System Adaptivity and Fault-Tolerance in NoC-based MPSoCs: The MADNESS Project Approach

Abstract

Modern embedded systems increasingly require adaptive run-time management. The system may adapt the mapping of the applications in order to accommodate the current workload conditions, to balance load for efficient resource utilization, to meet quality of service agreements, to avoid thermal hot-spots and to reduce power consumption. As the possibility of experiencing run-time faults becomes increasingly relevant with deep-sub-micron technology nodes, in the scope of the MADNESS project, we focus particularly on the problem of graceful degradation by dynamic remapping in presence of runtime faults.In this paper, we summarize the major results achieved in the MADNESS project until now regarding the system adaptivity and fault tolerant processing. We report the first results of the integration between platform level and middleware level support for adaptivity and fault tolerance. A case study demonstrates the survival ability of the system via a low-overhead process migration mechanism and a near-optimal online remapping heuristic.

*‐Predictable MPSoC execution of real‐time control applications using invasive computing

“…Adaptive fault tolerance where defective resources are replaced by functional ones has also been a focus of the MPSoC community. The MADNESS project 12 models processes with polyhedral process networks and provides a low‐overhead mechanism to migrate a process to a different tile in case a fault is detected. Migration techniques are also the focus in the work of Chou and Marculescu, 13 where Chou and Marculescu provide a theoretical framework to determine the impact of spare core placement on fault tolerance properties.…”

Uncertainty aware mapping of embedded systems for reliability, performance, and energy

“…[7][8]9 These tools facilitate flexible system-level performance evaluation by providing support for mapping a behavioral application specification to an architecture specification. 10,11 Also, reliability has become a primary design concern alongside traditional design objectives. 12,13 However, the majority of the previous work did not consider uncertainty or reliability in the design process of embedded systems.…”

A Low Overhead Self-adaptation Technique for KPN Applications on NoC-based MPSoCs

“…In terms of their classification criteria, our work can be classified as quality management since we address application-level performance goals. It can be incorporated into the generic selfadaptive run-time environments proposed in (Nollet et al, 2010;Derin et al, 2009) and reside alongside other self-adaptation services such as fault-tolerance (Meloni et al, 2012).…”

“…The platform proposed in (Meloni et al, 2012) is adopted as the hardware base for this work. In the proposed approach the system architecture can be seen as a network of tiles, interconnected by means of a NoC communication infrastructure.…”