Editor's Note: This is part two of this article. Read the first part of Mike Laverick's VMware View 4.5 speculation.
VMware and Microsoft are companies that pride themselves on being "customer focused and customer driven," but behind these corporate platitudes lies an important distinction -- neither company in the current economic climate will devote sqillions of precious R&D funds to merely having a product as a spoiler for their competitor.
Instead, they will focus their quality assurance resources where it counts: in features that different customers in different markets ask for time and time again. From the customers I speak to, the top priority for many isn't a client hypervisor. Sure, they are intrigued by the concept, and in the longer term it may be the way forward, but for many its regarded as somewhat "bleeding edge" for their needs.
The operative word in the previous sentence the word "may." It seems to me the driving force surrounding the client hypervisor is removing from the PC a layer of management that comes with having a regular operating system installed to it. But far from removing this layer, it merely supplants it with another layer. It seems to me the way forward is with much slimmer engines used to connect to a virtual desktop, such as dumb terminal or zero clients such as Panologic.
Despite the attractions of the client hypervisor, many of these customers are struggling with out-of-date computers; for them, the key selling point of the virtual desktop is the ability to deploy Windows 7 to PCs that wouldn't have the remotest chance of making a HCL for a client hypervisor.
This hardware compatibility issue is an enormous challenge. The only thing you can really compare it to is the painfully slow adoption of 64-bit operating systems. Despite the wide spread use of 64-bit CPU in most modern hardware, its still the case that many shop with a 32-bit operating system because of this hardware compatibility barrier.
Personally, I love the idea of the client hypervisor, but the reality is that it's a technology whose time will certainly come...just not yet. From speaking to VMware and my own customers, the verdict seems to be that offline or "Local Mode" desktops fulfill their need to be able to offer the small minority of their users who travel access to the corporate image, even when they are disconnected from the network (as with a long haul flight).
The key problem with previous versions of such offline desktops (as they were once called) is that they were only "experimentally" supported. And of course, the user must remember to synchronise their virtual desktop to their laptop/netbook before travelling, unless they want to download gigabytes worth of data through potential unreliable or bandwidth limited cable or ADSL internet access.
For VMware, working on enhancement and improvements (which I have seen in the beta program) in this "Local Mode" feature of virtual desktop seems more important because that's what customers were demanding. From the various websites I've scoured, it seems clear that "Local Mode" will be fully supported at the time of View 4.5 GA. Remember that this local mode feature works with any device, so long as it has the View client installed and the disk space to hold the virtual desktop and the delta.
At a stroke, the offline or local mode desktop addresses the compatibility concerns that the client hypervisor approach brings. The key sticking point remaining with all "offline" desktops is that the local hardware required to run the virtual desktop might not be up to the job. I don't think this issue will be going away regardless of what VMware (and its competitors) do to improve the performance of getting the virtual desktop on the machine.
Group membership and the virtual desktop
Personally, what I would like to see from all the vendors is the ability to identify roaming users by group membership (or even IP range), and allow the administrator to indicate that whilst the user is on the corporate LAN, a background engine downloads their virtual desktop for the first time. Right now, it is up to the user to remember to "checkout" their virtual desktop. A more seamless process that the user doesn't have to think about or interact with, controlled by the administrator, is a direction I could well imagine VMware looking at in the future.
I don't mean to poo-poo the idea of the client hypervisor, but I think its important to remember that all vendors (VMware included) play the feature-gap game -- flagging up features that their products have that others don't. It's all part of their sales, marketing and PR juggernauts. There's a trap here that I have seen customers fall into time and time again. "Look," they exclaim, "vendor A doesn't have feature B, whereas vendor B does." Dig a little deeper and you find out that it isn't a feature that they want or could even deploy.
ABOUT THE AUTHOR: Mike Laverick is a professional instructor with 15 years experience in technologies such as Novell, Windows and Citrix, and he has been involved with the VMware community since 2003. Laverick is a VMware forum moderator and member of the London VMware User Group Steering Committee. In addition to teaching, Laverick is the owner and author of the virtualisation website and blog RTFM Education, where he publishes free guides and utilities aimed at VMware ESX/VirtualCenter users. In 2009, Laverick received the VMware vExpert award and helped found the Irish and Scottish user groups. Laverick has had books published on VMware Virtual Infrastructure 3, VMware vSphere4 and VMware Site Recovery Manager.