Gajus - Fotolia
All VDI deployments should include disaster recovery, but IT admins should be aware of VDI's unique dependencies and risks that can complicate a disaster recovery plan.
When VDI sessions leave the data center, they become susceptible to issues -- especially when data spans continents. Even internal thin-client users would lose data if the VDI infrastructure suffered a complete failure.
There are a few options for IT to implement VDI disaster recovery (DR) without incurring massive costs.
Split it up
IT admins can reduce risk by splitting the VDI desktops and supporting infrastructure across two physical data centers. Then, if one data center becomes unavailable, the organization still has access to at least half of the VDI sessions. In this case, IT should ensure that the two data centers do not have shared dependencies.
An organization should have a redundant VDI, but those VDI instances still need to be able to connect to everyday applications. VDI instances should also connect to authentication and management tools, such as Active Directory (AD). IT should understand dependencies and test frequently to ensure that this is the case.
Up to the cloud
If an organization only has one data center, a small environment and a small budget, then splitting up the VDI instances isn't a viable plan for VDI disaster recovery. Organizations in this situation can use desktop as a service to protect their VDI. VMware, for example, provides a managed VDI service on several well-known cloud platforms. IT pros can also use their own Azure AD infrastructure to provide the underlying AD services.
These cloud services, however, cannot currently manage both on-site and cloud-managed infrastructure. To work around this, IT will need to set up two separate sets of infrastructure. It may seem problematic to administer two sites. If the worst should happen, however, IT can quickly and easily spool up new VDI desktops. That's the beauty of cloud environments.
Alternatively, if the organization has a secondary network location, it can set up a secondary VDI infrastructure. With hyper-converged infrastructure, IT can stand up a whole virtual infrastructure within several units' worth of rack space without needing to invest in expensive storage area network infrastructure.
IT cannot organize DR configurations after the fact, however. IT admins will need to be proactive and engage with VMware to take advantage of these setups. IT will also need to perform some legwork, which includes managing the images and configuring groups.
Instant failover products
IT pros can use traditional instant failover DR tools for VDI, such as Zerto and Veeam, but there are some caveats.
VMware Horizon, for example, uses linked mode or instant clones that use nonstandard methods of implementing a new VDI instance in terms of disk configurations and linking. Most DR tools don't react well to this and will likely fail.
Some instant failover DR tools do work with VDI, but IT admins should exercise extreme caution when they use and test these tools. They should also check with the vendors in question, as this may change over time.