Long Atomic Computations

Abstract

Distributed computing systems are becoming commonplace and offer interesting opportunities for new applications. In a practical system, the problems of synchronizing concurrent computations and recovering from failures must be dealt with effectively.