How OnCommand Insight Ethernet monitoring works

OnCommand Insight uses an Ethernet monitoring unit (EMU) to inspect Ethernet packet headers on a switch adjacent to storage, and determines the type of NFSv3 operation being performed, the destination IP address of the operation, and the volumes or shares being addressed.

The EMU continually monitors the mirrored traffic and builds top-k list of the "k" most heavily loaded volumes IDs and share names, sorted by total observed operation count. For each volume and share, the EMU generates a top-n list of the top "N" client IP addresses contributing to the load on that specific volume or share, sorted by total observed operation count. Top-k and top-n lists are maintained independently for each protocol. Monitored traffic is only examined to update internal counters relevant to per-volume and per-client operation accounting. The monitored traffic is not persistently stored.

Ethernet traffic monitoring is supported on hardware platforms running the following software:

Collected data is sent to the corresponding Insight server associated with the EMU. Since the Insight server is also acquiring the relevant storage details, for the same volumes and shares and potentially the host details, the server can merge the EMU data into the standard Insight data model, correlating Volume IDs and share names to discovered Internal Volumes and client IPs to resolved physical hosts or discovered Virtual Machines.

The information is available in OnCommand Insight, providing the following:

The information allows customers to pro-actively monitor NAS traffic and easily identify top contributing clients that are putting the heaviest loads on specific volumes or shares.