By now you've likely heard the news that Citrix has acquired Kaviza. It made quite a splash in the pre-Synergy news today, and it has a lot of people excited about the small to medium business applications of Kaviza's VDI-in-a-Box product. There's also a fair amount of people calling out Citrix for finally admitting (in an indirect way) that XenDesktop is too complex to be used for small businesses. If you're not familiar, VDI-in-a-Box is a turnkey VDI solution that makes it easy to deploy and extend a desktop virtualization system using a grid architecture, local storage, and management VMs. For a complete review of the package, see the review I wrote in December 2010.
We like Kaviza a lot, and have written about them several times. In April 2010, Citrix invested in them (at which point I suggested an acquisition might be forthcoming), and shortly thereafter the VDI-in-a-Box product added support for XenServer and the HDX protocol, which they licensed from Citrix. The elegance in the solution, combined with the fact that they licensed HDX from Citrix as an optional protocol in their product is what prompted us to award them the Best of VMworld award in the Desktop Virtualization category at VMworld 2010.
Why would Citrix want Kaviza?
The short answer to this is "because XenDesktop is too complex for SMBs," but I don't know that anyone from Citrix would go on record to say that. For years we've been saying that the complexity involved in deploying XenDesktop made it more or less an enterprise-only product, while Citrix has been quick to assure us that it's also possible to use in smaller environments. While it may be "possible" to do that, it's not necessarily "practical," and that's where Kaviza comes into the picture.
The long answer goes a bit more like this:
While the numbers being tossed around suggest that XenDesktop has always been targeted at 500+ user deployments, the fact of the matter is that VDI adoption typically starts with small deployments with the intent of growing them. So, while XenDesktop may in fact be targeted at 500+ user implementations, organizations start at much smaller numbers (like 10 - 50) while still dealing with the same complexities of a larger implementation. Citrix has tried to address this gap with XenDesktop Express Edition, which "simplifies" the product by removing all the cool things about XenDesktop Enterprise or Platinum, like Provisioning Services, HDX 3D, XenApp. Even the "Desktop Delivery Controller" part of XenDesktop Express is described as "Limited" in the product comparison chart. You could say that XenDesktop Express is just for simple evaluation purposes, but wouldn't a proper evaluation give you access to everything in an effort to show all of the Platinum features in hopes that you find something you can't live without?
Purchasing Kaviza means that Citrix now has a viable solution for the sub-500 user implementations that can compete against VMware View, which is arguably a less complex solution that XenDesktop. VDI-in-a-box will remain unchanged for the foreseeable future, which is in this case is a good thing. Citrix now has a VDI solution that can be test-implemented in just a few hours, then scaled up to full production in almost no time. The longest single part of setting up VDI-in-a-Box is building the desktop--it's that easy.
Not to be a downer, but...
There are some negative aspects, however, to leaving VDI-in-a-Box unchanged. For instance, in the Q&A published by Citrix after the acquisition, Citrix states that there is no license upgrade path because the solutions are "aimed for different audiences and needs." Some people may be alarmed by that, especially if they already have Kaviza and are on the cusp of needing to upgrade to something like XenDesktop. What is Citrix likely to sell if a company comes to them looking for a VDI solution for, say, 300 users? Do they suggest going with Kaviza, the simpler solution and potential "easy win," or do they go with XenDesktop, which may be too much for that size of a deployment? If a customer chooses Kaviza, then outgrows it in a year, they'll have to buy an entirely new solution.
I'm not sure that VDI has reached a point where adoption so widespread that the audiences can be classified differently based on business type or deployment size. In fact, I think that targeting VDI-in-a-Box only at SMBs with under 500 users would ignore a large part of the market that contains enterprises with small VDI implementations. My point is that the use-case should dictate the product used, not the type of business it is or how many seats are involved. That's a fine philosophy for competing products, but these two products are from the same company now, and I hope to see some sort of license exchange or trade-up program in the future.
Another reason leaving VDI-in-a-Box on an island is potentially disappointing is that there won't be any entitlement to FlexCast (that is, XenApp, Provisioning Services, XenClient, app streaming, etc...). There are workarounds for some things, like connecting to XenApp apps from the VDI-in-a-Box session. Still, that means implementing and managing two separate systems (not that that's much of a stretch). I'm certain this will all be worked out in the future, but as of now there are no formal plans for integration.
The Bottom Line
I was excited when I heard the news, and I'm very happy to see that Kaviza's product has received the ultimate validation that it deserves. I think this is a brilliant move for Citrix, and I'm hopeful that they move quickly to integrate this into the rest of the FlexCast without messing up the simplicity. I can't imagine they would, though, because Citrix already does VDI. They don't need Kaviza for their VDI capabilities, rather, they need Kaviza for their approach to implementing VDI. Who knows, maybe some of the best aspects of each product can blend together, tightening the relationship between the two products and putting in place both a license upgrade path and a hardware upgrade path.
Ultimately, I'm left with a few thoughts:
- Will we see a sort of XenDesktop Fundamentals like we have with XenApp Fundamentals, which is a turnkey solution for deploying XenApp that is aimed at SMBs?
- Will VDI-in-a-Box continue to run on VMware ESX or ESXi? Will it be extended to Hyper-V?
- Will VDI-in-a-Box still be sold at two price points, one with HDX and one without?
- Since VDI-in-a-Box (We need an acronym...I vote for VDIB) runs on XenServer and already supports local storage and image synchronization, can XenServer's Intellicache feature enhance that in any way?
- It's pretty obvious by now, but I'd love to see a license upgrade path. Switching platforms is probably a fairly manual process, especially if you're using different hypervisors, but a license upgrade path seems doable. Customers need to ask for it en masse, though, before it will happen.
As always, we want to hear what you think, and this is sure to bring out the commenters, so have at it! We'll be absolutely sure to catch up with Kaviza and Citrix this week, so if you want us to ask anything specific, let us know.