Abstract
Abstract: This paper examines client/server outage data and presents a list of outage causes extracted from the data. The outage causes include hardware, software, operations, and environmental failures, as well as outages due to planned reconfigurations. The study spans all client, server, and network devices in a typical client/server environment. The outage data is used to predict availability in a typical client/server environment and to evaluate various fault-tolerant architectures. The results are stated in terms of user outage minutes and have been validated by comparison with other outage surveys and data in the literature. The major results from the outage data study are: Each client in a client/server environment experiences an average of 12000 minutes (200 hours) annual downtime; Server outages, especially server software failures, are the dominant cause of client/server unavailability; Fault-tolerance in a client/server environment can reduce user outage minutes by nearly an order of magnitude.