BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Desktop as a service can offer an organization a number of advantages -- it supports distributed workers and avoids the upfront costs of VDI -- but before moving to DaaS, IT should take several important considerations into account.
A good place to start is to determine exactly what services a DaaS provider offers, while factoring in issues related to integration, interoperability, data security and systems management. Most importantly, IT should understand its users' needs and the total cost of moving to DaaS.
Know what services the DaaS provider includes
IT must know how each DaaS provider implements its platform, from whether the provider uses a VDI or Microsoft Remote Desktop Services model, to which hypervisor and display protocol it uses, to how it implements back-end communications.
Because so many factors can affect desktop performance and reliability, IT should learn as much as it can about the service's infrastructure, including the resources it allocates to each virtual desktop. For example, storage performance can have a significant effect on desktops, as can the number of virtual CPUs and RAM assigned to them. Even data center proximity can affect performance.
IT should also determine what mechanisms the provider has in place to ensure availability and to support disaster recovery. Also, it should learn about the provider's support model. If the DaaS provider offers various levels of support, IT might have to pay extra for premium coverage or optional services, such as application virtualization. Thoroughly review any service-level agreements (SLAs) to know exactly what to expect.
Understand integration and interoperability
IT must understand how back-end systems, such as directory services and management products, will integrate into the DaaS infrastructure.
The IT department must also determine whether IT will host back-end systems in the data center or let the DaaS provider host them. Hosting in-house provides the greatest control.
If IT cannot connect to back-end systems directly, it might have to come up with a workaround, such as implementing middleware in the data center to facilitate interoperability. Workarounds add more complexity to the deployment and can potentially affect performance.
If IT relies on other cloud services, it also needs to be sure it can interface with them. In addition, take into account factors such as whether users must access network printers or require multifactor authentication devices, such as fingerprint readers.
Don't forget the users
IT professionals must understand how their users work, where they work, how many users they have, how users are distributed, how many log into their desktops concurrently and the degree to which those numbers fluctuate.
DaaS delivers desktops through the internet, making it ineffective for users with unreliable or slow internet connections. Even if users have good connections, the provider must still deliver the necessary performance, availability and reliability, regardless of the applications users work with or the size of their files. For example, if users commonly work with massive graphics files, the provider's infrastructure must be able to support them.
IT must determine if users require persistent virtual desktops, nonpersistent desktops or a combination and whether it must support peripheral devices, such as USB drives.
Keep data security in mind
Like VDI, DaaS provides services without storing data on the users' devices. So, if a device is lost or stolen, the data is not compromised. But IT must also consider the DaaS infrastructure. Not only should IT account for how data is protected at rest and in motion but also how the entire deployment is protected as users carry out their daily workflows.
IT might also need to consider compliance and regulatory standards, such as the Payment Card Industry Data Security Standard. Depending on the nature of the data, IT might not be able to store any data off-site, out of state or out of the country.
Be aware of management tasks
DaaS providers make sure IT can implement its virtual desktops without having to deal with hypervisors, host servers or other infrastructure. The DaaS provider should also give IT what it needs to deploy images, set up desktops, scale desktops up or down and secure the deployment, but they don't do everything.
It's still up to IT to implement the desktop images, manage Windows updates, deliver the applications and carry out other desktop management tasks. IT should also determine when and how the provider handles updates and maintenance and how these operations might affect desktop performance and availability. It's just as important for IT to determine the level of control it has over the deployment and the degree to which it can customize.
IT also must take into account the endpoints users access their virtual desktops from. If an organization owns the devices, IT must manage them whether they're repurposed PCs, thin clients or zero clients. IT might also have to manage stand-alone PCs to accommodate multiple use cases. Even if users work on their own devices, there might be some management involved.
Another important consideration is how IT plans to migrate users. Moving to DaaS often ends up being a bigger project than IT expects, and it likely needs to support both the old and new deployments during rollout.
Evaluate total cost of ownership
Although IT can save in upfront costs, it still requires resources to manage the desktops and carry out day-to-day operations -- in addition to the ongoing subscription fees, which can add up quickly.
IT also has to ensure that it properly licenses all its software. Licensing can be somewhat tricky in a virtual deployment, especially with Windows because Microsoft has long made it difficult for providers to offer a virtualized version of the Windows desktop. Microsoft has eased restrictions, however, but just for specific Software Assurance customers. It also added virtualization rights for several versions of Windows 10 Enterprise, which makes it easier for providers to offer desktop services.
One big advantage with DaaS is that IT can scale up and down as needed, making the cost more predictable. With traditional VDI, IT often has to overprovision the deployment to accommodate evolving workloads. With DaaS, IT pays only for what it needs, when it needs it, although IT might have to pay extra for premium performance, top-of-the line support and optional services.
The ultimate DaaS vs. VDI comparison
What a DaaS SLA should look like
Custom services to expect from DaaS providers