Wow, BYOC and Client Virtualization is hard!

The dust is starting to settle from the successful launch of View 4.5 which includes the industry's first and only truly VDI-integrated offline desktop support via View Client with Local Mode.  As I emphasized in a post back in May, this solution really comes into its own when you consider Bring Your Own Computer (BYOC) and similar scenarios.  But there's an important observation I didn't make back then: "dang this stuff is hard!"
 
It's hard in two macro dimensions:

1. Doing BYOC successfully in the enterprise is hard
2. Building the right technologies that make #1 easier is hard

Let's consider #1 first.  There's a lot of reasons why many analysts remain skeptical of BYOC for laptops taking off soon — in spite of broader trends to device heterogeneity and end user demand for consumerization of IT.  A few good examples include:

• Risk of device failure = Productivity Loss (and IT left holding the bag)
• Protection against data loss (laptop left on the bus, taken form the back of a car, etc)
• Controlling Intellectual Property (IP) and application access (ensuring data or app licenses don't wander off)
• Ability to address all of the above AND still offer integrated, off-network, "work anywhere" mobility

We know that addressing this cohesively won't be a one-size fits all sort of approach.  An enterprise taking this seriously is going to need to bring together several, well integrated technologies to address the needs of different users.  View Client with Local Mode is one important piece of that story.  How we think about that is highlighted well in this little video:

http://www.youtube.com/watch?v=Qza_T06t41U

But why is full integration and mobility of the virtual desktop between client and server useful?  How does it make BYOC easier?  Well, on one hand it can enable occasional BYOC offline use: if you're using server hosted VDI for the typical office bound worker (quite likely with a zero-client), and they need to take that occasional trip to a conference or customer site, a quick checkout to their own laptop before they hit the road can enable several days of offline productivity with minimal IT overhead or concern.  But more profoundly, it can help solve one of the key problems that has plagued other BYOC offerings with offline connectivity — that risk of significant productivity loss if a user manages to mess-up or lose their device.  Consider the following scenario:

1. A roadwarrior checks out a desktop to their laptop of choice.
2. They use it 100% in Local Mode as they roam the world.
3. User changes are replicated back to the datacenter when they have network connectivity.
4. Something bad happens to the laptop (run over by a truck, etc).
5. They are able to immediately get up and running via any Internet kiosk in the world — their VM is brought up and running in the datacenter via the replicated bits — you can do this because the VM formats for View between server hosted and locally hosted workloads are the same.  No messing with in guest drivers, etc.
6. When the roadwarrior has an opportunity to get a new laptop from a local electronics store they do so.
7. They then re-checkout their desktop to run on that commodity laptop as if nothing has happened.

The important point here is that while IT doesn't own or manage the device, they are able to provide business continuity for the end-user in spite of the user's own mistakes in destroying or otherwise DOSing the device.  Productivity is actually improved, and IT is never left holding the bag for the reliability of hardware that is outside their control.
 
So that's a bit on how we're working hard to make BYOC easier for the enterprise.  Now what about the technology itself?  Yeah, this stuff is hard too.  As another vendor's recent attempts to offer BYOC-ish personal compute environments on corporate assets can attest, it's not easy to get this stuff right:

Brian Madden's Analysis of XenClient 1.0

I'm normally not one to wax too philosophic on other bloggers opinions, but Brian's observation at the end of his post does have an additional ring of truth to it:

"Oh, and by the way, how brilliant does VMware's cancel-type-1-replace-with-type-2 strategy seem now?"

I won't comment at this point about what VMware plans or does not plan to do when it comes to type-1 client hypervisors.  But I can say that we have invested over a dozen years of development now in the very platform that makes View Client with Local Mode do its thing.  When you check-out and run a desktop locally, you're leveraging the same battle hardened virtualization technologies used in VMware Workstation, Player and Fusion.  They're not perfect, but they've got some real treadwear.  Will the mouse cursor generally be responsive under normal conditions?  Of course.  Full (non-experimental) support for Windows 7 Aeroglass effects, or 3D apps like Google Earth with 3D buildings?   Yes, we support that on Vista and Win 7 hosts and on most GPUs — Intel, Nvidia and ATI are all fine.  When I unplug USB devices will everything work "normally" and not hang the whole machine?  Uh, yeah.  Will VMware be able to offer this same sort of goodness for the Mac too?  Well, you know by now that VMware folks never make commitments on yet to be announced products, but I can say that developers are hard at work on the right technologies for this in our labs — I think the existence proof of VMware Fusion makes such statements credible.
 
So bottom line, View Client with Local Mode opens up a lot of exciting new opportunities for managed desktop deployments — particularly those with Bring Your Own Computer overtones.  There will be many releases ahead where we continue to improve and refine this stuff.  It's indeed hard, but you can look to get real enterprise value for the right use cases today.
 
Happy View 4.5 piloting!