数据中心网络中一种半集中式SDN路由策略

在SDN体系架构中把网络控制功能从网络设备（交换机/路由器）里分离出来，集中到中心节点控制器上，交换机只负责数据平面的功能（通过流表进行数据转发）。在大规模的数据中心网络中，路由/流表的计算和分发完全由中心控制器完成，控制器成为网络的性能瓶颈和脆弱点。为了解决上述问题，本文提出半集中式SDN路由技术，其主要思想是每个交换机节点不需要控制器的参与，可以自主构建一个基础流表，基于基础流表，交换机可以完成基本的数据转发工作。而控制器负责更高级的路由选路（故障处理）工作，从而大大减轻控制器的负担。针对控制器的高级路由选路工作，本文通过对现有SDN网络中的故障恢复机制的特性以及限制的分析，在基础流表的基础上设计一套局部迂回故障检测恢复机制。基于该机制，控制器能够及时检测到网络故障，并在极短的时间内进行故障恢复，实现控制器的高级路由选路工作。

A Semi-centralized SDN Routing Strategy in Data Center Network

 In software defined network (SDN) architecture, the network control function is separated from network devices (switches/routers) and integrated to the central node controller, while switches are used to implement functions on the data plane (data forwarding based on the flow table). On large-scale data center networks, calculation and distribution of the routing/flow table are completely processed by the central controller, and the controller has become the bottleneck and vulnerability for network performance. To solve these problems, the essay proposes a semi-centralized SDN routing technology based on the data center network. Its main idea is that each switch node constructs a basic flow table for itself without involving the controller so that the switch can implement basic data forwarding based on the flow table. Meanwhile, the controller is responsible for routing (troubleshooting) of a higher level, thus greatly reducing controller’s loads. In this paper, advanced routing for the controller, a partial roundabout troubleshooting mechanism is designed based on the basic flow table through the analysis of the characteristics and limitations of the fault recovery mechanism in the existing SDN network. In this way, the controller can detect network failures in a timely manner and can recover the failures in short time to implement troubleshooting in advanced routing on controllers.