During the conversations we’ve been having with people about our upcoming DaaS book, it’s becoming clear that many customers aren’t thinking about something that seems like an important aspect of choosing to use DaaS - the exit strategy. I’m sure it’s easy to get caught up in the excitement of your first DaaS relationship, even thinking that this partner is “the one,” but sometimes things just don’t work out. Maybe you just decide you’re not that into it anymore, maybe a better provider comes along, or maybe they stop meeting your expectations. Or they just disappear...
Anyway, the fact that a company has chosen to move their desktops to DaaS means they’ve gotten over the trust barrier that exists when many companies consider cloud services of any type, but there’s a lot on the line when a large number of users are getting access to their apps and data through a third-party, no matter how trusted they are. If you’re looking at DaaS, you need to have two exit strategies, the Controlled Exit Strategy and the Uncontrolled Exit Strategy.
Controlled Exit Strategy
Controlled Exit Strategies are like prenups. You can work them out ahead of time with the provider so that you know exactly what has to be done before you actually get to the point where you pull your desktops out of their datacenter.
First, find out what happens with your data. Did you place your files (or a replica, or whatever) at the provider so users didn’t have to copy their documents across the wire every time they opened them? You’ll have to get that data back and make sure that no copies of it exist at the provider. Do they cache data on their end to speed up transfers and reduce the hit on the internet connection? You should find out how they purge that, not to mention get proof that it was done.
You also have to get a hold of your images. If you no longer plan on using them, deleting them is probably enough to satisfy you, but if you’re planning on using them in-house or taking them to another provider, you better find a way to get them, ensure their integrity, and ensure they’ve been deleted from the former DaaS providers environment. (Of course, if you’re pulling your DaaS desktops just to do your own VDI, you should probably figure out why DaaS failed so that you don’t make the same mistakes in your own environment.)
If you’re using non-persistent desktops, that could be more problematic because you need to get all the other bits as well. Plus, if a company is using a layering solution to enable non-persistence, you’ve got to figure out what to do with that. This might be one of those situations where you just start from scratch. You still have to make sure they’re deleted, though.
Uncontrolled Exit Strategy
The Uncontrolled Exit Strategy is the one you really, really don’t want to have. It’s the kind of thing nightmares are made of, and I can’t think of a worse scenario than waking up one morning to the phone ringing because nobody can access their desktops and the DaaS providers phone is off the hook. The provider doesn’t even have to disappear. There could be a government raid on the datacenter, for instance. Do you think when the FBI seizes the infrastructure related to a company associated with drug trafficking that their removal process will be surgical in its precision? Hell no!
No matter what the cause is, if your provider can no longer deliver desktops, what are you supposed to do? If you planned ahead of time, you might spread the desktops among a few providers, hedging your bets. Sure, that doubles your chance that one company will go under (see the Airplane Rule), but you’ve always got the other one you can lean on in that event. Even if they briefly exceed capacity, it still buys you more time to get things under control than if all of your desktops up and disappeared.
What are your other options? Honestly, I’m not sure. You’ve moved to DaaS, so you’ve probably repurposed your PCs as thin clients or replaced them entirely. You may have your old images that you could drop on to the PCs if you still have them, but that’s not an overnight sort of thing. Your management has likely changed, too, so the pieces to pull off this sort of maneuver might not even be in place anymore. Even if you have backups of everything from the DaaS provider, getting it re-setup at a new provider is going to take time.
You might be able to establish some sort of additional comfort level by looking at the providers financials (if they want to give up that sort of information), but good luck asking Dell or Amazon to do that. Still, maybe larger companies like that are too big to catastrophically fail, so maybe they’re the better bet.
You need to figure it out ahead of time
Look, nobody wants to think about the situation going south. You didn’t sell that part of the solution to your bosses when you pitched the DaaS (or when they pitched it to you). Still, as unlikely as it is, you need a plan. As unlikely as war with Canada is, you can bet the U.S. has a strike plan just in case those exceptionally polite neighbors to the North get unruly. Get your strike plan in order, and if you have suggestions, let us know. It just might make it into the book!