Cisco Blogs

Cisco and Citrix Deliver a Fresh Take On Desktop Virtualization

September 8, 2010 - 3 Comments

In talking to our customers, one of the things that I hear consistently is that customers appreciate our open solution stack.  What does that mean?  Well, a couple of things–first of all, it means that our innovations are individually available–you can connect a UCS to any upstream Ethernet and Fibre Channel switch, you can connect the server of your choice to a Nexus switch and still take advantage of OTV, FabricPath or unified fabric and you can run the Nexus 1000V in pretty much any hardware environment supported by vSphere.


The other thing this means is that we build solutions stacks with insertion points for our partners so our customers can take advantage our our solutions while still working with their favorite data center vendors.  Both Vblocks (Cisco, EMC, VMware) and SMT (Cisco, NetApp, VMWare) were initial examples of this approach.


Today, we are adding to the list with a newly announced desktop virtualization solution developed in partnership with Citrix.   Desktop virtualization is taking off with our customer for a number of reasons from reducing costs, to faster app deployment, to improving regulatory compliance and we have been listening to what our customers have been asking for.


The new solution with Citrix XenDesktop combines Cisco UCS and Citrix desktop virtualization technologies including FlexCast™ and HDX™, to deliver a cost-effective, scalable and high-performance solution for hosting, securing and optimizing the delivery of virtual desktops and applications.  


In this video Ben Gibson of Cisco and Gordon Payne of Citrix discuss the announcement:


Of course, we are not going to roll out a me-too solution….  


Our initial testing shows that our Citrix VDI solution can save at least 20% over competing solutions.  Much of this cost advantage can be linked to the fact that we can deliver 60% greater virtual desktop density with our extended memory technology without a loss of performance.  Beyond the infrastructure costs noted below, other Cisco advantages like UCS Manager service profiles also simplify and accelerate deployment and management and contribute to reduction of overall TCO.


1,500 Seats Cisco Vendor A Vendor B
Server Networking and Infrastructure $48,129 $64,437 $93,128
Compute – Management Nodes $13,621 $19,883 $20,157
Compute – Desktop Nodes $247,708 $289,325 $264,102
TOTAL $309,458 $373,645 $377,386
Cost per Seat for 1,500 Seats $206 $249 $252
Cisco Savings   21% 22%
5,000 Seats Cisco Vendor A Vendor B
Server Networking and Infrastructure $102,425 $128,873 $186,256
Compute – Management Nodes $27,242 $39,766 $40,313
Compute – Desktop Nodes $825,692 $964,417 $880,339
TOTAL $955,359 $1,133,056 $1,106,908
Cost per Seat for 5,000 Seats $191 $227 $221
Cisco Savings   19% 16%

This analysis is based on a comparison of like-to-like configurations with other compute manufacturers and assumes equivalent configurations result in the same number of seats.

The analysis is based upon ASPs as of 8/30/2010. Server Networking and Infrastructure includes chassis and server networking infrastructure necessary for a working system.

Management nodes are comprised of 48GB servers, X5670 CPUs and converged network adapters. Desktop nodes are comprise of 192GB servers, X5670 CPUs and converged network adapters. 


The solution is validated with Citrix XenServer and VMware vSphere™ today–you can expect support for Microsoft Hyper-V soon. For more details on the solution, including a validated design guide, check out this link where you will find a whole host of information. You should also check out this post by Raj Dhingra of Citrix.


To close, expect to see more of these kinds of solution stacks from Cisco and our ecosystem partners–for example, you can expect to see similar solutions in the near future with other members of our ecosystem.  From a broader perspective, you can also expect to see more in from us in terms of more tightly aligning IT innovation and investments with specific business results.  For more on this, check out this event.



In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.


  1. Thanks Omar. Very cool and affordable, compared to the competitors. How do they plan on competing with the open source and freeware market though?

  2. Ramesh: Ramesh:The differences stem from a couple of reasons. First, because of inherent differences in thinks like port counts, supported density, etc vendor to adds incremental networking equipment at different points. Since we were focused on number of Virtual Desktop seats, we allowed round numbers to drive the configurations. This was more equitable to each vendor than picking arbitrary numbers of servers based on what fit best into a Cisco configuration. As a result, the 1,500 seat configuration was slightly more efficient to the Cisco architecture than the 5,000 seat configuration. Second, the actual dollar savings on networking and infrastructure was greater at 5,000 seats, but the savings were a smaller percentage of the overall savings, so that the % impact was slightly less with that configuration.Omar

  3. Why is it that a 5000-seat solution delivers less [%] savings than the 1500-seat solution? Can the author, or someone else, explain? Thanks.