Introducing edge controlling to software defined networking to reduce processing time

Abstract

Edge computing enables computational services to be performed at the edge of a network on behalf of the main cloud services. The rationale of edge computing is applicable to any centralized system where the service should be provided at the proximity of the end-systems. Software-Defined Networking (SDN) is an evolving network model that runs a single-centralized controller to handle all network functions which may lead to limited performance capabilities with the recent growth of data communication. In this paper, we introduce a novel SDN Edge Controlling model that overcomes the performance-limitations by leveraging edge computing capabilities. It aims to bring the computing and storage resources close to network devices so that the load on the main SDN controller can be eased, and the delay between the forward plane and the control plane is minimized. We develop a simulation program to assess the effectiveness of our model. Experimental results show that the bandwidth usage is reduced by about 45% and the total processing time of the main controller is reduced by almost 62% for 10,000 requests. Therefore, the proposed model handles a higher network load and maintains lower latency when compared with a traditional SDN.