Yesterday Gabe wrote an article questioning whether Citrix would be able to catch up to VMware if VMware extended Horizon View to support RDSH and/or app publishing. Many of the reader comments on that article were variations of, "XenApp has many features and it's more than just a protocol on RDSH. So Citrix has no need to worry."
It's true that XenApp has a lot of features that are important for on-premises, traditional hardware-based deployments. But how many of those features matter in today's environment where everything is virtualized? And how much will they matter as we evolve from virtual datacenters to cloud-based datacenters?
What?? The cloud is years away!
Whenever I try to explain this to people, people are quick to disagree, protesting that the cloud is years away and that many companies won't trust the cloud. That's true for the public cloud. But remember those same platforms that run public clouds are making their ways into on-premises corporate datacenters. So when we talk about a particular technology running on a cloud platform, that doesn't automatically mean that we're talking about the public cloud.
For example, Microsoft has started talking about how future versions of Windows Server will be more like "mini on-premises instances of Azure." This is part of Nadella's whole "boundaryless datacenter" concept, (or what others are calling the hybrid cloud). The idea is that you run the same platform in-house as the cloud providers, which allows you to seamlessly move VMs and other workloads between your on-premises installations and public cloud providers.
But what most people miss is that this concept applies even when companies are afraid of the public cloud. If you don't trust the public cloud, fine, then don't use it. But everything you run in house will still be based on the same platform (PaaS or IaaS or whatever) as the public cloud providers.
So Microsoft is going after this big time with future versions of Windows Server. VMware is doing this with vSphere, Citrix has OpenCloud, etc. This is more than the pejorative "everyone is calling their traditional datacenters 'private clouds' now." This is the future on-premises datacenter that will run the exact same cloud platform as public cloud environments.
What's the value of XenDesktop / XenApp when running on a cloud platform?
Given that, let's go back and look at the main features of XenDesktop and XenApp today:
- HDX protocol
- Seamleass Windows / app publishing
- Session brokering / load balancing
- Server & app provisioning
- Image management
Now think about what features these cloud platforms provide:
- Image management & assembling
- VM deployment
- Connection brokering / load balancing
- Elasticity / VM migration
- Dynamic resource allocation
So if you're running a desktop virtualization environment on a cloud platform, what features of XenApp or XenDesktop do you really need from Citrix that your cloud platform doesn't provide?
- HDX protocol
- Seamless Windows / app publishing
Now, imagine if VMware wanted to enter the space to compete more broadly with Citrix XenApp? If VMware builds their solution for their cloud platform, suddenly the list of things VMware has to offer isn't a huge massive list. Instead it's pretty much just a protocol and seamless windows / app publishing. VMware already has the protocol. So now all they really need is app publishing. (Which itself is seamless Windows, a way to assign apps to users, knowing which apps are available on which hosts, etc.)
This is the only logical path for them, both in terms of the ease of building something and the chance of success with it. If were were VMware right now, would you really try to do a feature-by-feature replication of XenApp, especially when you're trying to elevate your customers to a true cloud platform instead of a bunch of Windows VMs running on hardware in your own datacenter? Of course not! Trying to build today's XenApp-for-VMware would be like Gretzky's whole "skating to where the puck is instead of to where the puck is going" thing. At this point in 2014 VMware specifically should not try to do that.
Instead VMware needs to leverage the capabilities of their cloud platform and just add the few bits they're missing instead of trying to build a XenApp clone from scratch.
What does this mean for the value of XenDesktop and XenApp?
The other interesting thing to think about here is that in that future world where everything (public and on-premises) is a cloud platform, what value is Citrix XenDesktop/XenApp really adding on top of a cloud platform? Are they just a protocol and seamless apps? If so, can that continue to be a multi-billion dollar business for them?
We've also speculated that Citrix might be able to tweak XenDesktop/XenApp so they can run on top of any DaaS offering that Microsoft would do from Azure. (i.e. "Mohoro.") But considering the points in this article, if Microsoft released Windows desktops and apps from Azure, why would Microsoft need Citrix? Microsoft already has RemoteApp and RDP 8. Seems like if they got their act together with some decent clients (including RDP zero clients), then seamless windows from Azure doesn't need Citrix at all?
Meanwhile Citrix is struggling to convince customers to move from XenApp 6.5 to XenApp 7.5, which is tough because 7.5 loses a lot of features and is still designed as a legacy platform. ("Legacy" in that XenApp itself handles many of the tasks that are built-in to the cloud platforms.) Customers have shown little interest in 7.5 (partly due to the lack of features and partly because it's a "rip and replace.") So it seems like this would be a great opportunity for customers to rip out XenApp and replace it with an on-premises cloud platform from VMware or Microsoft if such a product were to exist.
(Combine all this with the uncertainty around Citrix's new CEO, and yikes, there are a lot of balls in the air right now . . . )