Desktop Virtualization: Using and Abusing the WAN

Virtualization Reality

Desktop Virtualization: Using and Abusing the WAN

By TMCnet Special Guest
Alan Murphy
  |  May 01, 2011

This article originally appeared in the May 2011 issue of INTERNET TELEPHONY.


While most IT business initiatives these days are focused on on- or off-premises cloud deployments, a smaller subset of IT is looking at virtualization inside the data center as a streamline management and security tool: the desktop team. VDI – virtual desktop infrastructure – allows IT to virtualize end user desktops and move them from the cube or office and into the data center. This move centralizes desktop resources for more efficient support and, more importantly, for greater control. With VDI, IT has direct access to all user desktops at once and on their schedule.

Just like choosing a virtual platform environment, there are many VDI options available to fit the different needs of individual IT departments. The decision on which VDI solution to go with is typically a business decision – which virtual platform an enterprise is already using, which platform offers the best value for the deployment size, etc. – or a feature decision – which platform will provide the features needed by the enterprise and work in an existing deployment architecture.

The underlying protocols used by VDI solutions are very rarely a decision factor; however, they can play a large role in the success of VDI deployments. All three major VDI players lead with different primary transport protocols (although all support RDP as well): VMware with PCoIP, Microsoft (News - Alert) with RDP, and Citrix with ICA.

Regardless of which platform or protocol is selected, one common issue that many enterprise organizations run into during initial testing is a noticeable slowness of the desktop. More often than not this is traced back to a networking issue. Virtualized desktops can put a tremendous load on the network, especially if the desktop is in one location, the user is in another, and access is over the WAN. Two critical factors for deploying VDI over the WAN are latency and bandwidth.

VDI deployments are extremely sensitive to latency – the amount of time it takes to transfer data over the network connection. A better way to state this is that humans are extremely sensitive to latency. As modern desktop users we’ve become accustomed to watching the spinning hourglass or pinwheel (depending on your preferred desktop flavor) while applications are processing on our local desktop; we simply wait for the application to finish so we can go about our business. With a remote desktop, however, the latent elements are doubled: We’re still waiting for desktop applications to run on the remote server, and now we’re waiting on the desktop data to be delivered to our local thin client. And that delivery is where latency typically rears its ugly head. If there are any network issues between the user and the server, or the user is connecting to a remote desktop over a long physical WAN connection such as from Boston to Sydney, the user will notice the desktop lag immediately, and it will impact everything the user trying to do on his or her desktop.

Bandwidth is the other issue that users have to deal with: More VDI instances over one network means more bandwidth, and more bandwidth means more competition. A very common VDI deployment scenario is the branch office; a small cluster of employees will be located in one location, sharing one LAN and one WAN connection back to the corporate office. In a perfect VDI world, the virtualization servers would sit as close to the user as possible; in this case they would be physically located in the branch office, feeding VDI instances locally to users.

But that’s not always realistic. Sometimes there aren’t enough available resources at the branch to host physical servers (imagine a construction trailer at a worksite, for instance), or the enterprise network may not support extending the management domain to the branch. Ironically, one of the reasons that VDI may be deployed is because the enterprise can’t manage physical desktops on-site.

The solution suggested by all three major VDI vendors is to deploy a specialized WAN optimization solution along with virtualized remote desktops. These WAN optimizers typically work by optimizing both the application data – in this case RDP, PCoIP, or ICA protocol data – which typically helps with bandwidth, as well as the WAN connection between data centers or to the branch office, a solution to address latency. Although each VDI vendor has a WAN optimizing technology they prefer, there is no question that any VDI deployment will benefit from reduced latency and more intelligent bandwidth management. Moving those tasks to dedicated hardware or a virtual machine that’s part of the virtual infrastructure takes the burden off of VDI and lets it do what it was designed to do: deliver desktops to users without impacting their productivity while increasing IT efficiency. Much like load balancing has become a standard tool when deploying any new web app, WAN optimization should be a natural extension of any VDI deployment planning.

Alan Murphy is technical marketing manager of management and virtualization solutions with F5 Networks (News - Alert) (www.f5.com).


TMCnet publishes expert commentary on various telecommunications, IT, call center, CRM and other technology-related topics. Are you an expert in one of these fields, and interested in having your perspective published on a site that gets several million unique visitors each month? Get in touch.

Edited by Stefania Viscusi