Adaptive Performance Evaluation for SDN Based on the Statistical and Evolutionary Algorithms

Abstract

Being able to send different types of data (i.e. text, audio, or video) through the network is the most important aspect of networks. Different networks have different issues and restrictions while sending data. These restrictions are basically the QoS (Quality of Service) metrics and security. The recent Software-Defined Networking (SDN) that aims to separate the control plane from the data plane can be applied where Business requirements are not responsible for the way the network is configured; instead, it is the responsibility of the high-level business policies and objectives. SDN gives preferable techniques for centralized dynamic management and control configurations. In this work, a proposed model has been estimated and discussed to promote QoS requirements in some suggested topologies. Adaptive Resource Management (ARM) and control to send different types of data through different hosts have been investigated. The intended requirements are basically the capacity and delay of traffic metrics sent through different hosts through the network. It produces a mathematical model and implementation for three proposed algorithms to enhance the quality of a sample video sent from source host to destination host by Visible Light Communication (VLC)-media player in three different topologies. These algorithms (statistical, MOGA, and PSO) have been implemented using Mininet emulator, FNSS tool, PULP, and network libraries; with two types of controllers which are Floodlight and OVS under Linux operating system and in python programming language.