The future of virtual desktops
One day after returning from home I thought it might be a good idea to write a piece on the user workspace, the desktop of the future.
During VMworld it became clear that the desktop as we know is will cease to exist. VMware is working hard to build a new way of provisioning a container in which appllications and user settings are presented. With the latest acquisitions of Airwatch, Cloudvolumes and some internal projects they are heading to where no one has ever been.
VMware’s vision for a desktop is that it is a disposable container in which you add parts to add functionality. In IT terms that means you add VMDK’s to a virtual machine to provide it with applications (on-the-fly) or you add a VMDK to allow the user to access its user settings. The virtual desktop itself is created JIT based on a live running virtual machine you designated as the golden image.
So let’s dive into this proces a bit more, let’s draw this proces so that it becomes more clear to you. What do we need to build a complete workspace for our consumer users.
- Business applications
- User installed applications
- User settings, profile
- Operating system
So if we look at these components and see how VMware will be providing this workspace.
User Requests a workspace >> Desktop is hot-cloned >> Shared business applications VMDK is attached >> User installed applications + profile VMDK is attached >> Workspace is available for the consumer user.
User logs off from Workspace >> Business applications VDMK are released >> User installed applications + profile VMDK is released >> Desktop is destroyed
Let’s take a look at the different steps:
1 – First you need to have a Golden image, a virtual machine that you setup and tuned to your liking. Take a look at the last fling of VMware, they released a tuning tool to do the hard work for you. For most users this golden image will be a Windows desktop system running Windows 7 or 8.1.
..and now comes the fun part, that’s it, make sure the golden image is running, tuned, ready and you’re good to go, VMware will do the rest (in the future with Project Fargo).
2 – With project Fargo they’ve created a method to create a non-persistent desktop of a running machine, they hot-clone the desktop while its running.
3 – Once the newly created desktop is running, its still a child using the memory pages from the parent. Memory pages aren’t copied to the child to be as efficient as possible. All memory pages start out shared and will grow on usage.
4 – Once the user connects to the desktop VMware through Cloudvolumes will attach the neccesary VMDK’s for shared applications and personal applications as well as user settings. All this is done just-in-time.
The user will have access to his or her applications and settings the second he or she sees the desktop, the whole proces of cloning and attaching takes seconds. The user can install applications, change settings and these will be saved in a write-able volume through Cloudvolumes. The shared applications are only read-able and can’t be changed, they are shared among multiple desktops.
All this will happen in a matter of seconds, can you imagine….
6 – When the user logs off, the desktop is destroyed as it is disposable.
It makes for a hell of a desktop strategy if you can add context through Airwatch. With Airwatch it’s possible to determine who the user is, where he comes from and on which device he’s working. With that information VMware can attach a VMDK Just-In-Time.
You might wonder why I mentioned Airwatch, well take a look at this picture we got shown, if that’s not a hint I don’t know what is.
VMware just needs a tool to blok or allow certain functionality and they have a complete solution. They have NSX and with NSX you can create a “ring” around each workspace and determine what goes through or not. But that’s not what you really want (nice for connectivity control), it doesn’t control what you can start within your workspace. For instance, if you want to block the ability to browse the disk drive NSX is useless. What they lack so far is a UEM solution in this where they, based on the context, can handle the user environment, perhaps the next acquisition will cover that.
If you think how powerful the combination of Airwatch, Cloudvolumes and Project fargo are, you see a future that is very interesting.. there one thing, they have to integrate it all, let’s see how they’ll do that.
Hope we will talk about this again soon, I’ll keep you posted.