Conventional computation models restrict to particular data structures to represent states of a computation, e.g. natural numbers, sequences, stacks, etc. Gurevich's Abstract State Machines (ASMs) take a more liberal position: any first-order structure may serve as a state. In [7] Gurevich characterizes the expressive power of sequential ASMs: he defines the class of sequential algorithms by means of only a few, amazingly general requirements and proves this class to be equivalent to sequential ASMs. In this paper we generalize Gurevich's result to distributed algorithms: we define a class of distributed algorithms by likewise general requirements and show that this class is covered by a distributed computation model based on sequential ASMs.
展开▼
