Distributed Computation of a Spanning Tree

Abstract

A typical architecture that one encounters is that of a number of computers located in the nodes of a sparsely connected, but otherwise arbitrary network. One of the difficulties with such an architecture is that the flow of control and the information exchange between the various computers are not always easily orchestrated, simply because the network is so arbitrary. This brings about the so-called routing problem. There are, however, at least two general ways to reduce these difficulties.