K. Nguyen, H. Mahkoum, B. Jaumard, C. Assi, Marc Lanoue
{"title":"Toward a Distributed Control Plane Architecture for Next Generation Routers","authors":"K. Nguyen, H. Mahkoum, B. Jaumard, C. Assi, Marc Lanoue","doi":"10.1109/ECUMN.2007.53","DOIUrl":null,"url":null,"abstract":"As networks become more and more pervasive and broadband access technologies evolve to provide affordable broadband solutions for both homes and offices, one needs to rethink the architecture design for next generation routers in order to be able to deliver multiple services over high-speed interfaces and perform fast switching capabilities (e.g., multipetabit). The size and complexity of next generation networks requires a protocol transition from its classical centralized approach to a fully distributed and scalable paradigm. In this paper, we propose a new control plane architecture that is resilient and scalable where we distribute the OSPF/IS-IS, MPLS and BGP protocols using IETF-approved inter-protocol interfaces. In such architecture, forwarding tables will be updated by OSPF, IS-IS, RTM, BGP and MPLS routing protocols distributed onto multiple processors, in the interface cards themselves. We therefore revisit each of these protocols and redraw their implementation within an efficient, scalable, distributed, resilient, and fault tolerant scheme. As a proof of concept, we present the hyperchip router (PBR1280) that has been built on these grounds and discuss the performance of the proposed distributed architecture. Although some features (e.g., resiliency) still need further development, it is shown that the proposed distributed routing protocol scheme already offers a new generation of routers with significant higher switching capacity","PeriodicalId":202819,"journal":{"name":"Fourth European Conference on Universal Multiservice Networks (ECUMN'07)","volume":"105 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2007-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"18","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Fourth European Conference on Universal Multiservice Networks (ECUMN'07)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ECUMN.2007.53","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 18
Abstract
As networks become more and more pervasive and broadband access technologies evolve to provide affordable broadband solutions for both homes and offices, one needs to rethink the architecture design for next generation routers in order to be able to deliver multiple services over high-speed interfaces and perform fast switching capabilities (e.g., multipetabit). The size and complexity of next generation networks requires a protocol transition from its classical centralized approach to a fully distributed and scalable paradigm. In this paper, we propose a new control plane architecture that is resilient and scalable where we distribute the OSPF/IS-IS, MPLS and BGP protocols using IETF-approved inter-protocol interfaces. In such architecture, forwarding tables will be updated by OSPF, IS-IS, RTM, BGP and MPLS routing protocols distributed onto multiple processors, in the interface cards themselves. We therefore revisit each of these protocols and redraw their implementation within an efficient, scalable, distributed, resilient, and fault tolerant scheme. As a proof of concept, we present the hyperchip router (PBR1280) that has been built on these grounds and discuss the performance of the proposed distributed architecture. Although some features (e.g., resiliency) still need further development, it is shown that the proposed distributed routing protocol scheme already offers a new generation of routers with significant higher switching capacity
随着网络变得越来越普及,宽带接入技术的发展为家庭和办公室提供了负担得起的宽带解决方案,人们需要重新考虑下一代路由器的架构设计,以便能够通过高速接口提供多种服务,并执行快速交换功能(例如,多pb)。下一代网络的规模和复杂性要求协议从传统的集中式方法过渡到完全分布式和可扩展的范式。在本文中,我们提出了一种新的控制平面架构,它具有弹性和可扩展性,其中我们使用ietf批准的协议间接口分发OSPF/ is - is, MPLS和BGP协议。在这种架构中,转发表将由分布在接口卡本身的多个处理器上的OSPF、IS-IS、RTM、BGP和MPLS路由协议更新。因此,我们重新审视这些协议,并在一个高效、可扩展、分布式、弹性和容错的方案中重新绘制它们的实现。作为概念验证,我们提出了基于这些基础构建的超芯片路由器(PBR1280),并讨论了所提出的分布式架构的性能。虽然一些特性(如弹性)还需要进一步发展,但研究表明,所提出的分布式路由协议方案已经提供了具有更高交换容量的新一代路由器