Energy Efficiency in Main-Memory Databases

Abstract

As the operating costs of today’s data centres continue to increase and processor man- ufacturers are forced to meet thermal design power constraints, the energy efficiency of a main-memory database management system becomes more and more important. In this paper, we experi- mentally study the impact of reducing the clock frequency of the processor on the energy efficiency of common database algorithms such as scans, simple aggregations, simple hash joins and state-of- the-art join algorithms. We stress the fundamental trade-off between peak performance and energy efficiency, as opposed to the established race-to-idle strategy. Ultimately, we learn that database workloads behave considerably different than other workload types and that reducing the computing power e.g. by limiting the clock frequency can significantly improve energy efficiency.