Why Citrix Workspace Services is the future of XenApp and XenDesktop. (And why it's awesome!)

Last week at Synergy 2014, Citrix showed a prototype of something called "Citrix Workspace Services" (CWS) during their opening keynote. Based on conversations I had with fellow attendees over the next few days, it became clear to me that there's a lot of confusion about what CWS is and whether it matters.

Last week at Synergy 2014, Citrix showed a prototype of something called "Citrix Workspace Services" (CWS) during their opening keynote. Based on conversations I had with fellow attendees over the next few days, it became clear to me that there's a lot of confusion about what CWS is and whether it matters. So I want to clear up the confusion in today's post.

First of all, CWS is a big, big deal. It is the future of XenApp, XenDesktop, and DaaS.

The idea with CWS is that Citrix is using Azure to move the "management plane" of their various products into the cloud. This means that with CWS, you would log into the Azure-based CWS console to set up and configure your Citrix environment, which could include every product Citrix makes, including XenApp, XenDesktop, NetScaler, XenMobile, ShareFile, etc. Then you use the CWS console to select, buy, and connect resources. These resources can be in a public cloud, in your own datacenter, at a partner's site, at a colocation site, etc.

For example, if you wanted to build up a XenDesktop environment, you might link CWS up to a bunch of XenServers running in your own datacenter and a bunch of EC2 VMs on AWS. For disk images, you could pull them from your own systems or the cloud. For data you could plug in your own servers or providers like Box or Dropbox.

What's most important here is that this is not "XenDesktop on Azure." CWS means that you could run your XenDesktop VMs and store your files in Azure, or it could mean you continue to run them in your own datacenter. (Gunnar Berger had a good tweet where he said that CWS is not about moving your servers to the Azure, it's about moving the management consoles to Azure.)

To understand this, think about how you build a XenApp environment today versus how you'd build it in the future CWS world. Today you buy the hardware, install Windows, then install XenApp. Then you install all the other components that XenApp needs—the web servers and the databases and the management servers and all that. Then you use the various Citrix management consoles to get everything configured and setup as you want it. When you're done you have what's essentially a hand-built, self-contained XenDesktop environment.

In the future CWS world, you could imagine that all you do is buy the hardware, install Windows, and then install a simple CWS agent in which you enter your credentials for the cloud. Then you login to the CWS console in the cloud and you see your new server which you can click-and-configure right from there. The XenDesktop of the CWS future could run in your site. Or not. Or both. It really doesn't matter.

What's great about this is that from the CWS standpoint, there is absolutely no difference between a XenApp server running in your own datacenter versus a XenApp server running in a hosted location versus one running in the public cloud. You can even imagine that if you run out of capacity on your own on-premises servers that you could just click-click-click and add XenApp servers into your environment from Azure or AWS or a VMware cloud or SoftLayer or any one of thousands of partners. I'm sure we'll even see simple side-by-side comparisons showing pricing, SLAs, specs, etc. to help you make your decision.

The other cool thing about CWS is that since the physical location of the resources doesn't matter (at least from the platform standpoint), it will be simple to migrate resources from one location to another. I can imagine that a lot of VDI users might first feel more comfortable having their VDI servers on premises, but after a year of that they might say, "Eh, screw it," and add in the next batch of servers from AWS.

And of course, this will span Citrix's entire solution stack. In the future with CWS, it really won't matter where all the individual components live or who owns them. You can just make them all available via your CWS console and move them and change as needed.

CWS is also the future of DaaS

When Gabe and I wrote our book about DaaS a few months ago, we really focused on the fact that DaaS is nothing more than VDI that you pay for as a service. We also wrote several articles about how Citrix needs a DaaS strategy that's better than their current, "Well, we have VDI products which our partners can use to deliver DaaS offerings." We even wrote that Citrix should buy a huge DaaS provider so they can enter the DaaS market directly.

After seeing the CWS prototype, I have completely changed my mind, because building CWS puts Citrix in a much better position than if they had tried to compete as another "me too" DaaS provider.

Thinking about a CWS future, you can see that the "DaaS versus VDI" thing doesn't matter. CWS doesn't care where the servers are that run your desktops. It doesn't care about where the data is, where the gateways are, or where your apps come from. (As an architect for your company, sure, you'll care about all the standard stuff like making sure your servers are near your data, etc. But the CWS platform itself doesn't care.)

So there is no "DaaS" per se in CWS. It's just "desktops." Get them from a box in your company. Get them from AWS. Get them from a local CSP. Citrix, the platform, and the customers don't really care which option you take, or even if you mix-and-match or move them around as you go.

Pfft! This is vaporware!

Several people I spoke to last week accused Citrix of announcing vaporware. To be clear, this is true. What Citrix showed was a prototype that's far away from beta even. But that's okay. What's important is that we understand Citrix's vision for the future of the company and how their products will fit-in to the modern cloud-based world.

And remember, even when CWS comes out, it's certainly not going to be a requirement that means everyone has to use it. (At least not right away.) I fully expect that moving forward, if you want to completely build and deliver XenApp 8, XenApp 9, or XenApp 10 the traditional way where you build everything on your own servers—you'll be able to do that.

It's just that over the next few years as CWS emerges and matures and customers continue to get more comfortable in the cloud, eventually you'll ask yourself, "Why am bothering to buy, build, and maintain my own hardware when I can easily buy it as a service and plug it into the same CWS console I use for everything else?

Think about how CWS could fit into Microsoft's "Mohoro" DaaS offering

Last week Microsoft's Brad Anderson wrote a blog post where he talked about CWS and how it fits into Citrix's "Embrace and Extend" philosophy towards Microsoft's baseline products. You can imagine that the relationship between CWS and Mohoro in the future could be exactly like what XenApp is to RDSH today. (Or what MetaFrame was to Terminal Server in the 1990s.) If Microsoft came out with a base level desktops as a service offering in Azure, then CWS could be used to manage it, publish apps, add support for HDX, etc. Because in that case, a desktop running in Azure is just one of the many resources that could plug into CWS. So Microsoft can focus on making those as efficient as possible, and Citrix can focus on the features that customers need.

The bottom line is that I'm hugely excited about CWS. I'm sure this type of offering is not unique to Citrix, and I have no doubt that VMware will go in this direction too. Ultimately I love the idea that you use the cloud to pull together all the servers, instances, VMs, apps, settings, data, mobility, clients, monitoring, and everything else you need, and that what runs where doesn't matter. (Well, again, it matters to customers, but the platform doesn't care.)

I love that our future is not about "DaaS versus VDI" or "on premises versus off." Even though this is still several years away, I feel good about this direction and Citrix's direction with CWS.

[UDPATE May 19: We just posted a video from Citrix Synergy of Gabe Knuth talking to Citrix's Joe Vaccaro about CWS.]

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

no more product installation but keep the servers where I want... Great !


Thanks for the summary.

This sounds like ScaleXtreme will actually play a much bigger part in CWS than I expected.


Hot off the press, Mohoro (apps) is here - Azure RemoteApp.



I agree that this is an awesome idea.

The control plane design solves a whole bunch of problems that DaaS cannot solve. With DaaS, there are no good answers to how does my desktop get access to my apps & data, or how do you tolerate an extra hop and higher latency?

Plus it helps Citrix do the things it does well - HDX and Management - and leave the complexity of servers and storage to other folks who can solve them better.

The architecture is similar to the cloud based controller architecture for enterprise WLAN developed by  Meraki and Aerohive.

Of course, this is a brand new architecture, and will likely take a long time to be ready for prime time - at least 18 months, if not longer.


Showing powerpoint is easy. Proof is in the pudding.

Citrix copied the entire story from Azure. MSFT is doing the same thing for every product - HyperV can be on-prem or running on Azure. Exchange on-prem or provided as a service, etc. Long term every MSFT product will have a command center in Azure.  Microsoft has already delivered beta of Azure RemoteApp - they are introducing new services every month.

Citrix doesn't understand cloud - they see if as once a year delivery model.  This is the same team that has not delivered migration tools for XenApp. I have no confidence that they can deliver a real cloud product.

I bet they will buy a few more companies in the next 12 months to cook something up by next synergy. When was the last time Citrix actually delivered something new in-house?