2015 IEEE International Conference on Autonomic Computing (ICAC) (2015)
July 7, 2015 to July 10, 2015
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/ICAC.2015.39
Content delivery networks (CDNs) employ hundreds of data centers that are distributed across various geographical locations. These data centers consume a significant amount of energy to power and cool their servers. This paper investigates the joint effectiveness of using two new cooling technologies - open air cooling (OAC) and thermal energy storage (TES) - in CDNs to reduce their dependence on traditional chiller-based cooling and minimize its energy costs. Our Lyapunov-based online algorithm optimally distributes workload to data centers leveraging price and weather variations. We conduct a trace based simulation using weather data from NOAA and workload data from a global CDN. Our results show that CDNs can achieve at least 64% and 98% cooling energy savings during summer and winter respectively. Further, CDNs can significantly reduce their cooling energy footprint by switching to renewable open air cooling. We also empirically evaluate our approach and show that it performs optimally.
Cooling, Servers, Meteorology, Energy storage, Thermal energy, Pricing, Data models
S. Lee, R. Urgaonkar, R. Sitaraman and P. Shenoy, "Cost Minimization Using Renewable Cooling and Thermal Energy Storage in CDNs," 2015 IEEE International Conference on Autonomic Computing (ICAC), Grenoble, France, 2015, pp. 121-126.