Invited speaker at Datacenter Dynamics Converged Madrid 2012

Jose M. Moya attends as invited speaker to Datacenter Dynamics Converged Madrid 2012 (invited by ABB).


Current techniques for data center energy optimization, based on efficiency metrics like PUE, pPUE, ERE, DCcE, etc., Do not take into account the static and dynamic characteristics of the applications and resources (computing and cooling resources). However, the knowledge of the current thermal state of the datacenter, the past history, the thermal characteristics of resources, and the energy demand profile of the workload can be used very effectively to guide decision-making at all levels in the datacenter in order to minimize energy requirements. For example, the workload distribution in available computers, if done considering the architectures that better fits each job from the energy point of view, and taking into account the kind of jobs that will come after, can reduce IT energy consumption up to 30%.

Furthermore, to achieve a significant reduction in energy consumption of state-of-the-art datacenters (low PUE) is becoming increasingly important a comprehensive and multi-level approach, that is, acting on different levels of abstraction (job scheduling and resource allocation, application, operating system, compilers and virtual machines, architecture, and technology), and in the different domains (chip, server, rack, room and multi-room).

Slides (in Spanish):