Multi-Objective Genetic Algorithm for the Joint Optimization of Energy Efficiency and Rule Reduction in Software-Defined Networks

Abstract

Although the use of Ternary Content-Addressable Memories (TCAMs) in the flow tables of the Software-Defined Network (SDN) switches increases the efficiency of packets matching procedure, drawbacks such as their large power consumption and the limitation on the number of flow rules that can be installed must be taken into account. This paper tackles the joint problem of power consumption and TCAM size limitation in SDN. By exploiting the Rate Adaptation technique and compression methods, a Multi-Objective Genetic Algorithm is proposed to practically solve it. Simulations on a real network topology show that our proposed solution outperforms other state-of-the-art approaches, both in terms of power saving gains (20% at non-peak TM) and maximum TCAM utilization (5%).