Distributed computation for design aids

This paper is directed at users who are contemplating a move to distributed computing or are already there. It is a compendium of data and analytic models which has been assembled from experience on computers at Bell Laboratories and at Rutgers and Stanford Universities. It will be presented as a series of observations followed by either data or analysis to support them. The object of this paper is to help the reader to avoid some of the obvious pitfalls in moving to such a facility, and to provide some rational guidelines for planning. We will quantify the key parameters that characterize a distributed computing environment, and suggest how those parameters can best be used to take advantage of an existing environment, or to design an improved environment. Finally, we will relate these observations to the introduction of 'personal computing' into the computing milieu.