Energy Efficient Processing of Fine-Grained Loads in Heterogeneous Server Farms

The demand for energy is rising despite the fact that more and more efficient electronics are available. One cause for this is the increasing need for computing power. More and more data centers providing computing power, e.g. for ubiquitous computing. In former years, with decreasing transistor sizes, energy demand could be reduced while performance was increased. However, the effects in energy reduction achieved by miniaturization of semiconductor devices are diminishing. This leads to higher static power demand and thus, to rising energy consumption and heat development over time. There is an increasing need for energy efficiency in computing environments. In this paper, we present a two step approach to achieve an energy efficient operation of heterogeneous server farms: First, sophisticated load distribution algorithms are applied to generate distribution tables. Second, the potential of energy saving modes is leveraged. Idle times (so called gaps) are identified and the most suitable energy saving mode is chosen for each of these gaps.