There's a lot a wheeling and dealing in the virtualization software market right now -- but not quite as much as some would have you believe.
In the "real news" department, longtime Linux vendor Red Hat (NYSE: RHT ) just spent $107 million in cash to buy privately held Israeli virtualization house Qumranet. The acquisition will be somewhat dilutive to Red Hat's earnings this year, but should add some $20 million of revenue next year.
I'm using that company's main product, KVM, to run a virtualized media server in my bedroom closet, and can attest to the stability and versatility of this virtual machine solution. What KVM doesn't have today is a user-friendly interface. Five years of Unix system admin experience helped me get over a series of snags with setting this machine up, and a lesser geek might have given up or switched to a simpler hypervisor.
Under Red Hat's corporate wing, I'd expect KVM to gussy up with some graphical bells and whistles. For the fedora-sporting company, the attraction in Qumranet lies in controlling a serious virtualization platform of its own rather than relying on XenSource like many others do. Citrix Systems (Nasdaq: CTXS ) owns that solution, and is arguably the main target of this Red Hat attack. The mad hatters will support Xen in Red Hat Enterprise Linux until 2014, but it's never much fun to contribute to your competition's success.
Then there's the rumor mill, where Cisco (Nasdaq: CSCO ) CEO John Chambers got on CNBC last Wednesday and alluded that his company really should do something in virtualization. A few industry bloggers thought that meant taking VMware (NYSE: VMW ) off of EMC's (NYSE: EMC ) hands, and VMware's share price spiked briefly only to fall back down again. That's because there's no way in Hades Cisco is going to buy VMware.
Yes, the networking giant likes to grow by acquisition, like when it bought set-top box maker Scientific-Atlanta in 2005 or consumer networking specialist Linksys in 2003. But Linksys cost $500 million and Scientific-Atlanta $7 billion. VMware's enterprise value today, after a string of competitive challenges and tough-to-swallow news, is around $12.7 billion, before adding the customary 30% or so acquisition premium. That's much bigger than the small-to-medium bites Cisco likes to take.
No, I'm sure that Chambers would look elsewhere. Maybe he's thinking about virtualizing his enterprise-class routers and switches, which would be a unique feature that makes Cisco's hardware more manageable and flexible than competing products from Juniper (Nasdaq: JNPR ) or Nortel Networks (NYSE: NT ) . Or if that's too far left-field, Cisco could come up with new ways to optimize network traffic for virtualized environments, where one network card often plays host to several virtual machines. I don't know how much wiggle room there is left in the venerable Ethernet specification for getting more out of those virtual network interfaces, but someone on Chambers' staff just might.
In the first case, Cisco could pick up an existing privately held or community-supported virtualization expert like Open Kernel Labs or Trango Virtual Processors, both of which know how to handle the sort of microprocessors you're likely to find inside a large network switch or router. But don't expect Cisco to go out and buy something big you might already have invested in. And in the second case, I don't think that anybody is better positioned than Cisco itself to come up with an internal engineering solution to that very specific problem.