Price-Sensitivity Aware Load Balancing for Geographically Distributed Internet Data Centers in Smart Grid Environment

Abstract

In this paper, we investigate the problem of developing a geographical load balancing (GLB) scheme for distributed Internet data centers (IDCs) when they are price-makers in the deregulated electricity markets, i.e., GLB may impose impact on electricity prices in IDC locations due to the large power consumption of IDCs. Taking into account the dynamic characteristics (e.g., time-varying power demands and generation outputs) and actual physical constraints (e.g., active/reactive power flow balancing, transmission congestion, and network loss) of smart grids, it is difficult to obtain price impact models with analytical expressions accurately. As a result, it is challenging to design an efficient GLB scheme without requiring price impact models. To overcome the above challenge, we propose a price-sensitivity aware GLB scheme. The key idea of the proposed scheme is to impose proper limits on the workloads allocated to the IDC locations with high price-sensitivity coefficients by exploiting a number of interactive information between IDCs and main grids, so that the sudden increase in the total cost could be avoided. Here, the price-sensitivity coefficient is defined as the ratio of the percentage change in price to the percentage change in IDC power demand. Extensive simulation results show the effectiveness of the proposed GLB scheme.