Open Source Commentary from Navica's CEO, Bernard Golden

July 2006

In This Issue

• Open Source Adoption Patterns -- The Early Adopter/Mainstream User Paradigm is Broken

• Navica News (New Chinese Translation of Succeeding with Open Source)

Open Source Adoption Patterns -- The Early Adopter/Mainstream User Paradigm is Broken

If you've been reading my blog over at CIO Magazine, you'll have seen several posts about my experience at the Burton Group Catalyst Conference; in particular, about the very interesting interactions I had with a number of the attendees.

Many of them were aware of open source, as well as aware of its benefits like lower cost, source code access, and so on, but bluntly said that they need “one throat to choke,” which seemed to preclude using open source software.

One person I talked with said that his organization had done an enterprise portal bake-off, comparing the open source Liferay portal with several other commercial products. Their conclusion: Liferay is easier to install, configure, and manage. However, he continued, “we would never use it.”

On the other hand, the speaker after me, Adam Joffe of Sony Online Entertainment (SOE), described SOE's infrastructure, which supports tens of thousand of simultaneous gamers. SOE's infrastructure is almost entirely open source, comprised of Linux (both Red Hat and CentOS), EnterpriseDB (a Postgres spinoff), and Perl, along with other open source products.

I found the range of reactions to open source quite striking and rather confusing. SOE uses open source as a fundamental part of their infrastructure, while others at the conference refuse to countenance it, reckoning it totally inadequate for their organizations. How can such a divergence regarding the value of open source exist?

I've spent the past couple of weeks pondering this question, and have come to some conclusions about why this divergence exists. Certainly it goes well beyond the standard “early adopter/mainstream user” paradigm most of us rely on to assess product adoption. I think you may be surprised by some of the implications of this divergence, for both vendors and users.

If we look at the enterprises that have thus far adopted open source, even though it is fairly early in the lifecycle of open source, they do not all fit neatly into the early adopter profile.

There are, of course, many early adopters. A number of investment banks have jumped on open source, finding its malleability ideal for their constant search for competitive advantage. Being able to extend a base product via its source code instead of being locked into the vanilla functionality of a closed-source commercial product is ideal for these enterprises. Names like MorganStanley and Goldman Sachs are representative of this breed.

However, there are a couple of other types of companies that have moved to open source. First are companies that use open source as a fundamental part of their business model. SOE, described above, fits into this category. Google, which I wrote about a few months ago, also fits. SalesForce, which I believe uses open source in its SaaS infrastructure, lives here as well.

These companies represent new types of businesses, purely online services, that only exist on the Internet. In other words, they are pure-play information age companies. They have all designed businesses that literally would not be affordable if they relied on commercial software – they would go bust trying to build out their infrastructure. They exist only because free software exists; put another way, open source provides a new set of tools by which interesting businesses can be created that were previously unimaginable.

Another type of business – non-early adopter – has embraced open source. Far from being early adopters or business visionaries, these are companies in old line industries that are suffering intense competition and thereby under incredible pressure to cut costs. Industries like retail, travel, and telecoms. Each of them is rapidly changing, with new types of competitors and drastically shrinking margins.

Companies like Travelocity and FranceTelecom don't have a choice. They have to change their cost structure or face extinction. Somewhat reluctantly, they have embraced open source, and have begun to migrate their people and processes to open source.

So the organizations using open source don't really map to the classic early adopter profile. What about the mainstream organizations? Will they soon begin adopting open source?

The classic definition of a mainstream organization is one that begins using a technology when it is delivered as a “whole product” (what I call in my book a mature product). This is not just the product itself, but a bundle of product, marketing information, high-quality support, services, and so on. Mainstream organizations may even require industry-specific information and product (so-called “verticals”) before beginning to use a product.

Beyond the specific elements of the mature product, there is an unstated implication that mainstream organizations expect technology providers to be “ready to do business our way”: on-site visits, free proof of concept installations, lots of industry-specific information; that is to say, hot and cold running help.

So, the objection I heard at the conference about “one throat to choke” was really kind of a verbal shorthand for “whole product.” For these organizations, the issue is not just commercial support availability. The issue is more cultural. They will use open source when it's available like their other technologies.

There's just one problem with this scenario. Open source is never going to look like its commercial counterpart. The economics are totally different. MySQL does around $40 million per year. Oracle probably spends more than that on direct mail. The expectation that an organization can wait until open source is ready to do business the old way is completely misplaced.

Consequently, I expect that the move to open source within mainstream organizations will not follow the traditional smooth “crossing the chasm” profile. Organizations will adopt open source as they are forced to by company- or industry-specific factors. It's not a matter of the technology growing up; it's a matter of the adopter realizing it needs to do business a different way. My wife constantly quotes something she picked up while working in the psychology field: people only change when the pain of the present outweighs the fear of the future. That insight holds a lot of meaning for open source.

Navica News

I'm pleased to share the news that my book, Succeeding with Open Source, has been published in Chinese by Grandtech, a Taiwanese publishing company. Take a look at the cover.

You can hear me speak at these upcoming events:

August 14, 1:00 p.m.: LinuxWorld Health Care Day. Panel Moderator: "Opening Up Healthcare Markets with Open Source"

September 26, 9:00 a.m.: Virtualization Seminar, New York City

October 10, 9:00 a.m.: "Open Source ROI", Burton Group Catalyst Conference, Barcelona

October 26, 1:00 p.m.: "Using Virtualization in Your Data Center", Data Center Decisions Conference, Chicago