Get an inside look into the real world experiences and best practices of Intels IT group.
Intel IT studied the impacts of streamed client and Virtual Hosted Desktop computing models on server and network utilization. This is part of an ongoing evaluation of emerging computing models for enterprise operations. This white paper discusses the findings of the study which include: Network utilization was consistently low for Virtual Hosted Desktop; Server Utlization was significantly lower for streaming than Virtual Hosted Desktop (VHD); and streaming was more efficient for a wider variety of applications including graphics, multimedia, flash animation and real-time collaboration.
Tom,
Unfortunately, due to our license agreement with VmWare, we are unable to comment on any benchmarking that we may have performed internally with their products.
However, we are comfortable that generalizations can be made, especially with respect to the issues with RDP. The screen saver, for example, causing high server utilization is not an issue with virtualization. It's about how the user interface is distributed to the client.
I read your report and have a question re: the virtual desktop compared to streaming applications.
In your tests you found that virtual desktopa ate up more server CPU power than streaming, and that some desktops were disabled by a rogue screen saver. You also found that some conventional desktop functionality did not work with virtual desktops.
My question concerns the fact that all of virtual desktop testing was performed using Microsoft Virtual Server 2005 R2 SP1.
Did you consider how the test might have come out if you used VMware, the more mature and industry leading virtualization software? I just wonder if we can generalize from your test to all virtual desktop implementations, or just those that use Microsoft's technology?
Tom Farre
Freelance journalist