The rapid development of information and communication technologies (ICTs) and their penetration of all sectors of life make a lot of very positive effects possible, such as increasing efficiency, virtual products and services. The downside of this development is its part in the significant increase of CO2 emissions which already exceeds that of the air transport. So it is economically viable and environmentally necessary to improve the energy efficiency of ICT systems significantly.
In the area of energy efficiency in microelectronics OFFIS is already internationally established by its concepts and tools for low power design and part of a European network of the major semiconductor manufacturers and system integrators. In addition, OFFIS participates in national working groups and develops novel approaches to enhance efficiency, based on innovative load and energy management solutions to reduce the energy consumption of computers and especially data centers.
Many products today are only possible because of the integrated electronics. The design of modern embedded hardware- and software-systems presents industry and research with new challenges. These arise both by the rapidly advancing manufacturing technology that enables increasingly complex system architectures in a small space, but also from the increasing requirements of modern applications for processing speed and efficiency. OFFIS has been devoted so many years of modeling, analysis, optimization and of automated synthesis of embedded HW / SW-systems concerning performance, power consumption, robustness, space, and ultimately cost.Therefore the researc group EEI develops tools, such as the tool ORINOCO / PowerOpt which resulted in a successful spin-off, the ChipVision Design Systems AG.
Ecological, economic but in the meantime also technical problems force action in the optimization of energy consumption in data centers. Great potential for optimization lies in the hitherto largely static binding of services of data centers to the corresponding hardware resources working off the computational load, which resulted in severe underutilization. Load and Power Management can help here to greatly increase the efficiency of the use of hardware and energy by adapting dynamically the resources to the fluctuating demand. Necessary control and monitoring capabilities for the realization of such a solution already exist, however concepts for the use of these opportunities are still missing. Therefore, at OFFIS solutions are being developed as well for a data center's internal as a cross-data center load management.
