For VDI, coronavirus pandemic could require a new approach

Deploying DaaS or VDI during the coronavirus pandemic

One of the first questions that organizations without existing desktop virtualization infrastructure will have to answer is whether desktop as a service (DaaS) or traditional VDI is the best fit for their needs. Unfortunately, there is no simple or universal answer to this question.

With VDI, organizations host the desktops and applications on their own servers and data centers, but DaaS allows organizations to forego the hosting responsibility and pay a vendor to host the desktops and apps for them.

DaaS is generally best suited for organizations that do not yet have virtual desktop technology in place. This helps organizations avoid the large upfront investment of labor and resources in the infrastructure required for VDI. Deploying VDI is a major project, which might be difficult to undergo during the coronavirus pandemic.

Conversely, VDI offers far more flexibility for business-specific customizations than DaaS. VDI is a good choice for organizations that want to maintain control of the virtual desktop hardware configurations, OS, security and other factors.

Choosing a VDI platform

If an organization decides to deploy VDI technology, the next question is which VDI platform to adopt. Cost is a major consideration, but there are other factors that come into play.

For example, IT departments should evaluate how well a given VDI offering will mesh with their existing infrastructure. For an organization that is heavily invested in VMware, it may make sense to choose a VMware-based VDI offering. The pandemic requires organizations to quickly deploy technology that supports remote workers, so adopting technology from a vendor that an organization has an existing relationship with will minimize the learning curve for admins and reduce support issues.

Another key consideration is whether organizations will support persistent virtual desktops, which are mapped to a specific user. Persistent desktop users can make the same types of changes they would on physical desktop, such as rearranging desktop icons or saving files, and the desktop will save these changes for all subsequent sessions. These desktops can cause more work for IT, however, as these desktops require far more storage.

Conversely, nonpersistent virtual desktops are easier to deploy and maintain, because users receive a clean Windows deployment from one or a few golden desktop images each time that they log in. Nonpersistent desktops do not retain user customizations, so this may lead to some user frustration and hinder UX.

Provisioning and planning a VDI deployment

The process of deploying and provisioning VDI comes with a long list of tasks and additional considerations. One of the most important steps in this process is to ensure the underlying storage infrastructure can deliver enough IOPS to handle the needs of the virtual desktops. Inadequate storage performance is one of the most common causes of sluggish virtual desktop performance.

Another important consideration is what tasks users will do on the virtual desktops. Many VDI deployments are constructed based on the assumption that users will be accessing virtual desktops from within the corporate network. With the current pandemic forcing almost everyone to work from home, however, most users will be accessing the virtual desktops remotely.

This means that IT pros must consider VPN capacity in addition to VDI capacity. IT departments will need to properly license both the VPN and VDI deployments to support the required number of connections. Additionally, IT must provide enough Internet bandwidth to support all the remote VDI sessions.

Even if an organization already has VDI in place and, it probably isn't suitable to allow the organization's entire staff to work remotely. As such, there are some additional steps that IT needs to take to maintain good UX.

One of the biggest problems associated with VDI deployments is usage spikes.

One of the biggest problems associated with VDI deployments is usage spikes. VDI performance will often decrease at certain times of the day due to high usage. For example, virtual desktops often suffer from poor performance at the beginning of the workday because the system is under an increased load as everyone boots their virtual desktops and logs in.

There are a few steps IT can take to minimize the negative effect of the usage spikes. One option is to avoid hosting the maximum number of virtual desktops the VDI can handle. This leaves some extra hardware capacity available to absorb usage spikes but leads to additional costs for the organization.  

Another thing that IT pros can do is set the VDI to pre-boot virtual machines before the workday starts. This will cause the heaviest usage spikes to occur before users start logging in for the day and can reduce the desktop boot times.

A third option could be to work with management to encourage users to work a non-conventional schedule -- especially for workers who don't rely on collaboration. If users work when it is most convenient for them, as opposed to everyone working from 9 a.m. to 5 p.m., then user traffic will be distributed across a wider period of time. This will reduce the maximum load on the VDI at any given time.

Finally, it's important for organizations to think about what it will take to scale the VDI further to meet additional demand. Scaling VDI deployments usually involves more than just purchasing additional licenses.