Virtualization: A Case Study of the Impact of Open Source

In my most recent post, I wrote about the impact open source software will have on the proprietary software industry. In contrast to many others who seem to feel that open source and proprietary software operate in two parallel but separate universes — that open source is used by people who can’t afford “real” software, while proprietary commercial software is for organizations that need reliability, scalability, and all the other “abilities”, I believe that open source is already challenging the proprietary software world. The anemic growth rates and consolidation of the proprietary software industry reflect the subtle effects of lower cost and easily available open source alternatives.

Fine. That’s my opinion. You may think I’m spot on, or that I’m talking through my hat. How about a case study to test the theory?

Let’s look at virtualization. Virtualization is something I’m spending a lot of time on now, and I think it has tremendous potential with a very clear payoff: reduced costs for IT organizations, both hard (power, machines) and soft (admin and operations personnel). Virtualization evinces an undeniable fact: machines are improving so fast that they make possible — even dictate — a change to the traditional hardware infrastructure, breaking the bounds of the one machine, one application practice used by most IT shops. With such a clear value proposition, virtualization is red hot in IT circles.

In 2004, EMC decided to jump on this trend, figuring virtualization would be a complementary offering to its existing storage business. EMC paid over $600 million to buy VMWare, the leading virtualization provider. VMWare had a very capable, albeit pricy, line of products. VMWare offered them via a hands-on, expensive, direct sales force. These products and their sales strategy melded perfectly with EMC, which sells expensive storage solutions through a hands-on, expensive, sales force.

Today, less than two years later, VMWare has completely restructured its product line and its go-to-market strategy. VMWare offers a significant part of its product line available for immediate download at no cost. That’s right, EMC paid $600 million to buy a company that doesn’t charge for its products.

Why the big change in strategy? In one word: Xen.

Xen is an open source virtualization product emanating from Cambridge University, with a commercial arm called Xensource.

The entrance of an open source product into this market has caused the effective price of virtualization to head toward zero (Microsoft has also announced that its virtualization product will be free, although you have to purchase other Microsoft products to run underneath the virtualization product).

What’s really interesting about this market is how fast commoditization has occurred. Unlike databases, where Oracle has a huge installed base which it can milk at traditional prices, virtualization is a nascent market where user choices are being made today. VMWare faced its own choice: maintain its historical pricing and end up a bit player, or hold its nose, chop prices, attempt to establish a dominant market share, and figure out how to make money from the resulting user base. (For a fascinating look at the financial impact of much lower pricing entering a high-priced market, see Martin Fink’s discussion of the impact of generic drugs on sales of patent-protected proprietary drugs in his excellent The Business and Economics of Linux and Open Source.)

To its credit, VMWare recognized that attempting to maintain its existing pricing was a recipe for irrelevance, and cut its prices with gusto.

Of course, the flip side of this change is that VMWare expects you to download the product and do the work of figuring out whether it’s right for your purpose. They’ll be glad to engage in a sales conversation once you’ve done the exploratory work and decided the VMWare solution is right for you.

To my mind, this poses one of the great challenges open source presents to IT. Unlike databases, where the entrance of open source offerings was supported by a large trained technical workforce (the build up of which was supported by all of that expensive Oracle software), nascent markets like virtualization suffer from a lack of skilled expertise, which makes successful implementations much more difficult — with no skilled technical staff, and a hands-off vendor that can’t afford to spend much time with users, there appears to be an adoption dilemma.

I’m not sure what the answer to this dilemma is, but I’m pretty sure the momentum of commoditization through open source is unstoppable. The next five or ten years should be pretty interesting for the IT industry.