ERSA’2009
The International Conference on
Engineering of Reconfigurable Systems and Algorithms

Networked Self-Adaptive Systems: An Opportunity for Configuring in the Large.

Dr. Jean-Philippe Diguet, University of South Britanny, France

Self-adaptivity is a solution to give decision intelligence to embedded systems in order to dynamically adapt hardware / software architectures to environment variations, data changes and user requirements according for instance to energy efficiency. In our approach, this choice is done at run-time and based on a set of embedded configurations (binaries and bitstream in case of FPGAs). This choice offers fast reconfiguration but can also mean a restricted configuration space regarding multiple application systems. However usual network capabilities can enlarge this configuration space towards infinity. In this paper we present a combination of two techniques to propose a global configuration management. At the level of embedded system, we implement self-adaptivity by means of a close-loop approach and on a Borda vote. This aspect is demonstrated with a smartcam case study. Then, at the network level, we propose a hierarchy of caches based on system modeling to manage configuration in the large. A bi-processor architecture for audio coding illustrates this second point. Finally we draw some perspectives regarding this promising domain.

Jean-Philippe Diguet obtained his PhD from Rennes I University, France, in 1996. In 1997, he was a visitor researcher at IMEC. Between 1997 and 2002, he has been associate professor at UBS University in Lorient. Granted by Anvar in 2002, he has been in charge of a technology transfer in the domain of ambient computing. Since 2004, he is full time researcher at CNRS and a member of Lab-STICC. He has been involved in many research projects funded by UE, French Government, competitiveness centers and CNRS. In the domain of architecture for embedded systems, his main research topics are CAD tools for design space exploration, NoCs design and power consumption estimation, secure processor architectures, reconfigurable and self-adaptive architectures. Regarding applications, he focused on multimedia, digital communications and ambient computing for people assistance.