Fault Detection and Co-design Recovery for Complex Task within IoT Systems
Radia Bendimerad, Kamel Smiri, Abderrazek Jemai
Abstract
Internet of things (IoT) allows the implementation of embedded devices that guarantee multiple application processing. A device contains a processor core (CPU) but also hardware accelerator (FPGA) to permit a hardware-software (HW-SW) partitioning depending on its complexity. However, these devices have limited capacities, so they are exposed to faults. Thus, the system needs to be adapted in such a way to detect device faults, continues to function normally and perform tasks to produce correct output. In this paper, a holistic approach dealing with fault detection and recovery for complex tasks within the IoT was outlined. It offers a technique to diagnose the state of processing elements. Then it combines task scheduling in HW and SW parts (Co-design) in such a manner to ease the process of recovery when a system fault is detected. The proposed work ensures a better performance for the entire system. An experimental study validates the effectiveness of the present strategy without impacting system performances thanks to the contributions defined in this paper.
Download