Who doesn’t like to run a system in tip-top shape? And who doesn’t like to maximize user density and shave 5-10 seconds off a login time? Luckily our Citrix Consulting team specializes in this sort of thing, and we’ve got a few tricks today to optimize performance of your Citrix Workspace.
CPU Over-Subscription Ratio. We walk into a lot of XenApp environments and the CPU Over-Subscription ratio is all over the map. If you want to maximize user density or single server scalability (SSS) on the boxes hosting XenApp workloads, then I highly recommend a range of 1.5x to 2x. In addition to squeezing out a few more users on each box, this over-subscription range has also been proven to provide a great user experience (UX) for end-users. And as much as I love maximizing SSS, we can’t forget about UX either. If you’re curious how many VMs or users to expect on each box, then my “Rule of 5 and 10” is a great resource to quickly estimate density.
Graphics Codec. Many of you noticed that when we moved to the 7.x architecture, the “video codec” (or H.264) was the default graphics codec or engine used. And while that provides a great user experience due to the high frame rate, H.264 also consumes a lot of CPU. So, we rewrote Thinwire for modern operating systems (you might hear this referred to as “Thinwire Plus” or TW+) and kept improving it over the last 10 or so releases… and now TW+ is the default graphics codec used in the latest builds. The result is still a great user experience, but much better SSS at the same time. How much better? We typically see about a 10% improvement in scalability when toggling H.264 and Thinwire+.
WEM. Courtesy of the Norskale acquisition, this smart agent is quickly becoming one of our best friends. Why? It can shave valuable seconds off those long login times that might be frustrating your end-users. Your mileage will vary, but the more GPOs, GPPs, login scripts, drive and printer mappings you have, the more we can shave off. The trick is extracting these items into WEM and processing them in parallel (or after the shell loads), as opposed to processing them sequentially and adding to the login time. We just did one project for a customer and shaved off about 10 seconds from their login time. So, if you haven’t checked out WEM yet, it might be time to do so.
Those are just a few tips to optimize performance at a high level. Want to know a few more tricks or want some more details on these? Greg Davis and I hosted an action-packed 30 minute webinar last week, which you can watch now, on-deman! Be sure to check it out; after all, who doesn’t love a little free Consulting? 😉
Please feel free to drop us a note in the comments section if there are additional topics you would like to see covered. We look forward to seeing you there!
Nicholas Rintalan, Principal Architect & Sr. Director of Enterprise Architecture, CCS