Safe termination detection in an asynchronous distributed system when processes may crash and recover

The termination detection problem involves detecting whether an ongoing distributed computation has ceased all its activities. We investigate the termination detection problem in an asynchronous distributed system under the crash-recovery model. It has been shown that the problem is impossible to solve under the crash-recovery model in general. We identify two conditions under which the termination detection problem can be solved in a safe manner. We also propose algorithms to detect termination under the conditions identified.