Implementing Causal Ordering with Bounded Time stamps

This paper investigates a solution to the problem of causal ordering in message passing distributed systems Causal ordering is the restriction that messages are delivered in a fo fashion with respect to the global causal order between events in the system This is stronger than the condition that each local channel is fo In our algorithm causal ordering is implemented by having each process main tain as the computation proceeds its latest information about every other process in the system To achieve this messages are tagged with time stamps The novel feature of the protocol described here is that it allows for the reuse of time stamps Under certain conditions this permits an implementation of causal ordering using time stamps that are uniformly bounded