Election algorithms arewidely used in the primary-backup replication, and ithas been proved to be harder than the con-
sensus problem. In this paper, the relation between the election problem and the consensus problem is analysed, and a new fault-
tolerant election algorithm is given. This algorithm is stable, communication-efficient, and itcan tolerate the temporal failure on the
processors and links. The algorithms proposed in this papernotonly solve the election problem, butalso provides a newway to solve
the consensus problem
