VirtualWisdom SAN Performance Probe

Measure and Analyze Your SAN

VirtualWisdom SAN Performance Probe Fibre Channel probes analyze every packet on a Fibre Channel SAN. The SAN Performance Probe receives its data from the SANInsight TAPs and views a copy of all frames, extracts the frame header and ignores the data payload. Through the frame header information, the SAN Performance Probe can ascertain all critical information about the transaction in real-time. A SAN Performance Probe can be dedicated to a single SAN link or, using a Virtual Instruments SANInsight Rover physical layer switch, can rove (sample) among many SAN links to provide dozens of critical statistics about storage traffic, transmission errors, and SAN READ and WRITE latency. The SAN Performance Probe is the only product in the industry that can measure the impact of the SAN on application response times.

The Virtual Instruments VirtualWisdom SAN Performance Probe (Model ProbeFC8 or Model ProbeFC8-HD) supports Fibre Channel SANs at speeds for up to 8 Gbps. A single SAN Performance Probe can be dedicated to up to 16 SAN links, or, using a VirtualWisdom Rover, can rove among many more. The solution provides detailed measurement and statistics about storage I/O traffic as it relates to an application server, storage subsystem, remote replication, virtual or other key SAN device, or application. The SAN Performance Probe detects application performance slowdowns by measuring every SCSI I/O transaction from start to finish, for both reads and writes. Exchange Completion Time (ECT) metrics are maintained for every server/volume combination (initiator/target/LUN).

The SAN Performance Probe can measure latency through wide-area networks and from servers to storage subsystems. A baseline of metrics can be created to verify that there is no impact after server OS or device firmware upgrades, zone modifications, or other changes to the SAN.

Congested storage ports can be identified from the SAN Performance Probe queue depth metrics. VirtualWisdom measures queue depth by monitoring the number of active and pending SCSI exchanges to the same target port. Overrunning storage queues can impact the performance of the entire SAN. Failing SCSI devices can be proactively detected by monitoring link errors and SCSI operation errors using the SAN Performance Probe.

The VirtualWisdom SAN Performance Probe can simultaneously monitor up to 16 fibre channel links. Probes contain algorithms that correlate end-to-end storage traffic between a server and a storage volume. Metrics are calculated within the probes for every active Initiator/Target/LUN (ITL) nexus, including SCSI frame analysis, elapsed time from SCSI read to first data, time elapsed for each complete read or write exchange, and the number of simultaneous open exchanges. Probes also detect Fibre Channel faults, SCSI errors, and device faults. An Ethernet connection allows the probes to communicate with the VirtualWisdom Server. A summary of changed metrics is reported to the VirtualWisdom Server every second.