Aguilera, Marcos Kawazoe; Chen, Wei; Toueg, Sam
We consider partitionable networks with process crashes and lossy links, and focus on the problems of reliable communication and consensus for such networks. For both problems we seek algorithms that are quiescent, i.e., algorithms that eventually stop sending messages. We first tackle the problem of reliable communication for partitionable networks by extending the results of [ACT97a]. In particular, we generalize the specification of the heartbeat failure detector HB, show how to implement it, and show how to use it to achieve quiescent reliable communication. We then turn our attention to the problem of consensus for partitionable networks. We first show that, even though this problem can be solved using a natural extension of less than or greater than S, such solutions are not quiescent --- in other words, less than or greater than S alone is not sufficient to achieve quiescent consensus in partitionable networks. We then solve this problem using less than or greater than S and the quiescent reliable communication primitives that we developed in the first part of the paper. Our model of failure detectors for partitionable networks, a natural extension of the model in [CT96], is also a contribution of this paper.
computer science; technical report
Previously Published As