A Framework for Adaptive Software-Based Reliability in COTS Many-Core Processors

Conference: ARCS 2015 - 28th International Conference on Architecture of Computing Systems
03/24/2015 - 03/27/2015 at Porto, Portugal

Proceedings: ARCS 2015

Pages: 4Language: englishTyp: PDF

Personal VDE Members are entitled to a 10% discount on this title

Authors:
Alhakeem, Mohammad Shadi; Munk, Peter; Lisicki, Raphael; Parzyjegla, Helge (Technische Universität Berlin, Germany)
Parzyjegla, Helge; Muehl, Gero (University of Rostock, Germany)

Abstract:
Commercially available many-core processors offer the performance needed for computational-intensive safetycritical embedded applications, but are increasingly susceptible to soft errors because of their highly integrated design. While hardware-based solutions are costly, implementing softwarebased safety mechanisms is inexpensive, but complex and challenging due to the tight coupling between software and hardware, and due to the recent trend of mixed-criticality systems. To address these challenges, we present a framework enabling developers to specify safety requirements separately from functional aspects. The framework then automatically selects an appropriate safety mechanism and adapts it to the underlying system in order to achieve the specified level of reliability. We describe our framework and show its functionality based on an adaptive N-Modular Redundancy (NMR) mechanism.