The explosive growth of cloud computing services in recent years has led to significant expansion of data centers around the world and dramatically increased the overall electricity consumption, thereby resulting in a huge carbon footprint and severely impacting environment. As a consequence, data center operators have been increasingly urged to find effective solutions to achieve an overall net zero carbon footprint i.e., carbon neutrality. The state-of-the-art research addresses carbon neutrality based on accurate prediction of long-term future information that is typically unavailable in practice. In this chapter, we propose a provably-efficient online algorithm, called COCA (optimizing for COst minimization and CArbon neutrality), which minimizes the operational cost while satisfying the carbon neutrality without long-term future information a priori and in the presence of time-varying workloads and intermittent renewable energy supplies. We present a trace-based simulation study to validate the effectiveness of COCA, and the results show that COCA can outperform state-of-the-art prediction-based methods in terms of cost saving while achieving carbon neutrality. Moreover, we extend COCA to incorporate geographic load balancing to explore the geo-diversities of data centers for reducing the operational cost.
CITATION STYLE
Ahmed, K., Ren, S., He, Y., & Vasilakos, A. V. (2015). Online resource management for carbon-neutral cloud computing. In Handbook on Data Centers (pp. 607–630). Springer New York. https://doi.org/10.1007/978-1-4939-2092-1_20
Mendeley helps you to discover research relevant for your work.