The author analyzes client/server outage data, showing that user availability provides a meaningful way to define client/server availability. The data was measured in terms of annual user outage minutes, based on the concept of user downtime. Design engineers can use this concept to estimate the savings from client/server architectural and operational improvements. Based on client/server outage data, the author outlines the most important outage causes in a client/server environment and evaluates methods for improving client/server availability. Causes include physical faults of hardware components, design errors in hardware and software, operations errors caused by operations personnel or users, and environmental failures (such as failure of external connections, power, or cooling-system failures), and outages due to planned reconfigurations. The study spans all client, server, and network devices in a typical client/server environment. This analysis also addresses basic issues such as the definition of system failure and the appropriate measure of availability that need to be reconsidered in the client/server availability paradigm. The author presents some interesting results. Server outages were the most significant cause of client/server unavailability. Software caused the most user outage minutes, but all outage categories were important. Finally, fault tolerance, applied throughout the client/server architecture, can significantly reduce user outage time.
