Reducing Electricity and Network Cost for Online Service Providers in Geographically Located Internet Data Centers

Abstract

Online service providers(OSPs) have Internet data centers (IDCs) in multiple geographical locations in order to satisfy global user demand. Increased data centers consume a large amount of energy, and at the same time cause increased heat dissipation, greater cooling requirements, reduced computational density, and higher operating costs. It places a heavy burden on both the environment and energy resources. OSPs are now focusing more than ever on the need to improve energy efficiency. A new challenge has emerged besides the energy cost, the reduction of the carbon footprint. Although electricity is a clean and relatively safe form of energy to use, there are environmental impacts associated with the production and transmission of electricity. Meanwhile, increased online services require more Internet usage, especially with the trend of cloud computing, the network costs account for a large portion of operation costs for OSPs. This paper proposes an optimization load dispatching model to minimize the overall cost for OSPs. The electricity cost for IDCs and network cost are both taken into account. We conduct extensive evaluations based on real workload data and electricity price data. The results prove the effectiveness of our load dispatching model with a theoretically guaranteed quality of service(QoS).