Nathanael Iversen, Director of Technical Marketing, Xangati
As we’re talking about end user experience, this fourth tip is going to get directly at what’s necessary to keep track of end user performance. There’s no better way to track that than to record exactly what’s happening for a particular user at the exact time that they are having a problem. This is so important because it’s ultimately the end users who determine whether or not a VDI project will succeed. Because the users are now depending on this shared infrastructure to do their job, if there’s any problem with it, they are going to be very quick to complain about slowness, delays, etc. and any problem at all will get blamed on VDI. In order to move forward with a strong push from the user community, you have to take this into account early in the deployment. This is particularly problematic for IT because with end user issues, you basically can’t recreate them…an end user connects to a desktop in the pool but, the next time they connect, it might very well be a different desktop. And, when the reports of performance issues trickle in – maybe a day to a week after the incident, it becomes almost impossible to piece back together again what was the dynamic infrastructure used to support that user in a given point in time.
Xangati changes that model. We have something called a Visual Trouble Ticket that allows an end user to file their own infrastructure report. They put in their name, phone number, and a comment (ie. internet is slow, mail is broken, I can’t access my application, etc.) and then click the “create visual trouble ticket” button and then a 15-minute recording is made of everything that user is doing: on the network, what their connectivity to storage is, which datastores, which desktop, what host it’s running on – all of that information is collected and recorded with true to second accuracy so that during that time, you can simply grab a scroll bar and move backwards and forwards in the recording and isolate exactly the pieces of information you need to figure out what happened at that time. This is all done without any agents on the desktop and without any probes inserted in the network.
Check back tomorrow for the 5th Tip in our series where we look at establishing a cross-silo operational model to proactively manage performance.