计算机科学 ›› 2018, Vol. 45 ›› Issue (1): 205-210.doi: 10.11896/j.issn.1002-137X.2018.01.036
何璐蓓,厉俊男,杨翔瑞,孙志刚
HE Lu-bei, LI Jun-nan, YANG Xiang-rui and SUN Zhi-gang
摘要: SDN采用转发与控制分离的架构和集中的控制管理机制,可有效满足不同网络中不同粒度的管理控制需求。当高校科研人员进行SDN的教学和创新实验时,需要一个处理过程可感且可重新编程的数据平面来支持原理展示和自主研究。然而,传统ASIC交换机的内部实现流程不透明且转发查表架构固定,软件交换机的处理性能较低,因此无法充分支持数据平面的研究。目前,通过FPGA设计可编程数据平面,为满足不同科研场景下多样化的处理需求提供了一条可行路径。但是,在基于FPGA的可重构交换机架构和设计方法方面还缺少深入研究,主要表现在难以实现基于模块细粒度的SDN处理流程重构,现有工作复用程度低,同时无法为开源的SDN数据平面设计提供技术支持。为此,提出一种基于FPGA的SDN交换平面实现结构——RESSP(FPGA-based REconfigurable SDN Swi-tching Pipeline)。RESSP将报文处理流程拆解成多个可动态加载的模块,针对交换机具体的应用场景,利用FPGA可编程特性对硬件功能模块进行增加、删除或替换,从而针对实际需求设计出相应的报文处理逻辑。此外,基于RESSP实现了一个SDN交换机的原型系统MiniSwitch。MiniSwitch验证了RESSP在教学科研实验中快速重构所需SDN数据平面的可行性和可扩展性。
[1] NADEAU T D,GRAY K.SDN:Software DeRESSPd Networks[M].O’Reilly Media,Inc,2013. [2] BOSSHART P,GIBBZ G,KIMY H S,et al.Forwarding Metamorphosis:Fast Programmable Match-Action Processing in Hardware for SDN[J].ACM SIGCOMM Computer Communication Review,2013,43(4):99-110. [3] Intel Ethernet Switch FM6000 Series[EB/OL].http://www.intel.com/content/www/us/en/ethernet-products/switch-silicon/ethernet-switch-fm5000-fm6000-series.html. [4] 电子发烧友.http://www.elecfans.com/emb/fpga/20150907382835.html. [5] NAOUS J,ERICKSON D,COVINGTON G A,et al.Implementing an OpenFlow switch on the NetFPGA platform[C]∥Proceeding of the 4th ACM/IEEE Symposium on Architectures for Networking and Communication System.2008:1-9. [6] HU C C,YANG J,GONG Z M,et al.DesktopDC:Setting AllProgrammable Data Center Networking Testbed on Desk [J].ACM SIGCOMM Computer Communication Review,2014,44(4):593-594. [7] JIA C B,HUANG J F,SU Q,et al.OpenFlow Implementation on NetMagic Platform[J].Applied Mechanics & Materials,2012,198-199:516-522. [8] CAO C Z,MAO J B,SUN Z G,et al.Method of NetMagic hardware development[J].Computer Engineering & Science,2014,6(9):1678-1683.(in Chinese) 曹成周,毛健彪,孙志刚,等.NetMagic平台硬件开发方法[J].计算机工程与科学,2014,36(9):1678-1683. [9] FAST.http://fast-switch.github.io. [10] ONF.OpenFlow Table Type Pattrns Version 1.0[S].2014. [11] JAIN S,KUMAR A,MANDAL S,et al.B4:experience with a globally-deployed software defined wan[J].ACM SIGCOMM Computer Communication Review,2013,43(4):3-14. [12] BOSSHART P,DALY D,GIBBY G,et al.P4:ProgrammingProtocol-Independent Packet Processors[J].Computer Communication Review:A Quarterly Publication of the Special Interest Group on Data Communication.2014,44(3):88-95. [13] SONG H Y.Unaware Routing Protocol within OpenFlow2.0[J].Communications of the CCF,2015,1(11):35-41. [14] The world’s fastest and most programmable networks[EB/OL].https://barefootnetworks.com/media/white_papers/Barefoot-Worlds-Fastest-Most-Programmable-Networks.pdf. [15] GIBB G,VARGHESE G,HOROWITZ M,et al.Design Principles for Packet Parsers[C]∥2013 ACM/IEEE Symposium on Architectures for Networking and Communications Systems (ANCS).2013:13-24. [16] DNLAB.http://www.sdnlab.com/1710.html. [17] TECHCON A.Xilinx Introduces SDNet & ‘Softly’ DeRESSPdNetworks.http://www.xilinx.com/products/design-tools/software-zore/sdnet.html. [18] Xilinx SDNet:A New Way to Specify Network Hardware[EB/OL].http://www.xilinx.com/publications/prod_mktg/linley-group-sdnet-wp.pdf. [19] Specification,OpenFlow Switch.Version 1.1.0 Implemented(Wire Protocol 0x02)[S].2011. [20] SHELLY N,JACKSON E J,KOPONEN T,et al.Flow Caching for High Entropy Packet Fields[J].ACM SIGCOMM Computer Communication Review,2014,44(4):151-156. |
No related articles found! |
|