基于SDN的信息网络的设计与实现
|
摘要
在当前的互联网中,网络中的信息量越来越多,导致网络中的流量骤增。但是在当前的网络中,信息的传输完全基于信息的地址,信息的内容等对网络来说是透明的。因此在网络中传输的很多相同信息,严重浪费了网络带宽和能源。同时从社会经济学的角度来讲,当前的网络架构对于信息服务的提供者来说,利润是比较低的。因此未来的网络应该提供一种新的发布和订阅的方式,对不同的人群能够提供不同的服务,从而保障利益相关者的收益。因此需要一个新的网络架构来满足信息的有效分发。
本文首先调研了信息网络的研究现状,介绍了几种现有的信息网络模型：NetInf、NDN、PSIRP。并对信息网络中的关键技术进行了总结,分析了现有信息网络架构中存在的问题。同时介绍了SDN技术及研究现状,总结了SDN和信息网络结合带来的好处。
本文提出了一种基于SDN的信息网络架构来解决当前网络中存在的问题,同时尽量避免现有信息网络中存在的问题。论文中介绍了基于SDN技术设计的信息网络架构,在设计中,采用控制平面和数据平面相分离,控制平面负责信息的分发控制、注册、检索、缓存管理等功能,数据平面进行数据的转发。在本论文中,设计了信息的命名和分发路径构造方案,该命名即能够保证信息的自认证性又能够为信息的分发提供依据；分发路径的构造是整个网络架构中最重要的部分,也是本架构最有特色的部分。为了保证信息传输路径的有效性,在信息传输前,网络需根据信息的特性和网络状况为信息构造分发路径。另外,本网络架构也支持移动性管理和组播,从而为信息的有效传输提供了保障。
在完成架构的设计及实现后,本论文对架构进行了性能评估。在性能评估中,采用了理论分析与仿真结果相结合的方法,使用Open Flow和Mininet对网络进行了仿真。评估指标包括分发路径构造的开销、信息的传输时延、网络扩展性、网络管理等。分析和仿真结果表明,本架构能够对信息进行有效的传输,具有良好的可扩展性。
In the current network, the traffic increase very fast because of more and more services. But the transmission of the information is based entirely on information address, the content of the information is transparent to the network. There is so much information that is same transmitting in the network, resulting in a waste of bandwidth and energy. In the socioeconomic perspective, the current network structure makes low profits transmitting so mush content. So we hope the future network can provide different services to different groups and stakeholders. The future network needs a new structure to distribute information effectively.
The current network is not able to complete information distribution very effectively, so lots of research institutions study to design a new structure. As the core of the network is no longer the nodes, but the information, so the future network is information-centric network. Now some research institutions have some preliminary results, such as NetInf, NDN, PSIRP, COMNET. But there does not have a unified information-centric network, so this paper will analyze some problems that still exist in the information-centric network. For the existing problems, this paper proposes an information-centric network based on SDN. This paper defines the design of the architecture, the separation of the control plane and data plane. The control plane is responsible for the distribution of the information, registration, cache management and the data plane for forwarding. In the paper, we focus on the information distribution according to the information feature and network states. In addition, the architecture also supports mobility management and multicast, which has provided a guarantee for more efficient transmission of information.
After the completion of the design and implementation of the architecture, this paper evaluate of the performance of the architecture. In performance assessment, we use the theoretical analysis and also use OpenFlow and Mininet in the network to simulate including the configuration of the information distribution paths, mobility management, scalability. Analysis and simulation results show that the architecture has good performance on information transmission, mobility management, scalability.
本文首先调研了信息网络的研究现状,介绍了几种现有的信息网络模型：NetInf、NDN、PSIRP。并对信息网络中的关键技术进行了总结,分析了现有信息网络架构中存在的问题。同时介绍了SDN技术及研究现状,总结了SDN和信息网络结合带来的好处。
本文提出了一种基于SDN的信息网络架构来解决当前网络中存在的问题,同时尽量避免现有信息网络中存在的问题。论文中介绍了基于SDN技术设计的信息网络架构,在设计中,采用控制平面和数据平面相分离,控制平面负责信息的分发控制、注册、检索、缓存管理等功能,数据平面进行数据的转发。在本论文中,设计了信息的命名和分发路径构造方案,该命名即能够保证信息的自认证性又能够为信息的分发提供依据；分发路径的构造是整个网络架构中最重要的部分,也是本架构最有特色的部分。为了保证信息传输路径的有效性,在信息传输前,网络需根据信息的特性和网络状况为信息构造分发路径。另外,本网络架构也支持移动性管理和组播,从而为信息的有效传输提供了保障。
在完成架构的设计及实现后,本论文对架构进行了性能评估。在性能评估中,采用了理论分析与仿真结果相结合的方法,使用Open Flow和Mininet对网络进行了仿真。评估指标包括分发路径构造的开销、信息的传输时延、网络扩展性、网络管理等。分析和仿真结果表明,本架构能够对信息进行有效的传输,具有良好的可扩展性。
In the current network, the traffic increase very fast because of more and more services. But the transmission of the information is based entirely on information address, the content of the information is transparent to the network. There is so much information that is same transmitting in the network, resulting in a waste of bandwidth and energy. In the socioeconomic perspective, the current network structure makes low profits transmitting so mush content. So we hope the future network can provide different services to different groups and stakeholders. The future network needs a new structure to distribute information effectively.
The current network is not able to complete information distribution very effectively, so lots of research institutions study to design a new structure. As the core of the network is no longer the nodes, but the information, so the future network is information-centric network. Now some research institutions have some preliminary results, such as NetInf, NDN, PSIRP, COMNET. But there does not have a unified information-centric network, so this paper will analyze some problems that still exist in the information-centric network. For the existing problems, this paper proposes an information-centric network based on SDN. This paper defines the design of the architecture, the separation of the control plane and data plane. The control plane is responsible for the distribution of the information, registration, cache management and the data plane for forwarding. In the paper, we focus on the information distribution according to the information feature and network states. In addition, the architecture also supports mobility management and multicast, which has provided a guarantee for more efficient transmission of information.
After the completion of the design and implementation of the architecture, this paper evaluate of the performance of the architecture. In performance assessment, we use the theoretical analysis and also use OpenFlow and Mininet in the network to simulate including the configuration of the information distribution paths, mobility management, scalability. Analysis and simulation results show that the architecture has good performance on information transmission, mobility management, scalability.
引文
[1]Cisco System. http://www.cisco.com/en/US/netsol/ns827/networking_soluti ons_sub_solution.html#-forecast
[2]Randy H.K, David E.C, Seth S. An information-centric energy infrastru cture:The Berkeley view, Sustainable Computing:Informat-ics and Systems, vol. 1:7-22,2011.
[3]FIA-Future Content Networks Group. "Why do we need a Content-Ce ntric InternetProposals towards Content-Centric Internet Architectures".2009
[4]http://www.netinf.org
[5]http://www.named-data.net
[6]http://opennetsummit.org/
[7]https://www.opennetworking.org/
[8]H.Yin, H.Xie. SDNi:A Message Exchange Protocol for Software Defin ed Networks (SDNS) across Multiple Domains. IETF draft.2012
[9]H.Xie, T.Tsou. Use Cases for ALTO with Software Defined Networks. IETF draft.2012.
[10]J. Medved, A. McLachlan. MPLS-TP PseudoWire Configuration usi ng OpenFlow 1.3. IETF draft.2012.
[11]Li Erran Li, Morley Mao. Toward Software-Defined Cellular Netwo rks.2010
[12]Peter Dely, Andreas Kassler. OpenFlow for Wireless Mesh Network s.2011
[13]Lixia Zhang, D.Estrin, J.Burke. Named Data Networking (NDN) Pr oject. PARC Technical Report NDN-0001. Oct.2010
[14]B.Ahlgren, M.D'Ambrosio, C.Dannewitz. Second NetInf Architectur e Description. FP7-ICT-2007-1-216041-4WARD/D-6.2. Jan.2010
[15]http://www.psirp.org/
[16]M.Ain, D. Trossen, P.Nikander. Publish-Subscribe Internet Routing Paradigm:Architecture Definition, Component Descriptions, and Requirements. FP7-INFSO-ICT-216173-PSIRP-D2.3. Feb.2009
[17]P. Jokela, A.Zahemszky, S. Arianfar. LIPSIN:Line Speed Publish/S ubscribe Inter-Networking. TR09-0001, FP7-INFSO-IST 216173.June 2008
[18]http://www.comet-project.org/
[19]Ioannis P, George P, Gerardo G, D2.2 High-Level Architecture of t he COMET System, European Seventh Framework Project FP7-2010-ICT-24878 4-STREP,2010
[20]George K, Ning W, D4.2 Final Specification of Mechanisms, Proto cols and Algorithms for Enhanced Network Platforms, European Seventh Frame work Project FP7-2010-ICT-248784-STREP,2010
[21]严蔚敏,吴伟民.数据结构(C语言版),清华大学出版社,2007,187
[22]Nadira J, Arslanagic E, Kurtovic I, Dijkstra's shortest path algorith m serial and parallel execution performance analysis,2012 Proceedings of the 35th International Convention, page(s):1811-1815
[23]http://www.emc.com/domains/rsa/index.htm
[24]ITU-T Study Group 2. Definitions of terms related to quality of se rvice, Recommendation. ITU-T E.800. Sept.2008
[25]C. Dannewitz, J. Golic, B. Ohlman. Secure naming for a network of information. In Proceeding 13th IEEE Global Internet Symposium 2010, San Diego, USA, Mar.2010
[26]http://www.openflow.org/wp/learnmore/
[27]OpenFlow Switch Specification 1.0.0, https://www.opennetworking.or g/images/stories/downloads/specification/openflow-spec-v1.0.0.pdf
[28]http://www.noxrepo.org/nox/about-nox/
[29]https://openflow.stanford.edu/display/Beacon/Home
[30]http://yuba.stanford.edu/foswiki/bin/view/OpenFlow/MininetLearnMore
[31]YIN Chao, WANG Hongxia. Developed Dijkstra Shortest Path Sear ch Algorithm and Simulation. ICCDA 2010
[32]Jarschel, M.& S. Oechsner, D. Schlosser, R. Pries, S. Goll, Phuoc Tran-Gia. Modeling and Performance Evaluation of an OpenFlow Architecture [C]. ITC,2011:1-7
[33]Bianco, A.& R. Birke, L. Giraudo. OpenFlow Switching:Data Pla ne Performance[C]. Cape Town:ICC,2010:1-5
[34]Azodolmolky, S.& R. Nejabati, E. Escalona, R. Jayakumar, N. Efs tathiou, D. Simeonidou. Integrated OpenFlow-GMPLS Control Plane:An Overl ay Model for Software Defined Packet over Optical Networks[C]. Geneva, Swit zerland:ECOC,2011
[35]Beacon, https://openflow.stanford.edu/display/Beacon/Home
[36]DiegoPerino;MatteoVarvello;A reality Check for Content Centric Net working.InProc.of ACM ICN1011,August19,2011,pp:44-49
[2]Randy H.K, David E.C, Seth S. An information-centric energy infrastru cture:The Berkeley view, Sustainable Computing:Informat-ics and Systems, vol. 1:7-22,2011.
[3]FIA-Future Content Networks Group. "Why do we need a Content-Ce ntric InternetProposals towards Content-Centric Internet Architectures".2009
[4]http://www.netinf.org
[5]http://www.named-data.net
[6]http://opennetsummit.org/
[7]https://www.opennetworking.org/
[8]H.Yin, H.Xie. SDNi:A Message Exchange Protocol for Software Defin ed Networks (SDNS) across Multiple Domains. IETF draft.2012
[9]H.Xie, T.Tsou. Use Cases for ALTO with Software Defined Networks. IETF draft.2012.
[10]J. Medved, A. McLachlan. MPLS-TP PseudoWire Configuration usi ng OpenFlow 1.3. IETF draft.2012.
[11]Li Erran Li, Morley Mao. Toward Software-Defined Cellular Netwo rks.2010
[12]Peter Dely, Andreas Kassler. OpenFlow for Wireless Mesh Network s.2011
[13]Lixia Zhang, D.Estrin, J.Burke. Named Data Networking (NDN) Pr oject. PARC Technical Report NDN-0001. Oct.2010
[14]B.Ahlgren, M.D'Ambrosio, C.Dannewitz. Second NetInf Architectur e Description. FP7-ICT-2007-1-216041-4WARD/D-6.2. Jan.2010
[15]http://www.psirp.org/
[16]M.Ain, D. Trossen, P.Nikander. Publish-Subscribe Internet Routing Paradigm:Architecture Definition, Component Descriptions, and Requirements. FP7-INFSO-ICT-216173-PSIRP-D2.3. Feb.2009
[17]P. Jokela, A.Zahemszky, S. Arianfar. LIPSIN:Line Speed Publish/S ubscribe Inter-Networking. TR09-0001, FP7-INFSO-IST 216173.June 2008
[18]http://www.comet-project.org/
[19]Ioannis P, George P, Gerardo G, D2.2 High-Level Architecture of t he COMET System, European Seventh Framework Project FP7-2010-ICT-24878 4-STREP,2010
[20]George K, Ning W, D4.2 Final Specification of Mechanisms, Proto cols and Algorithms for Enhanced Network Platforms, European Seventh Frame work Project FP7-2010-ICT-248784-STREP,2010
[21]严蔚敏,吴伟民.数据结构(C语言版),清华大学出版社,2007,187
[22]Nadira J, Arslanagic E, Kurtovic I, Dijkstra's shortest path algorith m serial and parallel execution performance analysis,2012 Proceedings of the 35th International Convention, page(s):1811-1815
[23]http://www.emc.com/domains/rsa/index.htm
[24]ITU-T Study Group 2. Definitions of terms related to quality of se rvice, Recommendation. ITU-T E.800. Sept.2008
[25]C. Dannewitz, J. Golic, B. Ohlman. Secure naming for a network of information. In Proceeding 13th IEEE Global Internet Symposium 2010, San Diego, USA, Mar.2010
[26]http://www.openflow.org/wp/learnmore/
[27]OpenFlow Switch Specification 1.0.0, https://www.opennetworking.or g/images/stories/downloads/specification/openflow-spec-v1.0.0.pdf
[28]http://www.noxrepo.org/nox/about-nox/
[29]https://openflow.stanford.edu/display/Beacon/Home
[30]http://yuba.stanford.edu/foswiki/bin/view/OpenFlow/MininetLearnMore
[31]YIN Chao, WANG Hongxia. Developed Dijkstra Shortest Path Sear ch Algorithm and Simulation. ICCDA 2010
[32]Jarschel, M.& S. Oechsner, D. Schlosser, R. Pries, S. Goll, Phuoc Tran-Gia. Modeling and Performance Evaluation of an OpenFlow Architecture [C]. ITC,2011:1-7
[33]Bianco, A.& R. Birke, L. Giraudo. OpenFlow Switching:Data Pla ne Performance[C]. Cape Town:ICC,2010:1-5
[34]Azodolmolky, S.& R. Nejabati, E. Escalona, R. Jayakumar, N. Efs tathiou, D. Simeonidou. Integrated OpenFlow-GMPLS Control Plane:An Overl ay Model for Software Defined Packet over Optical Networks[C]. Geneva, Swit zerland:ECOC,2011
[35]Beacon, https://openflow.stanford.edu/display/Beacon/Home
[36]DiegoPerino;MatteoVarvello;A reality Check for Content Centric Net working.InProc.of ACM ICN1011,August19,2011,pp:44-49