Forschungsprojekte

SFB 912 - HAEC - Highly Adaptive Energy-efficient Computing

The way we use the Internet today has an enormous ecological impact on the increasing energy demand of modern society. For example, the electricity required by the servers that make up the Internet relates to about 2\% of the overall electricity consumption in the US1 and to about 26\% of the overall energy consumption in Germany. The mission of the collaborative research center "Highly Adaptive Energy-Efficient Computing" (HAEC) is to enable high energy efficiency in today's computing systems without compromising on high performance.

Certainly a straightforward way for improving energy efficiency would be to reduce the energy consumption of every individual hardware component involved. However, for many components the optimal energy performance point has already been reached. More important is an understanding of how software can adapt to hardware components and vice versa to address the computational problems of modern society in an energy efficient way. This requires new methods and tools to write energy-aware programs, new ways of interaction between the individual pieces that collaborate to solve a problem, new communication technologies to enable this interaction between pieces that are spread across highly scalable parallel systems, and a new multi-layer coordination infrastructure to bring together these technologies. In other words, we need an integrated approach for highly adaptive energy-efficient computing to approach energy-efficient computing at all involved technology levels.

The collaborative research center HAEC is a first attempt to achieve high adaptivity and energy efficiency in such an integrated approach. At the circuit level, we focus on innovative ideas for optical and wireless chip-to-chip communication. At the network level, we research secure, high performance network coding schemes for wired and wireless board-to-board communication. Innovative results at the hardware/software interface level will include energy control loops, which allow hardware to adapt to varying software requirements and vice versa. Software development in general is supported by energy-aware runtimes, energy-aware resource, stream and configuration management schemes and by an analysis framework for high performance/low energy applications. New internet applications are supported by innovations in energy-aware service execution. And, last but not least, formal methods are developed to offer a new quality of assurance in our systems of tomorrow. Demonstrating our results in a joint prototype - the HAEC Box - our goal is to become a pace setter for industry and academia on the design of future energy efficient-computing systems.

The objective of our subproject is to research into energy-aware software architectures. Such architectures are self-adaptive, based on a (dynamic) architectural language with a component model with energy-QoS-contracts, and appropriate run-time structures for dynamic QoS-contract negotiation (reflective meta-level with decider, reconfigurer). In such a self-adaptive energy-aware system, the reflective meta-level will, based on changes in the user requirements, context or energy availability, trigger reconfigurations of the software architecture at specific checkpoints at run-time and ensure safe migration of state of applications. Specific unique points of the envisaged architecture are:

The contract (re-)negotiation and reconfiguration will take into account energy monitoring, analysis and the evaluation of energy-utility functions in energy-QoS contracts. To this end, results from energy-QoS contract analysis will be taken into account. In the first phase of HAEC, mainly energy-QoS-contracts were in focus. In the second phase, we approach the scalability and efficiency of the developed multi-quality auto-tuning approach.

The architectural language will be used to synchronize the application architecture level with the operating system (subproject B.4/H\"artig), as well as the service management in the cloud (subproject B.6) so that reconfiguration is cross-cutting all layers of the HAEC software stack. Because from the architecture course-grain events can be generated to be communicated to these layers, it is expected that reconfigurations can be done in the right grain size and do not have to be performed too often.

  • Gefördert von: DFG
  • Kontaktperson: Sebastian Götz
  • Webseite des Projekts: link
  • Projektlaufzeit: 07/2011 bis 06/2019

Zu dieser Seite

Sebastian Götz
Letzte Änderung: 12.07.2016