Of all the downsides of challenges of DaaS that we wrote about in our book, the most common theme throughout was trust. Trust in the provider, trust in the platform, trusting something that isn’t in your walls, the list of trust issues goes on an on. For many people, the issue comes from simply not having their resources in their own walls, be it for the personal connection to their IT solutions, fear of government raids, or anything in between.
Then again, there’s a lot of benefits that are hard to ignore. Incremental scalability, for instance, means you only have to pay for what you use rather than over-buying resources in chunks. It also means you can dial back resources (and cost) on the fly, rather than being stuck with a VDI solution you built to scale to 2000 people, then scaled back to 1000. You still paid for the 2000-user system. With a subscription model, you simply stop paying for the extra 1000 desktops. That also means the costs are predictable.
Then there’s the ability to move to OpEx-heavy subscription model. We’ve played this sort of game with our physical desktops for years, leasing them instead of buying them outright, but with VDI we had to assume a sizeable CapEx spend and all the accounting complexities that come along with it. Subscription models take that away, which both companies and vendors like. Companies because they can return to OpEx, and vendors because they get recurring revenue at typically higher margins.
Still, over all the DaaS and subscription awesomeness is the dark cloud of trust that many companies can’t get past. So what if DaaS providers spun things around? What if there were cloud nodes that ran locally? What if there really was–GASP!–on-premises DaaS?!
Ok, first, if you build a VDI environment for your company and call it DaaS even though you manage everything about it, that’s on you. It’s not DaaS. Maybe you did it to sound important. Maybe you had to call it DaaS to get your boss to approve it. Whatever it is, you and I both know that it’s not DaaS, it’s VDI. I can keep a secret.
What I’m thinking is the boundaryless datacenter stuff that Microsoft is now talking about, blurring the walls between on-prem and the cloud. With this model, you’d place Azure nodes in your datacenter, and though everything is kept within your walls, it’s still part of the matrix, managed by Azure.
Now let’s take that a step farther and bring desktops into the mix. Imagine Microsoft announces Mohoro and it’s a full-on DaaS solution. Along with the “traditional” DaaS offering, there’s an on-prem version whereby you still pay Microsoft the $30/user/month, but instead of hosting the desktops somewhere in the cloud, they actually send you hardware that you plug into your network. It integrates with your AD, it’s close to your data apps, and users, but it’s owned by Microsoft and is included as part of the subscription. Best of all, it’s still part of Azure, and other than keeping the power on and network connected, managing the plumbing is not your problem.
It sounds like the best of both worlds. You can have your DaaS and eat it too? (Beat the joke...there’s something in there somewhere!)
Maybe Microsoft isn’t going to do a complete DaaS offering, but since they’re the ones talking most about the boundaryless datacenter, they get the nod here. Frankly, though, VMware and Citrix could do this, too. In fact, it could be how Citrix gets into the space. Brian wrote a while ago that Citrix should release their own DaaS solution to compete with VMware (although after the comments on that article, maybe not…).
The biggest problem with that, at least from a logistical standpoint, is that Citrix doesn’t have the massive infrastructure to pull that kind of a thing off quickly. They’d either have to acquire a service provider or build their own service provider solution from scratch. This kind of on-premises DaaS thing could be Citrix’s entrance into the space. All the really need is the be able to manage all the nodes without having to supply all the security, storage, and compute resources from regional datacenters. They could then build upon that later to add features like replication or even resources in the cloud.
Of course, they need a hardware partner to pull that off. So would Microsoft and VMware, for that matter. Who better to provide a solution like this than one of the several converged infrastructure companies? Nutanix, Simplivity, v3, and Pivot3 come to mind, but there are others. Big vendors like Cisco, HP, and Dell also have their own solutions. VMware is even rumored to be making their own converged infrastructure solution called Project Mystic.
The pieces of this are falling into place at the same time vendors and organizations are drinking the subscription model Kool-Aid. What do you think? Is this something that can catch on and help people get past trust issues by just keeping things on premises until they’re ready to move off to the cloud? It can certainly go beyond desktops, too.