We’ve spent the past year or so talking about when (or whether) desktop virtualization will take off and replace most corporate desktops. (And just a reminder, the term “desktop virtualization” on BrianMadden.com includes ALL methods of separating OS management from the client device, including VDI, client hypervisors, OS streaming, server-based computing, etc.)
Today’s full-on desktop virtualization successes are limited, with most people using Citrix or Terminal Server SBC solutions, and then a much smaller number of people using OS streaming and VDI.
So if you consider that there are something like 500-700 million corporate laptop and desktops in the world, we probably only have, what, 20 million virtualized today? (I’m saying 20 million even though Citrix has over 70 million ICA licenses sold because the vast majority of Citrix are traditional fat desktops or laptops connecting to seamless apps via XenApp. So in those cases, the end user's desktops are not virtual desktops.)
How does the world grow from 20m to 500m virtual desktops?
Since there's still so much growth ahead of us for virtual desktops, it's interesting to think what it will take to get us there. Certainly there’s a technology component that needs to be considered, in that today's technology just flat-out isn't good enough to handle the myriad of different use cases needed to get these other 480 million users onto virtual desktops.
Then there’s the fact that despite the technology not being ready, the vendors are talking about it like it is ready. And so far a lot of people are being disappointed. (Just to be clear, again, the disappointment is when people try to use this stuff for things it can’t really do. But there are lots of cases where it all works fine.) This notion led to that interesting conversation we had last week about desktop virtualization (or at least “VDI”) being in the “trough of disillusionment.”
But getting back to the question about how we grow this technology from 20m desktops to 500m desktops, the real challenge is going to be convincing “regular” desktop admins that it’s ok to do. The easiest way is to follow Jerry McGuire’s advice and “Show them the money!” Show them a reason to switch that is believable and demonstrable and makes good financial sense. We have a lot of examples of this already, even in the desktop virtualization industry.
For instance, remember how skeptical people were of Citrix and Terminal Server before seeing it for the first time? But both technologies are tremendously popular because they solve a slew of real problems. And remember how skeptical people were of server virtualization before seeing it for the first time? Again, server virtualization was so successful because it just plain worked and the ROI was so simple. In both cases the technologies first won with easy use cases, and then spread from there. (See “killer app.”)
So what’s the easy win for desktop virtualization? What’s it going to take to convince the administrators of the 480 million traditional desktops to switch? It’s not VDI. VDI is a nice use case, but VDI is nothing more than server-based computing with a bit more compatibility. So Citrix and Terminal Server already got the easy win on server-based computing a few years ago, and the incremental advantages of VDI are not interesting enough to reach the other 480 million users.
One potential easy win for desktop virtualization could be around client hypervisors. As I’ve written in the past, I’m most excited about client hypervisors because they provide a common piece of “hardware” for admins to deploy to. So if the worlds’ existing desktop admins can embrace client hypervisors, I think we have an easy way to bring 480 million additional users in the world of virtual desktops.
How do we get client hypervisors to the "other" 480 million desktops?
I generally frame the desktop virtualization industry in terms of five “big” vendors: Microsoft, Citrix, and VMware are the Top 3 (in terms of importance), with Symantec and Quest rounding out the Top 5. With regards to client hypervisors, I wrote about everyone last week. But as a refresher, Citrix and VMware are working on their own offerings, as are two small start-ups, Neocleus and Virtual Computer.
Of those “Big 5” desktop virtualization vendors, only Citrix and VMware are talking about client hypervisors. While I think they’ll both make quality products, I don’t think either one of them will make a dent in the 480 million traditional PC markets.
The problem with Citrix is that their users only know them as delivering Terminal Server-based server-based computing apps. So while that’s worked well for them, Citrix is not the “go to” company people think of outside of their existing niche. VMware has a similar problem, in that they're thought of as a "virtualization" vendor and not a "desktop" vendor. So both Citrix and VMware have to really work hard to get the regular admins who control the 480m remaining desktops to think of them.
Actually, of the Big 5 companies, only two have desktop influence over large swaths of those 480m desktops: Microsoft and Symantec.
So far Microsoft has no plans for client hypervisors. And when you ask them about client hypervisors, they kind of laugh them off, (probably because they want to protect themselves a bit, and let's face it, client hypervisors are disruptive, and existing monopolies are not exactly known for their ability to disrupt themselves).
So that leaves Symantec.
Here's how Symantec can extend desktop virtualization to the other 480m users
First, Symantec needs to buy one of the smaller client hypervisor vendors--either Neocleus or Virtual Computer.
Next (and this is the key part), Symantec needs to integrate the client hypervisor with their Altiris product line, and NOT focus on their Endpoint Virtualization Suite. Why? Let's take a step back before digging deeper into Symantec.
Symantec is huge now, with a lot of products spread across several groups. I want to focus on two right now: the "endpoint virtualization" group and the "Altiris Client Management Suite."
The “endpoint virtualization” group was just formed a few months ago. It owns four products:
- Symantec Workspace Virtualization (app virtualization, formerly Altiris SVS)
- Symantec Workspace Streaming (app streaming, formerly AppStream)
- Symantec Workspace Corporate / Workspace Remote (connection broker, formerly nSuite)
- Symantec Workspace Profiles (profile and user environment virtualization, an OEM version of RTO Virtual Profiles)
The "endpoint management" group, which owns several products, among them:
- Client Management Suite (the old Altiris stuff)
- Package Studio
In the endpoint management group, I'm mostly interested in the Client Management Suite. If you're not familiar with it, it's a lot like Microsoft SMS (now called System Center Configuration Manager) in that it combines software and hardware inventory, automated OS installation, software installation and patching, and remote support capabilities. So imagine if Symantec buys a client hypervisor and sticks that into the Client Management Suite product. That would be like saying, “Ok customer, you manage all these different computer configurations now. What if let you manage these things just like you do today, except every user magically hard the exact same hardware configuration."
Back in March when Symantec announced the new version of their Client Manaement Suite, I wrote a snarky article sort of laughing at them for focusing on this “old” way of managing desktops. And I was even sort of joking with the endpoint desktop virtualization guys about this “old way” and how crazy it was. But since then I’ve realized the “old way” is the perfect path to get to the new way.
The problem with VDI (or with moving an entire desktop to Terminal Server) is that these things are VERY disruptive. I've been talking and talking and talking about how "desktop virtualization" is different and it's hard and game changing.
And then it occurred to me:
As long as desktop virtualization is hard and game changing, no one is going to do it!
So that got me thinking about the easy wins that would apply to the environments of the other 480 million users. And imagine... I don't know how many Client Management Suite seats are deployed, but if Symantec gets a client hypervisor in there, then you'd have tens of millions of regular non-virtualization geeks adopting a form of desktop virtualization without even knowing it.
Then think about how Symantec could take the other assets and capabilities of the Enpoint Virtualization Suite and integrate them into the Client Management Suite. The Client Management Suite already includes the new Workspace Virtualization (their term for "app virtualization"). If they added profile virtualization and app streaming into the mix--wow! Actually the only thing from the Endpoint Virtualization Suite they wouldn't need would be the connection broker (although that might be cool if they wanted to bridge the gap between traditional desktops and server-based remote desktops).
This sets up Symantec for the ultimate solution
Client Management Suite plus a client hypervisor is all Symantec needs to really put a hurt on Citrix and VMware in the traditional client space. But just imagine how they could pull even more ahead.
Even though I like client hypervisors more for the hardware "white washing" than I do for the ability to run mutiple side-by-side VMs, a lot of people have been talking about how it would be cool to run utilities and stuff in parallel VMs on a client device. So for example, instead of installing and running antivirus software on top of Windows, you could download an antivirus virtual appliance that you're run on your client out-of-band. The same can be true for firewall services or for backup software.
Well guess what? Symantec just happens to sell antivirus, online backup, firewall, configuration management, and a host of other capabilities that would fit right into a client hypervisor. And since they'd make the hypervisor and the virtual appliances, they could hook-in via some cool methods, without having to create watered-down VMsafe-like interfaces.
Actually, heck, that would rock. Maybe they need a new CTO to make it all happen! ;)