Improving reliability through fault propagation scope in embedded systems

Abstract

In this work, we propose a technique based on task recomputation in order to improve reliability without incurring any performance degradation in embedded systems. The technique focuses on recomputing the task using the slack available on idle processors thereby maximizing the usage of the processing elements. In conjunction with task recomputation, we employed two metrics called as Fault Propagation Scope (FPS) and Degree of Criticality (DoC). Our technique, named as Hybrid Recomputation, improves the reliability even as the scope for fault propagation is reduced. The fault propagation scope of the task graph is reduced by incorporating the fault propagation scope of each task and its degree of criticality into the scheduling algorithm. Our technique of hybrid recomputation is analyzed using various automatically generated task graphs using TGFF. The experimental results clearly indicate the viability of the proposed approach under different latency constraints.