Open source to bust up Cisco Borg collective?
Can't beat that Sun Microsystems' feelin'
Open...and Shut Microsoft gets all the press for being a reformed monopolist, but in the hoary world of networking, no one has dominated longer or more tenaciously than Cisco Systems. And while Cisco has seen upstart competitors come and mostly go, perhaps none has the chance to up-end the networking giant's comfortable position more than Vyatta.
It's not for lack of trying, or for a lack of chutzpah. Vyatta chief execuive Kelly Herrell has been putting an expiration date on Cisco's fortunes since the small open-source startup was formed, but to little effect.
But something seems to be changing. That something is virtualization, and it's putting a strong wind at Vyatta's back.
A little context is necessary. Cisco competes in a dizzying array of markets. Even within the generic "networking" market, it has a broad portfolio. Probably the best way to describe the meeting point for Vyatta and Cisco is the "secure routing" market. Think of it as "when traffic enters or exits your building on the way to or from the Internet."
Sound niche? It's not. It's an $8bn enterprise market. The average price of traditional, proprietary Cisco hardware in this market ranges from $1 to $100,000.
But don't expect to pay Cisco $1.
It doesn't have to. This particular market exploded in the 1990s when everyone was building out distributed networks. In the process Cisco cornered the market and at its height had 90 per cent market share. With legions of Cisco-trained network engineers, proprietary software locks (it had its own IOS well before Apple's), Cisco had monopoly power that commanded a hefty price in customer conversations.
No one else, including Juniper Networks, was anywhere close.
But something happened on the way to Cisco's absolute dominance of the networking market. Standards happened.
Networks used to rely on Cisco-proprietary protocols, which were the company's technology lockup. Over time, however, networks shifted to Internet Protocol. Vendor interoperability became the reality. Cisco's grip started to loosen on its customers, dropping from 90 percent of the market to roughly 75 per cent in consequence. Other vendors are gaining ground, but it will take years (decades?) to see Cisco slip into the 30 to 40 per cent market share range.
But standards have started the ball rolling.
That ball has been helped along by an unlikely assistant: Intel, a monopoly force in its own right.
Intel back in the game
In the old days, Intel used to have a separate business unit that sold networking chips, which the company sold in December 2007. Pundits speculated that this meant Intel was killing its focus on networking, but they were wrong: Intel had a roadmap that absorbed network processing into the standard x86 architecture. It just wasn't on the market yet.
But it was coming.
The 2010 Intel Developer Event featured more than 10 separate tracks on x86 networking performance. Attendees could attend sessions on "Line rate 10Gb/s," "virtualized networking," "cloud networking." Suddenly Intel had networking in spades.
Chip prices in the networking market hit free-fall mode, to the point that today an x86 server costing $1,000 to $2,000 will outperform more expensive hardware that runs $15,000 to $50,000 per box. Vendors like Vyatta based its chip strategy on Intel, and has seen dramatic performance improvements over the past few years, all paid for by Intel's R&D budget.
Just as Unix gave way to Linux, in large part due to the cost benefits of Linux's x86 hardware, so, too, has Vyatta been banking on a shift from expensive Cisco to a low-cost, open-source Vyatta router platform (PDF).
Gravy train leaves the station
Not that it's been cheap trying to realize this marketing slogan. Vyatta has raised $35m since 2006 from mostly strategic investors: JPMorgan, Arrowpath (a venture capital firm) and Citrix. The company expects to raise its final round in 2011.
Vyatta needs the money. For the first two years of its existence, the company had roughly $0.00 in sales. Instead of generating cash, the company spent piles of the stuff improving its product and evangelizing the value of open networking. But enterprises weren't buying.
But then the company's evangelization efforts paid off. Downloads of the product - now nearly one million total - have increased to a respectable 30,000 per month. Putting this into perspective, the entire midrange router industry ships 150,000 units per month.
Yes, those units are paid for, unlike most of Vyatta's, but the company's downloads have finally started turning into paychecks: since 2008 sales have more than doubled each year, and should grow by 200 per cent in 2011.
All of which is fine, but Vyatta can hardly take credit for all of its recent success. Any good business involves a great deal of luck to become successful, and Vyatta has had luck in abundance.
The most fortuitous trend for the company was virtualization, which has been driving complete IT re-architectures, undermining Cisco's dominant position. Both virtualization and its kissing cousin cloud computing promise radically better economics and flexibility, generally delivered through a cost-effective x86 hardware foundation and both require the workload to run as a software-only virtual machine.
x86. Software-only. Sounds like Vyatta.
For years Vyatta pitched companies to run its software on their old servers, but few did, and essentially no one paid for them. But then virtualization came along and made Vyatta's software platform look very appealing.
Why? At the branch office, a Vyatta VM eliminates an entire Cisco hardware device. In the datacenter Vyatta VMs can be booted at the click of a button instead of having to acquire, rack, cable and install a separate network box.
This can be seen in the company's download numbers, which have not only accelerated in the past two years but have shifted increasingly toward Vyatta's VMs, which were initially a trickle. This bodes well for the startup, as Gartner and other analysts are now tracking networking virtual machines, and one analyst, Infonetics, projects the market for virtual security appliances will explode to $2.4bn in three years. Vyatta has been shifting its model to a usage-based subscription model for networking services, and sales growth suggests the company has finally found its sweet spot.
Cisco, for its part, is unlikely to be able to mount a direct response. Cisco's model is based upon selling expensive, proprietary hardware that generates a 70 per cent gross margin. And while Vyatta probably only challenges $3bn to $4bn of Cisco's gross margin, that's a significant chunk, one that Cisco won't give up without a fight.
Like Red Hat all over again
But how can it fight? Cisco has been expanding into a wide array of businesses that may prevent its network security routers from being hollowed out by Vyatta's open source, virtualization-heavy approach. But it's going to feel the sting, and it's not at all clear that it will be able to mount much more of a defense than Sun Microsystems was able to stage against Red Hat.
All of which is not to say that Vyatta's success is a foregone conclusion. It has significantly increased sales, but it's still a gnat to Cisco. Virtualization has been very kind to Vyatta after years of seeming lost in the wilderness, but others are entering the market and may out-execute the open-source upstart.
But these are risks Vyatta must be very happy to take. Before it had to deal with market indifference. It no longer has this problem. Not at all. ®
Matt Asay is senior vice president of business development at Strobe, a startup that offers an open source framework for building mobile apps. He was formerly chief operating officer of Ubuntu commercial operation Canonical. With more than a decade spent in open source, Asay served as Alfreso's general manager for the Americas and vice president of business development, and he helped put Novell on its open-source track. Asay is an emeritus board member of the Open Source Initiative (OSI). His column, Open...and Shut, appears twice a week on The Register.