In today's distributed computing environments, users are makingincreasing demands on the systems, networks, and applications they use. Users are coming to expect performance, fault tolerance, conguration, security, and accounting services comparable to those on traditional centralized systems. Meeting these expectations implies the need for management. While some successes have been realized in the areas of system and network management,only preliminaryinvestigationshave been conducted on management of distributed applications. At this stage, little is known of the costs associated with this type of management. To assess these costs, it is necessary to collect data about the performance and operation of distributed applications and their management systems. Furthermore, it is important to collect this data under varying constraints, congurations, and managementassumptions. In order to do this, we have developed a distributed applications management testbed. We present our initi...
Michael Katchabaw, Stephen L. Howard, Andrew D. Ma