Intel buys client hypervisor maker Neocleus for next-to-nothing

I doubt this will result in the embedded solution I've talked about, although I've never really thought about how that would be implemented. Hell, is it even possible to "port" something like Xen to the microcode that drives the hardware?

My thought was that vPro (well, VT at the time) would evolve into being the core hypervisor components, leaving the enterprise-class management to VMware, Citrix, Microsoft, Quest, etc...

Still, IF Intel were to do that, they'd need some sort of low level management, and that's one thing that Neocleus had. If I recall, when we talked to them at Synergy 2009 that's all they really wanted to talk about was the management capabilities, how they integrated with this and that, and so on. Maybe that's something Intel is interested in.

A few more thoughts:

I can totally see the point of acquiring Neocleus for a small price just to acquire engineers, but if the company was running out of money, those guys would've been on the market soon enough anyway. I feel like there had to be something in the product to make it interesting.

Also, if the price was so low, why didn't BigFix make a play. I guess we don't know for sure that they didn't, but the last we heard of Neocleus was that they were being OEM'd by BigFix as an out-of-band management technology:

www.brianmadden.com/.../hey-neocleus-has-a-strategy-now-plans-to-oem-their-client-hypervisor-to-others-first-is-bigfix.aspx

Maybe Neocleus' downfall is a casualty of IBM's acquisition of BigFix.

At any rate, it's put Neocleus and the hardwareification of hypervisors back into my mind for a bit. I'm interested to see what people think of this.

I just read the article of mine that Brian referenced, and I forgot about what Phoenix did with Xen, embedding it into the BIOS. Maybe that's a route Intel could take. Make it part of the BIOS or EFI.

In a situation like that all machines could come with processors capable of virtualization, but whether or not you could use it would be based on the motherboard or the firmware version or something.

I cannot imagine what they did with $22M.  I saw a few demos of the product and at least what I saw didn't really work at all.

This is great news. The "potential" possibilites are great to think about...

It's obvious given the research on Intel's Dynamic Virtual Client initiative that they see use cases for all forms of virtualization. Virtualization layer ontop of Hardware (Type 1), Virtualization layer ontop of local OS (Type 2), and Streamed OS/App (No Virtualization layer).

premierit.intel.com/.../DOC-5549

Maybe this is an opportunity to give a platform refresh for vPro and enabling the possibility for add-on features to CPUs by integrating a common platform, then allow the customer to purchase add-on management capabilities to it as "service VMs". Maybe this is where they can charge for future McAfee solutions?

Even though they purchased a client hypervisor doesn't mean they will utilize it, it was probably a move to swiftly pick up the expertise to develop Service VMs for XenClient. After all, you only need 1 client hypervisor per device and it doesn't make sense to work on XC for as long as they did then duplicate their work again...

As I posted the day the news got out (just read it here, www.wtslabs.com/blog) IMHO I see all these acqusitions by Intel as the tip of the iceberg of what is coming down the road.

No one better than Intel to come up with an embedded hypervisor that fully integrates and leverages the underlying platform.

The main thing now is to watch their next moves in terms of acquisitions. This will give us a clear picture on what they have in mind.

My next guess? Intel buying NVidia. :-)

If they do this, they will be in a position where they can easily dominate the hypervisor market and only leverage all the features to their own layer, excluding VMWare, Citrix and anyone else if they will, exactly like Microsoft is doing with RemoteFX and VMWare is doing for several of their solutions/products (that require vSphere, locking you in).

Not saying this is a good thing, having one vendor that monopolizes this segment. But again, this may happen. Let's see where we stand in 12 months.

CR

Like everyone else, I am waiting to see what Intel's specific plans for these aquired assets are, but I think the odds are that it will be a positive development for the client virtualization space.  

As many have pointed out in past threads, buy-in from the PC OEMs will be key to widescale adoption of client hypervisor technology.  Intel is in a unique position to offer platfrom level functionality to drive mindshare with the OEMs while keeping the playing field open to multiple management vendors, and I believe that is where this is all going.

It is also becoming increasingly clear that Xen open source technology will play a central role in the emergence of client hypervisors.  Intel having a direct stake in the success and evolution of Xen as the de facto "engine" for client hypervisors will positively impact the balance of power within the Xen community and provide important validation for Xen at a time when many in the Linux ecosystem are shifting focus to KVM.

A final point I'll make is that while by all appearances this was not an ideal exit for Neocleus, we should recognize the pioneering work that they brought to the client virtualization space.  I hope that the Neocleus team finds professional satisfaction in whatever comes out of Intel's plans.

Doug Lane

Virtual Computer

@Doug

That is some inspiring insight.

Given Intel's efforts in the Xen community and their recent acquisitions if they deliver "Service VMs" it *hopefully* would be open to other hypervisor offerings. If this is the case, do you see Virtual Computer offering the same as XenClient in terms of "Service VMs"?

A key indicator of the answer to my question is when you said that you believe Intel is moving towards a more open field to multiple management vendors, and that it's a good thing.

Too bad Citrix Receiver for XenClient is tied to the client hypervisor much like Citrix Synchronizer is tied to the server hypervisor. I know they want to ensure adoption of their Xen hypervisors however Citrix should take advantage of the community so I could then run Receiver on something say NxTop, and maybe Synchronizer on something say Hyper-V so I have the option to utilize server based offloading of remoteFX and client based offloading hdx on diverse enterprise platforms.

@Icelus

Yes, service VMs are part of Virtual Computer's strategy going forward, and we actually introduced a service VM architecture to our client hypervisor as part of the 3.0 release we just announced.  It is not yet what I would describe as an open service VM archecture, where any third party vendor can plug into it.  We started off with our own pre-built service VM, which can host things like an isolated web browser, as well as remote display client like the Quest vWorkspace client and Citrix Receiver.

Right now, the admin has some configurability over whether the service VM is deployed and which applications are enabled, but the service VM itself is built and maintained by us.  But the approach is certainly extensible to support third party service VMs, and we would definitely be interested in supporting them / enhancing our architecture as needed if companies like Intel invest in this area.

@Doug

If Intel wanted to purchase a Client hypervisor then they must have come talking to you.

Did they?  If so it means they may be serious in the space.

If not it just means they picked up an ailing company at a knock down price, maybe just as a defensive move.

@Jim

We talk with Intel regularly about our product and how we can best work together going forward, but this was not a situation where they looked at both companies from an M&A perspective.

That said, I also don't think that necessarily means that Intel doesn't have a proactive strategy for the space.  Neocleus was very focused on the security aspects of client hypervisor technology, so it's conceivable that they saw an opportunity to fill in some puzzle pieces around their strategy for McAfee but aren't necessarily trying to wrestle the whole client hypervisor stack away from companies like Virtual Computer and Citrix.