Deciphering the 2023 Gartner DaaS Magic Quadrant
The release of this Magic Quadrant has been anticipated for some time as the hosting of VDI workloads in the different public clouds has become increasingly popular, and companies are looking for guidance on which solution(s) they should consider.
Side note: the most critical consideration to make – and the most often not properly executed – is which VDI (or Remote Desktop Session Host (RDSH)) workloads should be hosted in the cloud, if any. "Just because you can doesn't mean that you should," is an architectural principle paramount to successful technology adoption and a delightful user experience, and nowhere does this apply more than with these inherently user-centric workloads. We have a Cloud VDI briefing that dives into this in detail to help you make the best decisions.
The debate surrounding this Magic Quadrant is not so much in who made the grade, but in the degree of separation among the solutions ranked as leaders – most notably the ranking given to Microsoft.
Another side note: before anyone makes accusations of bias or agenda in the following…we have partnered extensively with Microsoft, VMware and Citrix for years and have been going around the VDI/RDSH block and to VDI/RDSH county fairs since the early 2000s. We love them all and hate them all with an equal passion. Just ask us!
The initial reaction among many in the community was to question not only if Microsoft should be farther "up and to the right" of VMware and Citrix but also how there was such a large degree of separation between them. After all, as Gartner themselves point out in the report, Microsoft's VDI solutions have been around for a significantly shorter period than the others. The first iteration of Azure Virtual Desktop (AVD) was released in September 2019 and Windows 365 (Win365) most recently in August 2021.
Tenured VDI professionals know that the competing offerings have existed for several years prior, with Citrix XenDesktop (the first version focused on delivering desktop OS workloads) being released in 2008 and VMware VDM (later rebranded as View, then Horizon View, now just Horizon) in 2007. As such, the relative ranking and positioning of the Microsoft solutions invite serious questions given these significant disparities; for one, what ranked them so much higher after such a short period of time?
This historical bias – or, bias in favor of history – makes it essential to step back and determine precisely what this Magic Quadrant is and, more importantly, is not, and how it should then be viewed and used. In summary:
- What it is: a subjective rating of DaaS solutions that focus on "vendor assembled" and "vendor managed" options predicated by a foundational "strategic planning assumption" that those will account for 80% of the market by 2026.
- What it is not: a rating of "Cloud VDI" solution capabilities, value, maturity and "fit for purpose" determination.
What it is
Per the report, the DaaS market comprises three basic offering types:
- Self-assembled, where the customer assembles the aggregate solution from different components the vendor has available (e.g., compute, storage, networking, persona management technology). The vendor maintains the components, but the customer is responsible for integrating them and managing the hosted workloads.
- Vendor-assembled, where the solution vendor manages the components and the integrations between them and the customer manages the hosted workloads.
- Vendor-managed, where the vendor manages the entire solution and the workloads, up to and including patching and possible end-user support.
The strategic planning assumption that nearly all customers will purchase and consume the latter two options in the next two years leads to an implied devaluing of solutions that either focus on or leverage the self-assembled model. This results in an effective ranking baseline requiring a vendor to provide and manage the cloud hosting components and integrations between them for tasks such as desktop provisioning, power management and session brokering.
This lends to a substantial advantage for Microsoft due to its ownership of Azure. That ownership allows them the potential for the most in-depth end-to-end integration and is explicitly acknowledged in the report. Competing solutions will only ever be able to develop integration capabilities to the extent that Microsoft allows to their beholden public cloud. VMware and Citrix will not own a public cloud platform to compete with it.
100 points (or whatever) to Microsoft.
Critically, this total end-to-end ownership also extends to Windows and the availability of multi-session Windows 10 and 11. Multisession RDSH has always been the path towards the greatest ROI and lowest TCO for virtualized Windows workload provisioning (published apps and desktops), and the same value-add applies to multisession desktop OS workloads.
Microsoft has determined (decreed?) that multisession desktop OS workloads can only exist in Azure – full stop. This is not a technical limitation, only resolvable by the Azure hypervisor. This is a calculated policy decision to drive Azure consumption by making it the most financially compelling platform for these workloads. At present, Citrix and VMware can provision multisession Win10/11 workloads via their solutions on Azure as well, but Microsoft could rescind that at any time – simply because they can.
500 (or 5000, who's counting?) points to Microsoft.
What it is not
At first blush, it may seem the Magic Quadrant should apply to Cloud VDI in general, but it does not. Let's discuss…
By the definition and scope outlined in this report, DaaS does not encompass all the ways in which Windows desktop workloads may be running in various public (and private) clouds. Additionally, the weighted focus on "vendor-assembled" and "vendor-managed" options devalues solutions such as Citrix or Horizon on VMC on AWS, AVS, or GCVE, which would be considered "self-assembled."
Further, the capabilities listed as must-haves or standard in the report (total of 13) may be considered table stakes with no reported differentiation between the competing solutions and no corresponding impact on rankings. However, the capabilities listed as optional (total of 20) are not appropriately weighted, yet they represent the most significant value differentiators! Items such as display protocol performance and features, audio/video optimization, app layering, and support for multi-cloud and hybrid architectures are essential to successful user adoption and streamlined IT administration, and they are not adequately represented in the report.
Even more, solution maturity is not weighted but has an obvious impact on the depth and breadth of capabilities and value. A newer solution focused on "catching up" feature parity with more mature solutions may affect their ability to innovate and truly lead.
Most notably, this report does not address "fit for purpose," which is critical above all (but, in that Cloud VDI briefing mentioned previously, we discuss this in extensive detail to help you make the best decision for your purpose). A given solution may rank high on this report but be entirely wrong for your use cases, and the impact of these shortcomings on your success may not be adequately reflected in this type of report.
Given all the above, the rankings should be no surprise. By scope and definition, Microsoft is going to lead. But, it also should be no surprise that the report's usefulness must be similarly constrained by its scope and definitions and used accordingly. Any form of VDI needs to be done the right way, for the right reasons, with the right user experience to be successful, regardless of what any Magic Quadrant or Mysterious Rhombus (TM by me someday) may say.