How Updated Firmware Makes a Difference
by Team AppNeta on

Virtual Desktop Infrastructure seems to be popping up everywhere these days. I remember the days when RDP was used solely to save time from having to go down to the server room. Now it’s blossomed into a multibillion dollar industry with real businesses relying on it. Microsoft, VMware and Citrix, among others, are trying to put stakes in the VDI landscape. Even with all the players trying to get in on the game, there is one static among them all – the network.

I was giving a trial recently to some folks who wanted to understand why mouse/keyboard freezes were existing in their VMware View deployment. To make things simple for the deployment, there were two distribution switches on either side of a router. Plugged into the first switch were the image servers, the other housed the connections for the thin clients. All around, everything was set to Gigabit. Upon first inspection, this seemed simple and straight forward. There were only three network devices, a thin client and a server – not many areas where the protocols could get hung up.

Of course, nothing is that easy.

Using the PathView component of PathView Cloud, we were able to monitor the paths to the thin client as well as to the servers. While we initially did not see much activity, things got interesting when utilization spiked. In our monitoring, we saw utilization grow up to 95% to the image server and then go offline. The disconnect lasted approximately 90 seconds, but once that time passed everything was fine.

Firmware 1

Turns out during those 90 seconds the user was resetting the thin client – because of mouse/keyboard freezes – thus terminating all connections and starting back up again. You can see on these screen shots that the disconnect cleared the error, albeit at the end user’s expense.

As is often the case, network devices were not the problem in this instance. Instead, the problem stemmed from a firmware issue. Under high load, the thin client would have disconnects occur from the image – complete outages. The firmware on the thin client was a couple versions out of date.

Firmware 3

Once the reason for the problems was determined, we turned to curing the issue. We confirmed with the vendor that an upgrade was necessary and implemented the necessary changes. We then used PathView again to confirm that the issue was in fact cured.

In conclusion, the network is a critical piece to any application’s performance. However, we can never overlook the basic troubleshooting phrase – is your firmware up to date?

Filed Under: Performance Monitoring

Tags: firmware , PathView Cloud , VM technology