A Distributed Energy aware Controller Placement Model for Software-Defined Data Centre Network

Abstract

The single controller setup, in software-defined data center, is purposed to support dynamic load and scalability of switches. As such, multiple controller implementations become a promising approach to handle huge volumes of traffic. Co-locating the controllers along with switches becomes mandatory. In this article, we propose a controller placement model using independent dominating set, which identifies suitable positioning of controllers and assures minimum response time between controllers and switches. This creates an efficient switch to controller mapping with less latency. Since the control traffic influences the performance of software-defined network intensely. Moreover, with link energy as a metric, an energy aware routing mechanism, through the novel path ability component, is been proposed. It maintains the levels of path energy through which we route the traffic. Sustaining a unique module to maintain path energy metric increases the routing efficiency of the controller. Based on the dynamic update of path energy status, controllers periodically do the operations like switching off/on links and rerouting of packets. Traffic rerouting prevents congestion and avoids controller overhead during enormous traffic arrival. This adaptation ensures that our routing procedure utilizes less link energy during burst traffic compared to other existing approaches. The objective of this research work is to develop a controller placement model with an efficient routing strategy. It is capable of accomplishing low response time and prevents link energy depletion. The simulation outcomes indicate that the proposed algorithm achieves 70 to 227 watts less energy compared to the existing mechanisms with 52% increased throughput.