Minimization of energy consumption in IP/SDN hybrid networks using genetic algorithms

Abstract

Although most of the efforts to tackle the energy consumption problem on wired communication networks have been made on traditional IP networks, energy-aware protocols require of several coordination tasks that are not easy to implement on current networking equipment. By exploiting the flexibility and programmability provided by the emerging Software-Defined Networking (SDN) paradigm, this paper investigates the problem of saving energy in hybrid IP/SDN networks where traditional IP nodes are incrementally replaced by SDN ones. To solve it, a Genetic Algorithm is proposed and evaluated through simulations over realistic network topologies. Results show that a reduced percentage such as 60% of SDN nodes would be enough to achieve significant reductions in energy consumption during the transition from IP to SDN, with a performance gap less than 7.5% compared to fully-deployed SDN networks.