Of all the DaaS topics we've covered this year, perhaps the most important -- and often overlooked -- aspect of...
formulating a DaaS plan is the exit strategy.
You might wonder why you'd plan to stop doing desktop as a service (DaaS) as part of moving to DaaS, but the fact of the matter is that when you place your desktops in someone else's hands, you open yourself up to certain risks -- that are more likely to happen than you'd probably care to admit. For instance, what if your DaaS provider goes down for an extended period of time, or what if it goes out of business altogether?
There are two ways you can part with DaaS if the time ever comes: controlled and uncontrolled. Controlled is where you decide that DaaS isn't right for you or your DaaS provider alerts you that it's shutting down operations well ahead of the actual event. Uncontrolled is when you wake up one morning to no desktop sessions and a 404 error on the provider's website. It could be a protracted outage or it could be that the provider vanished. The reasons for the shutdown are almost irrelevant because whether it was an SEC raid or a tornado, its effect on you is the same. You do not want this situation.
To create your plan, you'll need to learn a few things. Find out what happens to your data, desktop images, delta files -- if you use nonpersistent desktops -- and any other services you placed with the DaaS provider should you stop doing DaaS. You'll likely have a backup of this information on-premises, but if not, how are you going to get that data back? Ensure that the vendor purges the data from its system after you get it back, including any caches, so your stuff isn't floating around after you're gone.
But that's the easy stuff. It's nearly impossible to plan for an uncontrolled separation between you and your desktops. If that happens, you're going to be less concerned about where your images are than you are about getting users connected to their work applications and data. Even if you have a backup of everything, standing up a local desktop virtualization environment as a stopgap takes more than just restoring the information locally. You'll need hardware, software and expertise, all of which you washed your hands of when you moved to DaaS in the first place. Even if you decide to simply go back to PCs, it could take days or weeks to get them all imaged and configured appropriately again, assuming you have the hardware.
Perhaps the best way to mitigate the risk associated with an uncontrolled exit is to spread your desktops around to multiple providers. With some DaaS options -- VMware Horizon DaaS comes to mind, but there might be others as the dust settles -- you can place your desktop images at different providers and dynamically distribute users across the systems. If one DaaS provider vanishes, the other can -- hopefully -- absorb the load while you search for another redundant provider.
The key takeaway from this is that you need to spend some time considering how you'll deal with these situations. Even DaaS providers think you should have an exit strategy and several of them have mentioned that they were surprised at how few customers do. Having an exit strategy is not a lack of confidence in the provider or the overall DaaS implementation; it's just doing a thorough job. Any comprehensive plan should involve risk assessment and contingency planning and backing out of DaaS altogether should be in the discussion.
How are cloud-hosted desktops different than VDI?
Five questions for DaaS providers