Energy-Efficient Controller Placement in Software-Defined Networks: A Heuristic Approach

S Mohanty and B Sahoo, ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING, 50, 20285-20316 (2025).

DOI: 10.1007/s13369-025-10553-8

The escalating concerns over energy consumption, carbon emissions, and environmental sustainability present critical challenges for large-scale data networks. Software-defined networking (SDN) has emerged as a solution, offering a paradigm shift with its centralized controller and separation of control and data planes. This paper delves into energy- efficient controller placement within SDN infrastructures to minimize active link utilization while meeting traffic demands and latency requirements. Heuristic algorithms provide an effective approach for tackling NP-hard problems such as energy-efficient controller placement (EHACP). To address this, three heuristic methods are evaluated: genetic algorithms (EHACP-GA), grey wolf optimization algorithm (EHACP-GWA), and harmony search algorithm (EHACP-HSA). Through a thorough analysis of real network topologies, this study demonstrates the efficacy of EHACP- HSA in attaining substantial energy savings, often surpassing 45% across several scenarios, while considerably minimizing the number of active links compared to other competing algorithms. In the AT & T network, EHACP-HSA achieves energy savings of up to 69.64%, alongside an approximate 15% decrease in active links relative to Greco and an 11% reduction compared to the enhanced genetic algorithm (EHACP-EGA), indicating its effectiveness in reducing active link usage. EHACP-GWA confirms its stability in latency performance by consistently exhibiting smaller latency gaps than EHACP-HSA, with an average gap of -\documentclass12ptminimal \usepackageamsmath \usepackagewasysym \usepackageamsfonts \usepackageamssymb \usepackageamsbsy \usepackagemathrsfs \usepackageupgreek \setlength\oddsidemargin-69pt \begindocument$$-$$\enddocument0.0012 and a low standard deviation. The results demonstrate that EHACP-HSA is the preferred choice for scenarios emphasizing energy efficiency and faster computation, whereas EHACP-GWA is better-suited for latency-sensitive applications.

Return to Publications page