A Survey of Lifetime Reliability-Aware System-Level Design Techniques for Embedded Multiprocessor Systems
Abstract
A Survey of Lifetime Reliability-Aware System-Level Design Techniques for Embedded Multiprocessor Systems is emerging as a major concern for system design as escalating power density and hence temperature variation continues to accelerate wear-out, leading to a growing prominence of device-defects. This has attracted a significant attention both in industry and in academia to investigate on wear-out mitigation techniques, from micro-architectural adaptations to systemlevel optimization. Task mapping and scheduling-based system-level design techniques provide a low overhead approach for reliability optimization at design-time as well as reliability management at run-time. This Lifetime Reliability Aware System Embedded Systems paper provides an overview of the developments on reliability optimization using task mapping and scheduling over the past decade, since the introduction of the first reliability-aware system-level design technique in 2004.
This A Survey of Lifetime Reliability-Aware System-Level Design Techniques for Embedded Multiprocessor Systems paper presents an automatic reliability-aware system-level design methodology to tolerate hardware defects caused by manufacturing tolerances as well as destructive agents and aging processes at the places of activity of the system components.
System Configuration
Speed : 1.1 GHz
Platform : MySql,Embedded system
Conclusion
The work at hand proposes a system-level design methodology for a reliability-aware optimization of embedded systems. A formal analysis based on Binary Decision Diagrams (BDDs) [4] is proposed to determine the reliability of a system implementation. This analysis technique is capable of automatically taking arbitrary system structures into account and deriving the overall lifetime reliability of the system based on the reliability functions of the system components.A technique to reuse existing resources for the multiple instantiation of software tasks called multiple binding is proposed that avoids costly resource replications.
