Anyone who administers or maintains a critical Enterprise Storage Area Network (SAN) has most likely experienced the frustration of what happens when problems occur on their network. With a critical SAN, usually supporting a large number of users, downtime is expensive in terms of both lost productivity and data access. Adding to the frustration, the techniques used to solve these problems often require stopping the entire SAN, including unaffected areas, to install monitoring equipment or analysis equipment. This paper describes a flexible, centralized, and cost-effective method to instrument critical SANs for analysis and monitoring that not only creates a superior troubleshooting environment, but one that allows administrators to track down problems while the network is up and running. And, one that guarantees not to introduce prospective points of failure during normal operation, even in the event of power loss.
