Red Hat buys Qumranet for $107M. What does this mean for KVM and SolidICE?

Red Hat announced that they've acquired VDI vendor Qumranet for $107M in cash. Qumranet only has one product-an end-to-end VDI product called SolidICE that includes a connection broker, a remote display protocol called Spice, and a hypervisor called KVM.

Red Hat announced that they've acquired VDI vendor Qumranet for $107M in cash. Qumranet only has one product—an end-to-end VDI product called SolidICE that includes a connection broker, a remote display protocol called Spice, and a hypervisor called KVM.

The connection-broker features of SolidICE are great, but really not that different than Citrix, Ericom, Quest, or VMware in the grand scheme of things. The Spice protocol is amazing, and much more appropriate for VDI than ICA or RDP. And the KVM hypervisor is a direct competitor to VMware ESX, Citrix XenServer, and Microsoft Hyper-V.

This takes us to the million-dollar question for this acquisition: Which "Qumranet" did Red Hat buy? Did they buy "Qumranet, the SolidICE vendor," or did they buy "Qumranet, the KVM creator?"

The whole first half of the press release is about Qumranet, the KVM creator. Actually, only two of the ten paragraphs in the while press release are about VDI and SolidICE. The rest are all about KVM.

Red Hat also posted an FAQ about the acquisition which is also more focused on the general virtualization value of KVM. However, one of the questions listed is: Does this mean that Red Hat is getting into the Windows desktop business?

Their answer: Red Hat will be able to offer a secure and scalable virtualization platform to Windows desktop customers. Red Hat is focused on providing the best infrastructure upon which to run the complete spectrum of enterprise workloads. This will range from server virtualization to desktop virtualization, which includes Linux servers, Windows servers, Linux desktops, and Windows desktops — all running on and managed by a Red Hat infrastructure.

So I guess really Red Hat wants to be another virtualization infrastructure provider, like VMware, Microsoft, or Citrix. Fine enough. I just can't help but wonder what this means for us in the application and desktop delivery space? I wonder if this will be like Symantec buying Altiris or Microsoft buying Softricity, where the portion that we care about sort of loses focus as The Borg concentrates on the parts of the acquired technology that are more relevant to them?

The upside to Red Hat only really caring about KVM is that they might decided to open source Spice. How cool would THAT be? Imagine being able to snap Spice into any Windows VM running on any hypervisor for free!

A few other random thoughts about this acquisition

Regarding the purchase price, $107M for Qumranet is about is crazy / far out as Citrix paying $500M for XenSource. The press release says that Red Hat doesn't expect Qumranet to contribute materially to Red Hat's FY08 revenue, this despite the fact that Red Hat's fiscal year doesn't end for another five months! Also, this deal is all cash, and $107M isn't too much lower than the cash portion of the Citrix / XenSource deal.

Second, it's getting harder and harder for people to argue that KVM is not a real hypervisor. Even though Red Hat was already committed to KVM, now they're obviously more determined focus on KVM over Xen.

The white paper that Gabe and I did analyzing SolidICE and studying desktop workloads on KVM is just about finished. Hopefully we'll get it published next week. As a quick preview, yeah, we like KVM for desktops a lot!

And finally, this is just another in the ever-growing list of BriForum exhibitors that has been acquired. It just confirms that after six BriForum events, if you want to see the future of our industry, BriForum is the place to do it!

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Fundamentally, I do not think the RH acquisition makes any sort of a big impact. Qumranet is yet another VDI product that will need to compete with Citrix, Microsoft and Vmware whose channel depth is so much bigger that RH's it's not even funny. RH has no channel to speak of so I do not see how it will compete with C,M,V. 

Furthermore, Qumranet's tevhnology is not game-changing. Essentially, it is the same value prop as C,M,V with the big caveat of being new, no channel,  etc. I do not see how RH+qumranet can even come close to Vmware+Pano Logic.



Buy my silence? Didn't I pretty clearly say that I'm not too happy
about this for what it means in the app / desktop virtualization space?

for what it's worth, I found out about this news via a Google news
alert on the press release. No advanced heads up at all.
One of the articles I read last night said that Red Hat is going to "open-source" the SPICE protocol.  That could become very interesting to all the VDI players if you can now use Qurmranet's special sauce for free.
What article was that? Could you give us a link?
The acquisition makes sense and keeps ahead RH ahead of the virtualization curve in the Linux world.  But I'm curious: how would SolidICE work in a large enterprise where teams are segmented by operating system and function?  Who would own the implementation, the Linux team or the Windows desktop team?
Nevermind, SPICE is patented by Qumranet.  I read that twice last night and re-read it before I posted it but now I can't find it...I'm sure it was pulled.
what about the test results that they paid you for. Are they gagging you or are they still a work in progress?

 Virtualization Info had a pretty good write up on this here:

and then they followed up with a good update of where we are today and who does what, similar to Brian's only newer: 

Just becuase Spice is patented doesn't mean that Red Hat can't open source it.
Did you even read the article? I talked about that paper in the article, but in case you missed it, the tests were fine, and we're just working on the paper and documentation. Hopefully we'll get that published this week.
I'm not following you on this statement.  If you've got GigE to every desktop, I think Spice is fantastic, but there are no enterprises with this luxury.  Most enterprises are geographically disbursed and bandwidth is a serious consideration.  I'm not saying that Spice is not a good remoting protocol, but it is not "better" than these other protocols unless it can be used with a greater percentage of the enterprise desktops.  It's new, so hopefully it will continue to evolve to address its shortcomings.

Are we not back to the discussion about which is best? I mean from what i read in the articles brian published on Spice, it seems like a great protocol, and from my point of view it is great to have another live competitor in the market, and seeing RH buying them will most likely give a better competitive edge than if Qumranet had to make their market shares themselves. This way RH can help bring Spice and KVM to the people.

I think it all comes down to needs, features and gut-feeling. RDP, ICA, Spice they all have their advantages, and i think we are back to Ferrari, Lambo, Koenigsegg.. its a matter of taste.

Out toolbox expands good for all of us :-) and even for the competitors i think this makes adoption of VDI even easier with more vendors speaking for the cause :-)

/Rene Vester

I thought they mentioned you only needed 100MB?

Any chance Provision will support the SPICE protocol? If RedHat open sources it? I guess the question is that would it really be beneficial? SPICE seems to take advantage of the fact that it was developed recently and clearly designed to deliver a media rich desktop experience in highspeed low latency network environment. Where ICA and RDP have always been better a media-less low bandwidth average latency environments. Are the optimizations being delivered by Provision capable of deliverying the same experience SPICE does over highspeed low latency connections? If the answer to that question is yes then who cares about SPICE. But, if SPICE is superior to Provision enhanced RDP and ICA + SpeedScreen in LAN environments, I think there should be an effort by Provision to support SPICE for LAN connections and Provision enhanced RDP for WAN and remote connections. How difficult would it be to support KVM virtual machines and broker the SPICE connection? Is this another case of if our clients demand it we will build it?

Adding to Brian's comment - we're in the final phase of the document right now that requires some back and forth between us and Qumranet.  They've been (understandably, now) pretty busy lately.
Haha.. There is NO WAY Spice  protocol will be open sourced. Wihtout Spice, what is there to Qumranet?? KVM is small fries compared to ESX and they don;t have the backing that XEN has. RH wants to take advantage of the VDI "wave" and remain relevant in the marketplace.

Provision Networks will always be on the fringe, they have nothing revolutionary or ground breaking just a mix of advanced login scripts and a virtual channel for RDP. They will fade away or at best remain a product used by that admin guy who hates Citrix, Microsoft and VMware. Go with Vmware dude. Forget Provision.