T-Engine: Japan's Ubiquitous Computing Architecture Is Ready for Prime Time
IEEE Pervasive Computing
ElastIC: An Adaptive Self-Healing Architecture for Unpredictable Silicon
IEEE Design & Test
Thermal sensor allocation and placement for reconfigurable systems
Proceedings of the 2006 IEEE/ACM international conference on Computer-aided design
The Erlangen Slot Machine: A Dynamically Reconfigurable FPGA-based Computer
Journal of VLSI Signal Processing Systems
SAMOS '08 Proceedings of the 8th international workshop on Embedded Computer Systems: Architectures, Modeling, and Simulation
Self-Adaptive Networked Entities for Building Pervasive Computing Architectures
ICES '08 Proceedings of the 8th international conference on Evolvable Systems: From Biology to Hardware
RECONFIG '09 Proceedings of the 2009 International Conference on Reconfigurable Computing and FPGAs
Decentralized control for dynamically reconfigurable FPGA systems
Microprocessors & Microsystems
Hi-index | 0.00 |
The progress in hardware technologies lead to the possibility to embed more and more computing power in portable, low-power and low-cost electronic systems. Currently almost any everyday device such as cell phones, cars or PDAs uses at least one programmable processing element. It is forecasted that these devices will be more and more interconnected in order to form pervasive systems, enabling the users to compute everywhere at every time. This paper presents a FPGA-based self-reconfigurable platform for prototyping such future pervasive systems. The goal of this platform is to provide a generic template enabling the exploration of self-adaptation features at all levels of the computing framework (i.e. application, software, runtime architecture and hardware points of view) using a real implementation. Self-adaptation is provided to the platform by a set of closed loops comprising observation, control and actuators. Based on these loops (providing the platform with introspection), the platform can manage multiple applications (that may use parallelism) together with multiple areas able to be loaded on-demand with hardware accelerators during runtime. It can also be provided with self-healing using a model of itself. Finally, the accelerators implemented in hardware can learn how to perform their computation from a software golden model. Focusing on the low-level part of the computing framework, the paper aims at demonstrating the interest of self-adaptation combined with collaboration between hardware and software to cope with the constraints raised by future applications and systems.