A Predictable performance Multi-controller-based Monitoring Framework for SDN

Abstract

Software defined networks (SDNs) have been released for dynamic operations of the networks and for controlling scalable networks. In SDN, the operating processes are controlled by a centralized controller.  The performance of the centralized controller is decreased by increasing the scale of the network.  Monitoring functionality is an essential element of any network system.  The monitoring performance under centralized controller is decreased with large scale networks.   In this paper, a proposed monitoring framework for multi-controller software defined networks is introduced. The introduced framework enhances the reliability and performance of large-scale software defined networks. Many copies of the proposed monitoring framework can run with SDN multi-controller for monitoring large-scale networks, detecting failure in any controller, and failover of the network failure.  All copies of the introduced monitoring framework run in parallel, each on a slice to enhance the performance of the SDN network.  All copies receive the requests from network applications, collect considerable amounts of measurements data, process them, and return the results to the network applications. The contribution of this paper is introducing a reliable and high-performance multi-controller-based SDN monitoring framework. The introduced monitoring framework monitors large scale SDN networks with good performance and enhances network reliability. It has the capabilities to monitor in parallel many network applications where each application runs on a network slice. Each slice is controlled by an SDN controller and monitored by a copy of the introduced monitoring framework. The copies of the introduced monitoring framework are communicated in a synchronization scheme.