This paper presents a sequential traversal of a network in an environment with crash-failures. This communication primitive is basic and may be used to solve easily many problems like identification, termination detection, and agreement, etc. Contrary to a possible intuition, this primitive is non-trivial since crash-failures involve tricky problems. Moreover, the implementation proposed in this paper is optimistic in the sense that, in the absence of fault, it is as efficient as the underlying non-fault-tolerant one.
Adoption protocols for fanout-optimal fault-tolerant termination detection
