Model framework to facilitate robust programming of distributed workflows

摘要

A method is disclosed. A finite state machine model for a single system workflow is replicated across a plurality of distributed nodes associated with a leader election protocol. A leader is determined amongst the plurality of distributed nodes to perform a next action of the finite state machine model based at least in part on the leader election protocol. One or more nodes amongst the plurality of distributed nodes are configured to submit a timeout ticket if the next action of the finite state machine model does not appear to have been performed by the leader within a prescribed time.