Green scheduling algorithm for a computational grid environment

Abstract

Recently, the increasing demand for computing has led to the development of computational grids. Subsequently, the increased demand has also resulted in the significant heat management problem in the computational grid environment. The increased in computational workloads caused the servers to produce more heat, which consequently upsurges the cooling load. The extant research stated that the power supply for cooling such systems consumes quite a substantial portion (approximately 40% to 50%) of the total power consumption in a Data Center (DC). Current approaches addressing this problem include new cooling technology or optimizing cooling delivery. Alternatively, this paper proposes a green scheduling algorithm with the objective to minimize the cooling load by avoiding high thermal stress circumstances such as large hot spots and thermal violations while still maintaining the competitive performance in DC. The experimental results significantly showed that the proposed green scheduling algorithm was able to save the consumption of cooling load by 22.02% in DC's server room.