What customers need to know about Windows Virtual Desktop

WVD brings distinct advantages to the table, with a few caveats.

Last week, Microsoft finally announced the public preview of Windows Virtual Desktop (WVD).

You have probably read plenty of posts about WVD already. Today, I will lay out my personal take on WVD, its impact upon our industry, and what precisely you need to know to plan your strategy to deliver desktops and applications to your users.

Where my perspective comes from

For those who don’t know me, I co-founded FSLogix and served as CEO until the Microsoft acquisition. I “retired” to avoid having to work at Microsoft, but that's another story. Since the mission of FSLogix was to make remote desktops suck less, I was interested in all platforms, and I participated in briefings back when WVD was called RDmi.

Now, even though I don't work at Microsoft and do not have any inside information about what’s going on currently, I have spent plenty of time analyzing WVD, and my role as head of FSLogix enabled me to discuss strategy with all the players in this market. Plus, since I’m independent now, I can write about my thoughts on BrianMadden.com. Here we go!

First, a caveat: Things could change

Before being christened “Windows Virtual Desktop,” RDmi went through several name changes, including "Azure Virtual Desktops," and then for about a week, "Microsoft Virtual Desktops." I mention this only to point out that if something as minor as the name has changed multiple times, it is possible that other aspects will also change before WVD ships.

Speaking of the ship date, several people are asking me if I know when it will be. I have no definitive answer (and for their part, Microsoft last week stated that it would be in the second half the of year) but here is what I tell people: A well-run sales organization is laser focused on achieving quota. As an executive, I knew not to derail this by introducing major new products in the middle of the year. Therefore, I expect WVD to ship in time for Microsoft's fiscal year 2020, which begins on July 1, 2019. (Technically that’s the second half of 2019!)

Does the world need another remote desktop solution?

I would have never asked this question out loud while peddling FSLogix solutions, as FSLogix required companies to use remote desktops, or they wouldn't need our software. However, I will ask it now. Why do we need yet another remote desktop solution? Moreover, what does this mean for your application delivery plans? Let's examine this from several perspectives:

  • What does Microsoft have to gain?
  • How does this affect you if you are using a current cloud desktop offering, such as Nutanix/Frame, Workspot, Amazon WorkSpaces, or an MSP like CloudJumper?
  • How does this affect your plans if you currently do desktop virtualization on premises? (Via Citrix, VMware, and so on)
  • Finally, if you manage physical desktops (which, by the way, is 92% of the Windows market*) how does this affect you?

(*At FSLogix, we believed that remote desktops of all flavors make up 8% of the Windows desktop market, with Citrix commanding half, or 4%.)

We know that some vendors cross the lines between these areas, and that some vendors and customers do hybrid deployments, but for now, we’ll just look at this from the perspective of their flagship products.

What does Microsoft have to gain? One word: Azure

One reason I believe that Microsoft thinks that a new remote desktop solution is necessary is due to the amount of existing usage on Azure.

I spoke to someone from Microsoft at Synergy 2017 about Azure remote desktop usage, and they indicated that there were 4 million Azure monthly active users (MAU) using Windows Server Remote Desktop Session Host (RDSH). According to this source, usage was growing at 25% per year. And back at Microsoft Ignite 2018, we heard that that RDS workloads were responsible for 10% of Azure compute hours, and had grown 230% YoY.

Microsoft’s primary mandate, of course, is to grow Azure usage.

Whenever I speak with the leaders of cloud desktop vendors, they all seem to have a similar strategy, which is something to the effect of: "We're going to increase our market share by taking 'N' number of seats from Citrix." (“N” ranges from 1 million to 10 million (!) seats depending upon the vendor.)

It makes sense for Microsoft to want these "taken" seats to go from on premises to Azure. Regardless of the numbers, they do not want any users migrating to a competing platform, such as AWS.

Follow that line of thought, and you will conclude that the same would hold for any physical desktops that could potentially migrate to the cloud.

Another reason is to move on from their previous (unsuccessful) attempt, Azure RemoteApp.

How Microsoft will attract customers to Azure

There are several aspects of WVD that are designed to attract users to Azure. I’ll go over them individually.

Support cycles: Windows 10 multi-user is attractive in several ways. One of the benefits is more frequent release cycles. Windows Server (2013, 2016, 2019) is suitable for many workloads, but remote desktops need to update more frequently than every three years, especially to maintain feature parity with physical desktops.

Resources: Another benefit of Windows 10 multi-user is lower resource usage. The administrators of those 4 million RDSH MAUs will undoubtedly want to trial Windows 10 multi-user. I would guess that this should be easy since WVD is a bring-your-own VM solution.

Licensing: There are many reasons why remote desktop’s market penetration is less than 10%, but the number-one support call involves licensing. This is another issue we’ve talked about forever. Windows 10 multi-user finally addresses this, but again it is Azure only. To those who will say that this will bring up anti-trust issues, I can imagine Microsoft pointing to Apple's iOS operating system as precedent. Have you heard anyone propose requiring Apple to make iOS available on competitive hardware? I haven't.

(By the way, it is important to note that WVD is not required to use Win 10 multi-user—just Azure. CSPs such as Citrix will be able to use it on top of Azure, but with their own management, VDAs, etc.)  

Windows 7 support: As folks still supporting it know all too well, Windows 7 goes EOL next year (January 14, 2020, to be exact). One way to extend your security updates for free is to migrate to WVD.

Office 365 support on Windows Server 2019:  One item that I hear is causing the most consternation is the announcement that Windows Server 2019 does not support Office 365 ProPlus. I would be more concerned, but today there are alternative ways to deliver Office 365 to your users.

Years ago, I used RDSH solely to publish PowerPoint to my iOS device. Now I use the Office 365 iOS app to access PowerPoint. Again, this is another ploy by Microsoft to get you onto Azure, but the Office team still has quotas to meet. I wonder if there is an account large enough to force the Office Customer Success rep at Microsoft to change the policy?

So, that explains some of the reasons why Microsoft thinks we need another remote desktop solution, and how they’re making it attractive.

How does this affect you if you’re a cloud desktop customer?

I already like a lot of things about cloud desktops—both existing ones and WVD:

Low barriers to entry: Getting software deployed is another story, but unlike building your own on-premises VDI solution, getting up-and-running on a cloud desktop platform is much simpler, hardware wise.

Microsoft makes it even easier with WVD licensing. Have an M365 E3 /E5, Windows E3/E5, M365 Business/F1 license? Then WVD is included. Again, it is a bring-your-own-VM solution, so once you have your VMs created (WVD supports N-Series VMs) and populated, you are on your way.

Adjacent resources: The general theory is to keep your data close to your application. If your data is in the cloud, then your desktops can be right next to them.

No capex: Why buy hardware that quickly becomes obsolete? Let’s just say it’s better to rent.

Cloud agnostic: With some obvious exceptions (AWS WorkSpaces, Workspot, etc.), many cloud desktop vendors tout their ability to support multiple clouds, and give you the freedom to move between clouds providers. You may think this is unrealistic technically since it seems difficult to imagine migrating 10,000 VMs from Amazon to Google to Nutanix Cloud and back again. But keep in mind, though, that if one of these vendors believes they can relieve Citrix of 10 million seats by migrating them from on premises to the cloud, what is to stop them from doing the same thing between clouds?

In the long run, this is great for you because it forces the vendors to compete on price. (Obviously, the cloud-agnostic part does not apply to WVD.)

Now, aside from the benefits of today’s cloud desktop solutions, I can think of a few downsides as well. WVD should be able to mitigate some but not all of the others. Let go over them.

Immaturity: Today's solutions are relatively immature. If there were 50 million cloud desktop users on Azure, Microsoft wouldn’t need WVD. The majority of the vendors shipping a solution have very few enterprise customers. Look at their customer list. It is okay for them to be a who's who of SMB shops, but nobody who values their career wants to be a vendor's largest customer.

(It is also okay to ask the salesperson for a private list of significant customers. Lots of corporations these days refuse to allow vendors to say their name publicly.)

Limited availability: The other thing that bothers me about some of the cloud desktop solutions is that they are not available in all regions. Sure, your home office is on the East Coast, and your VMs are in the Azure East data center, but hypothetically, what happens when your salesperson is on the West Coast and needs to do a super important demo and the performance lags? They don't blame the cloud desktop vendor; they blame you (the desktop virtualization admin). The glossy brochures don’t typically point this out, so make sure your solution is in all the cloud data center regions you need before deploying.

Traditional remote desktop shops

Based on my own experience, the majority of customers use Citrix or VMware. WVD has features that you will find compelling, as well as some flaws that may stop you from being able to take advantage until Microsoft fixes them.

Security: WVD utilizes reverse connections, so there are no inbound ports open, and it encrypts the connection from end-to-end. (I am not sure what the encryption will do to your performance, so you will have to pay attention to this during your evaluation.) WVD is also multi-tenant, so you don't have to worry about other customers in your Active Directory, and ADFS will facilitate single sign-on.

Load balancing: This will be a welcome feature. WVD supports both depth-first and breadth-first schemes.

Partner opportunities: WVD provides cloud vendors and other third-party vendors with opportunities to fill in current gaps in the solutions. Monitoring, life-cycle management, and image management are the first ones that I see.

There are a few other things that you’ll have to follow closely:

API limits: First, Azure/WVD designed its APIs (there are both REST and PowerShell) to limit the possibility of denial-of-service attacks. While this is fine for servers, desktops have different requirements. Have 10,000 VMs? You are not going to be able to determine the up-down status using the Azure APIs in a time frame your users need (User: "Hey Support! Is my VM up? Support: "Well it was two hours ago when I checked. Will let you know in 20 minutes..."). 

Clients: Also, because of the reverse proxy, WVD is not compatible with existing RDP clients, so I have asked thin-client vendors, and as of this writing, WVD doesn’t have a Linux client. If you want to utilize thin clients in your WVD installation, make sure to check with your vendor when they will be supported.

Physical desktop shops

Products such as Microsoft Intune/SCCM continue to improve, meaning that applications are easier to deploy to physical desktops. This should be considered in any plans, as Jack pointed out recently.

Lastly, since WVD is “free” with many types of licenses (you just pay for the workload compute costs) many shops might consider it for certain use cases that may have been too marginal in the past, like using remote apps to support device choice programs.

Wrap up

Keep in mind that the value of this "tiny" market is greater than $10 billion a year, so you can see why all of these vendors are anxious to help you spend your dollars.

This is my take on WVD and its impact upon the industry. Let me know what you think!

Dig Deeper on Cloud-hosted virtual desktop and application strategy