Also visit my website Haarausfall Stoppen

   0 likes

Take a look at the Cisco Validated Design (CVD) from the link in the post. It will answer a lot of your questions.

To address your specific questions, we booted 5000 Windows 7 SP1 HVDs from shut down to ready to log in 40 minutes.

The Login VSI workload test logged in 5000 users who all started using eight applications immediately after login with a random delay of application startup after login of up to 15 seconds in 30 minutes.

Every user is working productively on all of their applications concurrently within that 30 minute period.

See page 187 in the CVD for a detailed description of the user workload simulation.

The VNX7500 configuration is detailed in Chapter 6 on page 101.

All of the HVDs ran on a single dedicated VLAN. See Chapter 6, page 49 for a list of all VLANs configured in the environment. We utilized our Cisco UCS Manager, Cisco Nexus 5548UPs, and Cisco Nexus 1000Vs to provide VM to L3 QoS priority for HVD traffic.

Thanks for your comments!

   0 likes

We did use Citrix User Profile Manager to manage the end user data. We recognize that AppSense and others offer expanded capabilities in this area. You are correct that we did not use AV protection in this Cisco Validated Design. We did not utilize published or streamed apps into the virtual desktops, generally considered double-hop scenarios.
Our objective in the Performance Solutions Team, is to demonstrate real world solutions at large scale. Our sister group, the Cisco VXI team, focus on integration of voice, AV, EUM, collaboration, and end points into their testing. Here is a link to their latest publication:

http://www.cisco.com/en/US/docs/solutions/Enterprise/Data_Center/VXI/configuration/VXI_Config_Guide.pdf

They will publish an update in early calendar Q2 2013.

Stay tuned for the next phase of our scale testing that will feature XenDesktop 5.6 HVDs with PVS write cache on local SSDs, XenDesktop 5.6 with Personal vDisk, and virtual XenApp 6.5 at a combined scale of 4000 seats.

Thanks for your comments!

   0 likes

I agree,in a small part, with what David says above however; I feel solutions already exist to resolve most of the concerns his comment implies. For instance, there are agent-less security solutions such as Trend Micro’s DSVA, as well as VMware vShield which will resolve most of those concerns.

The question I might have is what was the average boot times for end users and what impacts could users expect to experience as they ramp up to steady-state? Perhaps their session boots quickly, but can they open Outlook, a GIS solution and a complex linked spreadsheet without too much impact? With say 50 people doing such in a couple separate VLANs; who all is impacted? Or maybe there are no impacts?

I’m familiar with the VNX 5300 however; I’m curious how you designed the storage itself. Surely you are using an auto-tiered solution, but how was the storage provisioned? Thin/Thin Thick/Thin?

Again, thanks for the great post. And if you don’t mind answering a few questions, then thanks again.

   0 likes

I do like the fact that it does include test configurations and key metrics for post deployment testing which could aid service and support teams across any design.

   0 likes

   1 like

   0 likes

Best,

ET

   0 likes