VNETWORKING:  THE CHANGE FACTOR

A Contributed Article by Kelly Herrell, Chief Executive Officer of Vyatta

The evolution of the IT industry is like a pendulum that sometimes has Russian roulette consequences.  An architectural theme starts, gains speed, and then ultimately stalls until a new theme starts.  It takes years to get through a single swing of the pendulum, and the existing industry structure may or may not take a bullet in the process.  It all depends on how disruptive the theme is with respect to existing technologies and vendor business models.

Virtualization is one of these major pendulum cycles.  The compute-side vendors are clearly moving to accommodate it and have been able to take it in fair stride because architectures were already modularized.  Hardware and software separated as open systems during the ‘90s, so inserting a new software layer to virtualize it wasn't difficult to accommodate.  Result:  The pendulum swings, but no bullet.

Not true for the networking side of the industry - you can almost hear the hammer cocking,  The products in the networking industry never modularized in the first place; they're closed, monolithic systems - essentially, mainframes.   That makes them sitting ducks.  As a result, virtualization is ushering in far greater structural change to the networking industry than it did to its computing counterpart.

Some vendors in the upper networking layers (e.g. Citrix, F5, Checkpoint) are responding by modularizing, offering their solutions as virtual machines.  Their reward will be new opportunity in terms of new consumption models such as cloud computing and virtualized branch office network solutions.

But in the lower layers of the network - routing/security, and switching - the impact on the industry will be far more traumatic.  These markets are measured in $10B increments, an order of magnitude larger than the upper networking layers, delivering 70% gross margins to the major vendors.   Those businesses have far too much to lose by modularizing their architectures in an open-systems environment.

That doesn't mean it isn't happening.  It just means the incumbents aren't the ones doing it.  More on that in a moment.

First, acknowledge that virtualized networking's day has come.  Just read what the analysts are now publishing.  IDC:  "It's clear to the industry that, because of server virtualization, a new network needs to emerge."  Yankee Group:  "Virtualization levels the network playing field.  The vendor that solves that problem first has a huge upside."  The pendulum can't be stopped.  The big vendors, however, are trying to stuff a cork in the gun barrel.

The response is fascinating:  Everywhere else in the industry, vendors are modularizing and talking up the benefits of dynamicism and cost-effectiveness resulting from openness and modularity.

But the big network guys?  The opposite.  They're turning their SUVs into Supertankers.  They're trying to trap more and more functionality inside their proprietary systems by subsuming virtualization.  Why?  Because their business models won't allow them to modularize.  If they use the same approach as the rest of the industry, it will destroy their hardware-dependent business models.  And it's not just the vendors who are at risk:  I recently spoke to a multi-billion-dollar distributor of network hardware who confided that virtualization threatens their very existence since software virtual machines simply bypass them in the value chain.

There's more than one bullet coming into this part of the industry; it's an impending war zone.

The networking side of the IT industry needs more nimble solutions, not more unwieldy ones.  We need hardware cost curves that mimic efficient servers, not gold-plated Hummers.  We need standards and openness, not proprietary architectures.  Today's server infrastructure has such great efficiencies - including virtualization, mind you - precisely because the mainframe model broke into components. 

Today there are enterprises, ISPs, telecoms and clouds all running virtual machines at their network foundation.  These solutions are not coming from the incumbents; they're coming from innovative, focused attackers with pure-play software business models and partnering strategies that fit the business economics of virtualization perfectly -- form-fit for the pendulum.

These innovators loaded the gun.  And their massive new partners are trigger-happy.

About the Author, Kelly Herrell

As CEO and member of the Board of Directors, Kelly provides the strategic leadership and vision for Vyatta and drives it through to rapid execution. Kelly has a proven track record for growing companies based on open source in the systems, embedded and telephony industries. Before joining Vyatta, Kelly was the SVP of Strategic Operations at MontaVista Software, the world's pre-eminent embedded Linux supplier, where he led the strategic focus into telecommunications equipment and mobile phones. Today operators around the world such as NTT DoCoMo have network infrastructure and tens of millions of mobile phones in use that are based on MontaVista Linux. Prior to joining MontaVista, Kelly was Vice President of Marketing for Cobalt Networks, the dominant provider of open source-based server appliances for web hosting. After helping drive the company growth through an explosive IPO and beyond, he played a key role in driving its successful $2.3B merger with Sun Microsystems. Previous to Cobalt, Kelly was VP of Marketing for CacheFlow (Nasdaq: BCSI), directed worldwide marketing for Oracle's database products, and served in various product- and market-focused roles at NCR, Teradata and AT&T. Kelly holds a Bachelor's Degree with Honors in Marketing from Washington State University, and an MBA from Cornell University.