Scalable workload management for water efficiency in data centers

The huge demand for data center computing nowadays has resulted in a significant amount of electricity consumption as well as environmental impacts. While current works mainly focus on the energy cost of data centers, the severity of water consumption problem in data centers is largely neglected. In this paper, we propose an optimization framework for the workload management of data centers, which takes the efficiency of water usage into account. The workload management is formulated as a revenue maximization problem. To solve the large-scale optimization problem with scalability, the alternating direction method of multipliers (ADMM) is utilized. The optimization problem is decomposed into independent subproblems, which can be solved in a parallel fashion on distributed computing units and coordinated through dual variables. We evaluate the performance of proposed algorithm by simulations, and numerical results validate the effectiveness of the proposed algorithm.