Details of the project:
According to CISCO, from 2007 our society consumes more energy by browsing the Internet than by air travel. It is also predicted that the energy footprint of computation and data traffic will steadily increase in the near future: Data centres will grow and so will the network infrastructure together with the number of terminal nodes of the global information network such as computers, phones, gadgets and other connected cyber-physical devices (so called Internet of Things). Energy-efficiency of components at all levels of the computation hierarchy thus became a major concern for the microelectronics industry. A serious factor that impedes progress in addressing this concern is a wide gap that exists between the ways of how energy efficiency is approached by hardware and software engineers, and this gap is matched by the lack of common understanding between the two communities.
This project aims to make an initial exploration into the challenge of bridging this gap by developing a shared design criterion, called power-proportionality, on the basis of which both electronics and programming solutions can be judged. The hypothesis that this criterion, albeit appearing to be quite specific, may be of fundamental nature for computing systems, needs checking in a proof-of-concept fashion. We thus believe that this makes this project ideally suited for eFutures, before it can be put forward for a more substantial grant application to EPSRC.
Published: 1st October 2012