Windows 7 will be huge for VDI. Why? Because it’s not Vista.

If storage prices remain high and vendors expect us to use these expensive solutions with complex connection brokers why move to VDI, vs just staying with Terminal Services on Win2K8 R2 on local disk? If indeed with application virtualization and user workspace management, the TCO will be a huge question mark that VDI needs to address for broad adoption. It seems to me if all I want to do is connect to Windows 7 for single user use, there is a lot complexiity and cost in the current solutions on offer this basic feature. We are being told by the marketing folks that new management models in VDI will answer the TCO question, i.e layer OS, then apps etc, and I question how long those will take to mature to where they need to be, and even when they do, why not apply these management tools to SBC. In the mean time I think most people can do VDI just fine on Win2k8 R2, and have a high level of confidence that since this is same code base as Windows 7, they will have very few app issues. So June 2010, I doubt it will be that soon. In addition I think the client side virtualization platforms will be a huge market and making management tools cover the client and backend will take time as well to mature. Regardless it seems there will be an evolution that will be great for the industry.

The Windows 7 image we use for our VirtualStorm VDI solution is build with Vlite and is only 2.2 GB in size.

Still large but smaller then Vista.

The only thing we really looked at very well are the drivers.

Windows 7 comes with over 2 GB of drivers that are not needed, and make the image slow. Without these drivers and Vmware tools installed it is faster and smaller.

To bad we did not mentioned it on Manage Fusion. I hat it with me.

Regards

Erik

www.virtualstorm.org

@Brian

Great article.  I do have to disagree with you on this point tho: "Even if a customer can use today’s “New Desktop”-type products, the vast majority of people are still using Windows XP. And customers really don’t want to go through all the headaches of a major desktop change and yet still end up on Windows XP."

VDI is become a strategic solution in many company thanks to VMware's marketing engine.  While company are not ready to update their OS platform they are still faced with PC refresh.  There is a ROI by investing into VDI over the traditional desktop.  Sure the ROI is not as good as TS/RDS but the selling point for VDI is the organization can continue to use the "Corporate Image".  That Corporate Image is tried and trued.  Try moving to a TS/RDS/Traditional SBC model and you open up a can of worms.  Everything from application compatibility to skill sets and management.

As far as Win 7 goes.  I totally agree.  I plan on upgrading all my XP instances when the RC is available.

@AppDetective

If the organization is already running TS/RDS it would absolutely make sense for them upgrade to w2k8r2.  This we agree on.  But also remember w2k8r2 is 64-bit only.  There are still 16-bit applications and 32-bit applications that use drivers (this document management/imaging software).  To move from TS/RDS to VDI does not make sense unless there is a specific use case for it.

As for traditional desktops.  Why do they exist today?  Why isn't everyone using TS/RDS?  Most TS/RDS implementation are tactical solutions (remote access, for example).  Cost is a factor, but the ROI shows to be better for VDI than traditional desktop.  I think someday we will get to TS/RDS as a strategic solution for the majority of organizations,  but to get there we have to do VDI first.

Brian's June 2010 number is not far off.  For the last 6 months,  we've been doing nothing but XenDesktop deployments and POCs.

Joe

@Joe

I'm glad you mentioned hardware refresh. Quite often all the talk is about the services that are delivered (SBC/TS and/or VDI) but there is little mention of the desktop device itself.

Please don't think that VDI and SBC/TS always equate to a thin client at the desk!!!

Some requirements such as multi-mon expect a level of "local intelligence" in the desktop device which means that the device is actually something that needs to be managed (patched, upgraded, montiored etc etc). Is this "management" less just because the apps (or even the OS) are being hosted elsewhere? I'm not sure.

Don't get me wrong, I'm a subscriber to SBC but with VDI I'm still not sure. Why virtualise the OS if you have virtualised the applications.

The ROI aspects of all of this are interesting....and often quite emotional !!!

I agree that Win7 is catalyst to VDI adoption... but I dont see VDI as a traditional desktop replacement solution... until client bare metal hypervisor and subsequent management software are released and robust.

@Joe,

I understand your x64 point, but i have found the number of apps that I can't get to run on x64 is very small. It is cheaper for me to deliver those via traditional x32 SBC models as opposed to buying expensive overly complicated storage for VDI as the vendors sell it. Additionally I would consider offloading these kinds on apps to x32 clients using client side virtualization. Management there just is not there yet.

Every time I run the numbers VDI, is never cheaper than SBC. VDI makes sense when you need session isolation and even there, methods exist to provide session isolation without the storage costs, and connection broker overhead for ltd benefit. So I don't buy any of the vendor implementations, they are just inflexible and add to cost.

I could go on about many reasons why everybody doesn;t use TS/RDS today for desktops, but it boils down to protocol performance (multimedia), traditional arhitectures that don't scale and fail and frankly most people don't get the benefits of having remote desktops that can be accessed from anywhere. I am big buyer of VDI (remote desktops accesed from anything, anywhere etc), just not the vendor "do it my way, and buy all this extra crap at greater cost that I don't need"

I get a kick out of every VDI thread that eventually turns into a VDI vs. Citrix/TS debate. Remember folks, VDI does natively what a Terminal Services/Citrix desktop struggles to do even with additional software to "help" it along, especially under a desktop that has a significant load based on application workflow. Published apps yeah well VDI can’t touch it.

As far as Windows 7 driving VDI, it will certainly help but it won’t be the nirvana many claim it to be.

What VDI needs is the same thing that hardcore pc gamers like myself have been saying for years, "We need an OS just run our games" In the case of VDI, we need an OS just to run Windows 32bit Line of Business applications. Idealistic? Yes!

I am sure if VMware could, they would run the apps right in some shell scheduled by the hypervisor. Essentially a hypervisor is nothing but a resource scheduler, same as any OS.