Energy management and profit maximization of green data centers
MetadataShow full item record
The demand for Internet and Cloud Computing services has been constantly increasing in the United States and elsewhere over the past few years. In response to the users' growing needs for Internet and computational services, major companies such as Google, Microsoft and Amazon have built several data centers at a variety of geographical locations, each one including hundreds of thousands of computer servers. There is a huge amount of power consumption associated with each data center due to running several servers, plus the electricity needed for the cooling and lightening equipments. While a large body of work has recently focused on reducing data center's energy expenses, there exists no prior work on investigating the trade-off between minimizing data center's energy expenditure and maximizing their revenue for various Internet and cloud computing services that they may offer. In this study, we seek to tackle this shortcoming by proposing a systematic approach to maximize a green data center's profit, i.e., revenue minus cost. In this regard, we explicitly take into account practical service-level-agreements (SLAs) that currently exist between data centers and their customers. Our model also incorporates various other factors such as availability of local renewable power generation at data centers and the stochastic nature of data centers' workload. Furthermore, we propose a novel optimization-based profit maximization strategy for data centers for two different cases, without and with behind-the-meter renewable generators. We show that the formulated optimization problems in both cases are convex programs; therefore, they are tractable and appropriate for practical implementation. Finally, we propose a workload distribution algorithm to utilize locational diversity of data centers that are built at different regions and may face different prices of electricity and different renewable power availability at different times of day. Using various experimental data and via computer simulations, we assess the performance of the proposed optimization-based strategy and show that it significantly outperforms two comparable energy and performance management algorithms that have recently been proposed in the literature.