Renewable and Temperature Aware Load Balancing for Energy Cost Minimization in Data Centers: A Study of BRT, Peshawar

Abstract

Power management in data centers (DSs) is a pressing contemporary concern, particularly in the context of cost reduction. DCs are the key source of energy consumption as they strive to meet customer demands, leading to negative impact on environment and elevated energy expenses. Existing research in this domain primarily focuses on task allocation techniques that leverage renewable energy sources and lower electricity rates to mitigate energy costs. In our research, we address the cost reduction problem in data centers and consider Bus Rapid Transit (BRT) service system as a case study in Peshawar, Pakistan. We introduce a novel strategy called Renewable and Temperature-aware Load Balancing (RTLB), which employs an online greedy algorithm design technique to optimize the processing of user requests within a DC. Our proposed algorithm considers various factors, including ambient and internal temperatures, on-site renewable energy availability, conventional energy consumption, active server count, and compliance with predefined constraints. The experiments and their results, conducted using real-world data, validate the higher performance of RTLB when compared to existing workload allocation strategies, ultimately reducing the overall operating expenses of the DC.