Building Fault Tolerant, Scalable and High Speed Multicontroller Environment in Software Defined Networking
Abstract
SDN controller is prone to failure at a single point as there is usually a breakdown of some issues in the controller; due to this, the entire system's performance is bottlenecked and limited. To improve the overall security, scalability, and reliability of Software-defined networking, many security utilities and some of the applications must be built upon SDN capabilities. In this implementation paper, we aim to build a fast, highly scalable, efficient fault-tolerant multi-controller environment with the help of RYU controllers. We look into the problems of the SDN controller with a focus mainly on efficiency, performance, and also the availability of the controllers. In this paper, we have modified the single controller environment with that of a multi-controller environment also we have put into practice Apache Zookeeper which runs the master-slave algorithm to always elect a master controller in a pool of controllers to take over the previous controller's data, in addition to this a Redis Server has also been used to reduce the processing time, all this ensure that the environment is fault-tolerant.
Keywords: Openflow, RYU, REDIS, Software Defined Networking, ZooKeeper.