Incorporation of optimal timeouts into distributed real-time load sharing

The authors address the problem of designing and incorporating a timeout mechanism into load sharing (LS) with state-region change broadcasts in the presence of node failures in a distributed real-time system. They formulate the problem of determining the best timeout period T_out^<i> for node i as a hypothesis testing problem, and maximize the probability of detecting node failures subject to a prespecified probability of falsely diagnosing a healthy node as faulty. They outline the LS algorithm and the proposed timeout mechanism and establish a theoretical basis for the calculation of optimal T_out^<i>. The simulation results show that the LS algorithm, which combines online parameter estimation, the timeout mechanism, and a few extra, timely broadcasts, can significantly reduce the probability of missing task deadlines