Locating Distributed Information

As part of research into transparent execution of computational services in a distributed environment, two problems we have encountered are how to manage and locate information about service availability and the status of server machines. The key observation about these two problems is that they are not isolated problems, but each is an instance from a class of distributed information problems. The class of problems is characterized by the inherent distributed nature of the information where each server machine supplies a portion of the total information. In this paper we first examlne general techniques for managing distributed information and how various evaluation criteria, when applied to the techniques, are affected by the parameters and scale of a specific problem. We then illustrate the use of the techniques and criteria for the problems of locating service and server status information, as well as determining the current status of a user in a distributed environment. In each case we find the best technique to use for the problem based on its specific parameters. '"This work was supported in part by grants from the National Science Foundation (MCS8219178), Sun Mierosystems Incorporated, and Digital Equipment Corporation.