everythingpossible - Fotolia

Five reasons cloud-hosted desktop projects fail

Desktops as a service often bring benefits such as easier management, but they don't fit every situation. Tackling a DaaS project onto an OS migration, for instance, is doomed to fail.

On the conference circuit this year, I've noticed that interest in desktops as a service is increasing.

Despite the fact that there aren't a ton of organizations making the leap to cloud-hosted virtual desktops, people are actively searching for a reason to use DaaS. But even they are at the bottom of the hill staring up.

With that in mind, I put together a list of reasons cloud-hosted desktop projects fail. For those who are trying to find a use case -- or they think they already have one -- this can help you kick off your project in the right direction.

Thinking DaaS is about saving money

I'll be the first to tell you that there are financial benefits to DaaS, but saving money is not on the list. You can start doing DaaS without the huge up-front costs that come with building VDI, but the entire cost of DaaS includes the cost of migration, services and additional infrastructure, in addition to the ongoing monthly DaaS service payments.

A cloud-hosted desktop deployment offers predictable costs over time, but knowing what your costs are going to be is a lot different than saving money.

Thinking DaaS is about easier management

You can break a typical VDI or DaaS deployment up into two management groups: the desktop delivery infrastructure and the Windows desktop itself.

DaaS providers take care of all the delivery infrastructure -- the stuff you'd have to do with Citrix or VMware for in-house VDI. For many organizations that could mean easier management, because it takes the complexities of storage, hypervisors, SQL databases, application catalogs and web portals off IT's plate. But it only addresses part of the problem.

Windows management still falls to IT, so if you consider managing applications, profiles, Group Policy Objects, application compatibility, Windows updates, printing, antivirus, browser plug-ins, and more to be the hard stuff, DaaS won't help at all. Organizations can pay someone else to do that -- it's called managed services -- and you don't need DaaS to get that.

Not respecting data center 'remoteness'

Knowing what your costs are going to be is a lot different than saving money.

On-premises IT infrastructure is like a giant Jenga puzzle. Each part is leaning on the other, arranged in such a way that it's unlikely to topple over. But if you start removing pieces, the tower becomes unstable. Moving desktops to the cloud when there is nothing for them to lean on can result in a bad user experience, or worse, a failed cloud-hosted desktop project.

When you move desktops to the cloud, the applications they use have databases that reside on premises. Their authentication and profiles are stored on premises, too. Every time a user logs in, her profile is shipped over not just the LAN but the internet. When a user starts an application with an on-premises data store, that information has to be shipped over the internet, too. All of a sudden, you're using the internet connection more than you used to, and the user experience is far worse than it was with local desktops.

For this reason, it's important to treat DaaS differently. Cloud-hosted desktops are more like branch offices than extensions of the LAN. You may have to relocate some services to the DaaS provider to support the desktops at that location and avoid any significant effect on the connection to the provider.

Forcing DaaS where it doesn't make sense

It's easy to get swept up in the idea of handing off desktop management to someone else, but before you go running to the nearest DaaS provider, ask yourself these questions: Did you fail trying to deliver this workload via VDI or Remote Desktop Session Host (RDSH)? If so, why?

If you failed because you couldn't manage the delivery infrastructure, you might have success with DaaS, but if you failed because data center-hosted desktops don't make sense for the use case, DaaS will fall short, too. There are use cases for DaaS, but don't try to force it where it doesn't belong.

Trying to do too much at once

Just moving to DaaS is a large project, so don't get grand ideas of bolting it onto the end of another project, such as a Windows migration. If the pre-DaaS environment is 32-bit Windows XP with unique images, locally installed apps and a few published apps from RDSH and you're trying to move to 64-bit Windows 7 with nonpersistent images and an application management platform at a DaaS provider, you may have bitten off more than you can chew.

Instead, treat DaaS like climbing Mount Everest. You wouldn't stand at the bottom and make a beeline for the top. Assemble what you need for the OS migration and get to a base camp to get acclimated. Once you've got everything accounted for, then begin the rest of the ascent.

Next Steps

How VDI and cloud-hosted desktops compare

Should you migrate legacy apps to the cloud?

What are the downsides of DaaS?

Dig Deeper on Cloud-hosted virtual desktop and application strategy