The focus of the ecoDB project is on designing, implementing, and evaluating methods to reduce the total operating cost (TOC) for data-intensive workloads. This project focuses on various aspects of data center operations including the selection of the initial hardware, impact of energy consumption, selection of different component for the IO subsystem, etc. — all with a focus on the implications and impact of these choices on the TOC.

People: Willis Lang, Yinan Li, Ning Zhang, Jignesh M. Patel

• When does building data centers using mobile processors (e.g. ATOM) versus traditional server processors (e.g. Xeon) make sense? Published in DaMoN’10.
• How does one reduce the energy consumption of clusters running MapReduce workloads, with performance-based SLA constraints? In VLDB’10.