority, low-jitter scheduler[C]. ACM/IEEE Symposium onArchitectures for Networking and Communications Systems (ANCS’10),2010:1-3
[97]Zhang J, Jin Y, Ansari N. Dichotomy Slot Allocation: A Low-Jitter Scheduling Schemefor Input-Queued Switches[C]. IEEE High Performance Switching andRouting(HPSR’07), April,2007:1-6
[98]John A. Stankovic, Marco S, et al. Deadline Scheduling for Real-Time Systems: EDFand Related Algorithms[M]. Kluwer Academic Publishers,1998
[99]Baruah S, Buttazzo G, Gorinsky S, et al. Scheduling periodic task systems to minimizeoutput jitter[C]. Real-Time Computing Systems and Applications(RTCSA '99),1999:62-69
[100] Mohanty S R, Bhuyan L N, Guaranteed Smooth Switch Scheduling with LowComplexity[C]. IEEE Globecom’05, June,2005:626-630
[101] Boyd S, and Vandenberghe L, Convex Optimization[M]. Cambridge University Press,2003
[102] Boyd S. CVX toolkit [EB/OL]. http://cvxr.com/cvx/,2008-03-01/2009-10-20
[103] Detti A, Parca G, Carrozzo V, Betti S, Optical Packet Network with Limited-rangeWavelength Conversion: A Novel Formalization of the Optimal SchedulingProblem[J]. Journal Lightwave Technology,2009,27(24):5607–5618
[104] Qin X D, Yang Y. Multicast Connection Capacity of WDM Switching Networks withLimited Wavelength Conversion[J]. IEEE/ACM Transactions on Networking,2004,12(3):526–538
[105] Eramo V, Listanti M, Spaziani M. Resources Sharing in Optical Packet Switches withLimited-range Wavelength Converters[J]. Journal of LightwaveTechnology,2005,23(2):671–687
[106] Li H L, Thing I J, Cost-saving Two-layer Wavelength Conversion in OpticalSwitching Network[j]. Journal of Lightwave Technology,2006,24(2):705–712
[107] Eramo V, Listanti M, Germoni A. Cost Evaluation of Optical Packet SwitchesEquipped with Limited-range and Full-range Converters for Contention Resolution[J].Journal of Lightwave Technology,2008,26(4):390–407
[108] Eramo V, Germoni A, Raffaelli C, et al. Multifiber Shared Per-wavelength All-opticalSwitching: Architectures, Control, and Performance[J]. Journal of LightwaveTechnology,2008,26(5):537–551
[109] Okonkwo C, Almeida R C, Martin R E, et al. Performance Analysis of an OpticalPacket Switch with Shared Parametric Wavelength Converters[J]. IEEECommununication Letters,2008,12(8):596–598
[110] Kitsuwan N, Rojas-Cessa R, Matsuura M, et al. Performance of Optical PacketWavelength Converters[J]. Journal of Optical Communication Networks,2010,2(8):558-569
[111] Raffaelli C, Savi M. Sharing Simple Wavelength Converters in Optical PacketSwitching: Performance and Power Consumption Evaluation[C]. IEEE Globecom’10,2010:287-296
[112] Liu H, Zhang B, Shi S A Novel Contention Resolution Scheme of Hybrid SharedWavelength Conversion for Optical Packet Switching[J]. Journal of LightwaveTechnology,2012,30(2):303-312
[113] Eramo V, Germoni A,Cianfrani A, et al. Evaluation of Power Consumption in LowSpatial Complexity Optical Switching Fabrics[J]. IEEE Journal of Selected Topics inQuantum Electronics,2011,17(2):356-368
[114] Raffaelli C, Savi M, Stavdas A. Multistage Shared-Per-Wavelength Optical PacketSwitch: Heuristic Scheduling Algorithm and Performance[J]. Journal of LightwaveTechnology,2009,27(5):1899-1910
[115] Rogiest W, Turck K D, Laevens K D et al. Optimized Channel and Delay Selectionfor Contention Resolution in Optical Networks[C]. IEEE Intenational Conference onCommunications(ICC’11), June,2011:1113-1119
[116] Kitsuwan N, Rojas-Cessa R, Matsuura M, et al. Performance of Optical PacketSwitches Based on Parametric Wavelength Converters[J]. Journal of OpticalCommunication and Networking,2010,2(8):558-569
[117]沈建华.光网络生存性及控制平面技术研究:[博士学位论文].南京:南京邮电大学图书馆:2012.
[118] Kitsuwan N, Oki E. Dynamic Pump-Wavelength Selection for Optical Packet Switchwith Recursive Parametric Wavelength Conversion[C].IEEE13th InternationalConference on High Performance Switching and Routing(HPSR’12),2012:174-178