Mohoro at last! My analysis of the impact of Microsoft Azure RemoteApp:

On Monday, Microsoft finally announced the details surrounding Azure RemoteApp, or what we previously called Project Mohoro.

The days where we talk about how Microsoft “gets it” or is cool are few and far between around here, but for all the strife they’ve caused us with licensing over the years (among other things), it’s worth calling them out for doing something right once in a while. On Monday, Microsoft finally announced the details surrounding Azure RemoteApp, or what we previously called Project Mohoro. I’m sure you’ve read about it since then, but I wanted to give my take on it.

In short, Microsoft nailed it.

Early on in the rumor cycle, Mohoro was described as “RemoteApp from the cloud,” which, being a rumor, did next to nothing to dispel talks about Microsoft releasing a full-blown DaaS solution with, depending on who you ask, a new version of Windows that only Microsoft can run to avoid licensing issues or sweeping Windows VDA and SA licensing changes that make everyone on Earth happy. It wound up being just as it was described. Microsoft successfully assembled a Windows-apps-from-the-cloud solution without fundamentally changing anything about on-premises enterprise Windows licensing.

There are two versions of Azure RemoteApp, one that is cloud-only and another that is hybrid. The cloud-based solution is fully managed by Microsoft. As the admin, you set up your account, choose from a list of Microsoft applications (mostly Office, but a few others) that you want your users to access, and assign those apps to users. Users log in with their Microsoft account credentials or Azure Active Directory account (if you’re into that sort of thing).

The benefit of the cloud-only model is that all the updates and patching are handled for you. Since these are Microsoft apps running on Microsoft’s cloud, they know what will and won’t work, so it’s 100% hands off after connecting the users to the applications. Of course, it also means you’re limited to a relatively small number of applications.

For more advanced use cases or ones that require more control over the apps and OS, the “hybrid” solution is available. On the surface it seems similar to what you’ve been able to do in Azure for a few years: run your own template images in the cloud. You can install whichever applications you want, connect it to your Active Directory with ADFS, or sync your AD to Azure with DirSync. With this added power comes, however, comes more responsibility on the part of the admin. No longer will Microsoft do all the updating for you. This is your image with your applications, and they don’t want to break anything.

Where it differs from the tradition RDSH server in Azure is that this is a true cloud service. In the past, not only have you been responsible for the image, patching, and updating, but you’ve also been on the hook for scaling the solution. By making this a cloud service with a subscription model, you are now handing over the platform reigns to Microsoft and letting them take care of scaling the solution out. So yeah, you have to create and maintain your image, but once you get it into Azure and provision the apps to users, the rest is taken care of. It becomes a true cloud service that you only need to maintain a small part of.

Focusing on the applications themselves serves to reinforce two things that we’ve talked about for a while: Windows as middleware and the inverse bell curve of RDSH adoption. As we get into the cloud era and further away from caring about desktop applications that require Windows to run, the effort we put into running Windows on-premises becomes greater on a per-application basis. At some point, the number of Windows apps we care about will be so few that it will make sense to put them in the cloud rather than deploy and manage them ourselves.

Azure RemoteApp is currently in Tech Preview, and during that time it is free. Microsoft released a bunch of new clients for Mac, Android, iOS, and Windows, and (at least for the Mac version) the demo of Azure RemoteApp amounted to launching the new Microsoft Remote Desktop app, clicking RemoteApp, and logging in with my Microsoft account. After that it pretty much just works, which is the best part. Notably, there are no Windows Phone or Windows RT clients yet, so even Microsoft doesn’t see developing for those platforms as a priority. (Zing! Had to get in one burn.)

We’ll see what the cost works out to be. Unless things change, the licensing will be per-user for persistent desktops with 50GB of storage on Azure Storage. Assuming this is available through partners, there could also be a base price point with various value-adds.

What do you think? How will this work with Citrix Workspace Services? Will it go beyond a simple “embrace and extend” approach at some point, or will CWS simply do for Azure RemoteApp what XenApp does for RDSH?

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

You missed talking about this versus CSPs (aka ASPs).  These folks provide an option that is in the mix of both forms.  They have the advantage of providing more custom management of the stack and people that can customize whatever you need, although presumably at a higher cost.


@Gabe, time of the market here is key. There are still many new Windows apps still being actively developed and maintained in the enterprise and will continue for many years. There will be modern Windows apps also, reality…cloud only apps will be another app type for customers. What changes is the delivery model since distributed computing models are so cumbersome and the app mix will only get more complex. So it's great that the core infrastructure for delivery is being moved to the cloud for these use cases to ease adoption. This means that the value will shift to how to deliver services (apps/data) and how they will be managed in the future. Simply managing the image in the cloud won't be easy. A good article here talking about CSP opportunities through increased automation. That said I still don't buy that all of a sudden end users will fall in love with this. I've worked in XenApp only environments, where users just don't like it because they are used to working in a desktop style workspace. Microsoft has spent billions of dollars making sure that is what people are familiar with. Published apps are a foreign concept to most users still. So I think the delivery model will have to address both desktop and app only interfaces to keep users happier unless you are talking about basic mass market use cases which is what Microsoft is clearly focused on.


Nice summary Gabe.  

I like the idea of set it and forget it app delivery and not having to worry about scale.  Of course Microsoft really stands to benefit by being the ultimate checkpoint for license use.  It has been historically much easier to violate licenses on traditional Windows desktops.

The future of Windows delivery and  application delivery is heterogeneous. No one app strategy will win out hands down but there will be leaders.

I am in agreement with Harry that the app mix will only get more complex.

You'll always have the difficult apps that need a particular delivery method for any given reason.

Windows (virtual or physical), Citrix Workspace, VMware Horizon Workspace, etc all stand to benefit by pulling together disparate ways to receive apps into a central area.