Abstract:
This paper proposes a faulty node detection scheme that performs collective monitoring of a distributed system. The algorithm is designed to observe every node's behavior collectively based on the pseudo-random Bose-Chaudhuri-Hocquenghem (BCH) code. Each node sends a single BCH bit simultaneously as a replacement for the cyclic redundancy check (CRC) code. This method can be used for detecting and preventing serious damage caused by node failure. Simulation results show that the high-polynomial CRC retains its performance level, although its code is modified for the distributed BCH code.

Available online: https://ieeexplore.ieee.org/document/6459486