Some time ago I posted a first look post on XenClient. My overall impression was that it was an interesting technology but the hardware support was too limited. I’ve continued to keep my eye on it from a distance because VDI is always promising but the expanded use case I keep running into are client that want to “checkout” desktops to allow offline use of the “VDI” session. The primary reasoning is not to streamline desktop deployments but to allow a BYOD solution. So, I thought I’d do a quick hit comparison of the two solutions based on the use case.
What do they do?
From this use case both clients seem to do the same thing. They allow offline use of VDI sessions hosted on their respective backend solutions. You can, in theory and sometimes even in practice “checkout” a desktop VDI session from you vSphere or XenServer environment. The image runs as a VM within the respective client giving “offline” access to the corporate image on potentially resources not owned by the organization.
Both solutions allow the administrator to setup basically DRM to require periodic checkins to ensure the end user has continued rights to the corporate image and ensures that the image is up to date with the latest software and security patches maintained by the administrator.
How do they do it?
Now this is where the differences get exposed between the two solutions. Both clients are based on hypervisor technologies but leverage two different approaches to virtualization. VMware View Client is based on VMware Player/Workstation technology. This means that the View client is a Type 2 hypervisor which means that it would run under Windows (or Mac) and require the end user to have the view client installed on the OS that is factory installed. This allows View to be installed on a wide range of desktops and laptops. There’s not a modern Windows laptop or Desktop configuration that I can think that wouldn’t support a standard corporate installation of VMware View.
On the other side, XenClient is a Type 1 hypervisor. This is also known as a “Bare Metal” hypervisor meaning that XenClient is standalone OS. Type 1 hypervisors generally should perform better than Type 2 hypervisors. This means in theory your corporate image should perform better under XenClient than View. In practice, I don’t know how much truth if any there is in this statement. If you are worried about performance then this approach to VDI more than likely isn’t for you. One major drawback is that the machine needs to be re-imaged and have the consumer OS re-installed. Another disadvantage is that there is a bit less driver support for XenClient. One major requirement had been the requirement to have a certain video card platform.
So which is better?
It depends. Did you really think I was going to make it that easy for you? I doubt you would find that either solution will meet all of your requirements just around this use case. If you are considering either client it’s because you have a larger VDI effort in play. I can almost guarantee that other requirements would drive your decision just as much if not more than checking out a VDI session. I’ve used some variation of both solutions. XenClient is a cool technology but requires a lot of heavy lifting on the backend and a lot of disruptive installation for the client. VMware View traditionally hasn’t had the large enterprise management tools Citrix is known to provide.
If you aren’t satisfied with this answer I’m more interested in why you want to use a VDI solution for BYOD. I’d love to hear your argument for or against VDI for laptop BYOD strategies.