Cutting the electric bill for internet-scale systems
Proceedings of the ACM SIGCOMM 2009 conference on Data communication
INFOCOM'10 Proceedings of the 29th conference on Information communications
Greening geographical load balancing
Proceedings of the ACM SIGMETRICS joint international conference on Measurement and modeling of computer systems
Foundations and Trends® in Machine Learning
Proceedings of the ACM SIGCOMM 2012 conference on Applications, technologies, architectures, and protocols for computer communication
Dynamic provisioning in next-generation data centers with on-site power production
Proceedings of the fourth international conference on Future energy systems
Hi-index | 0.00 |
Datacenters consume an enormous amount of energy with significant financial and environmental costs. For geo-distributed datacenters, a workload management approach that routes user requests to locations with cheaper and cleaner electricity has been shown to be promising lately. We consider two key aspects that have not been explored in this approach. First, through empirical studies, we find that the energy efficiency of the cooling system depends directly on the ambient temperature, which exhibits a significant degree of geographical diversity. Temperature diversity can be used by workload management to reduce the overall cooling energy overhead. Second, energy consumption comes from not only interactive workloads driven by user requests, but also delay tolerant batch workloads that run at the back-end. The elastic nature of batch workloads can be exploited to further reduce the energy cost. In this work, we propose to make workload management for geo-distributed datacenters temperature aware. We formulate the problem as a joint optimization of request routing for interactive workloads and capacity allocation for batch workloads. We develop a distributed algorithm based on an m-block alternating direction method of multipliers (ADMM) algorithm that extends the classical 2-block algorithm. We prove the convergence and rate of convergence results under general assumptions. Trace-driven simulations demonstrate that our approach is able to provide 5%--20% overall cost savings for geo-distributed datacenters.