The energy consumption of IT and communication infrastructure is dramatically increasing. For instance the energy consumption of mid-range, and high-end servers in the U.S. and worldwide has doubled from 2000 to 2005.
In the year 2000 the power consumption of servers, routers, and PCs in Germany was about 5 billion kWh per year. For the year 2010 a power consumption of more than 55 billion kWh per year is expected for the IT and communication infrastructure in Germany.
The G-Lab infrastructure also consumes a significant amount of power. Even if these servers operate with a very low utilization, they often need about 70% of their maximum power consumption. Half of the energy needed in G-Lab is used for cooling the equipment (similar to data centers).
Two major goals are considered in the Ener-G project:
1. The first goal is to analyze the energy consumption of the current G-Lab infrastructure and to optimize it with respect to energy efficiency. This aspect is not yet explicitly addressed in the G-Lab project, therefore a high optimization potential in the G-Lab energy consumption is expected. A reduction of the energy consumption directly reduces the operational costs of G-Lab and also its environmental impact (CO2 emissions). This increases the sustainability of the G-Lab project significantly. When the funding of the project ends after 3 years, G-Lab is intended to stay operational as long as possible. If the operational costs of G-Lab are decreased, it will be easier to finance it after the funding ends. Additionally, an energy-efficient operation of infrastructures is a key feature of the Future Internet. Increasing costs of energy and the intended reduction of CO2 emissions are driving forces in the development of future technologies. The envisioned energy efficiency of G-Lab will be a unique feature which is not explicitly addressed in other comparable Future Internet platforms like GENI.
2. The second goal of Ener-G is to explore the energy-efficient operation of other future IT and communication infrastructures. Not only G-Lab will benefit from this research, the results can also be adapted to related areas (e.g. data centers, cloud computing, grids, super computers, etc.). The Future Internet and its infrastructures have to be structured in a way that inherently supports an energy-efficient operation. The results of this project might strongly influence the hardware and software development of such future infrastructures. G-Lab is an excellent opportunity to investigate energy-efficient future infrastructures. Since G-Lab is still under development in terms of hardware, network, protocols, and services, it seems to be an ideal platform to integrate energy efficiency from the beginning. It is not sufficient to just optimize hardware, as it is mainly done today. An energy efficient operation of an infrastructure like G-Lab can only be achieved if all relevant layers are addressed, including communication protocols, virtualization of resources, and application-layer services.