I was in San Francisco for Oracle World. I even spoke briefly in Thomas Kurian's keynote on Java middleware. But Neelan and I had to leave on Tuesday and missed the Big Deal: Larry Ellison announcing that Oracle are offering support for Linux.
This is an interesting event from the perspective of the open source business. What are the wider implications?
Oracle are offering support for an open source product that they did not create and don't control.
This is possible for a number of reasons:
- Linux is not a product. It is a class of technology, and companies or organizations assemble, document, distribute and support products.
- Linux support is already commoditized to some extent. Red Hat is only one of several distributors offering support.
- The leadership of Linux is diffused. Linus Torvalds does not work for a big distributor; Red Hat does more than most but no single company provides clear leadership.
- The leadership of Linux matters less than you think. Linux is not primarily an engine of innovation, but an engine of commodization.
However, it's interesting to think about what the limits are for companies such as Oracle in providing open source support, and what the implications are for the open source business.
Let's consider Interface21 and Spring as a case in point. (Or JBoss and Hibernate, or MuleSource and Mule, or LogicBlaze and ActiveMQ, as examples.) Why is this different?
- Spring is both a project and a product. The market does not want different versions of Spring, when they already have one that works out of the box.
- The leadership of Spring is tight and respected by users. Many of the leading developers of Spring were leaders before Spring existed, and would remain leaders in enterprise Java if Spring did not exist.
- The leadership of Spring matters. Spring is not a manifestation of commodization, but of innovation. It has changed the way people develop enterprise Java applications, and will continue to do so. Users don't want just what Spring has done to date, but what it will do tomorrow.
All these protect Interface21's investment in Spring. That magic quality of leadership and creation of the technology adds a lot of value to customers: the ability to provide support from the people who take the project forward, training written by people who have commit rights and talk regularly to the core developers, professional services from people who recognized the problems Spring solved and are continuing to advance the industry. In short, Spring from the source.
Obviously, I spend a lot of time thinking about Spring and the business situation around it. But there is a general point here. Ownership of open source IP is not that important. What matters, is creation of open source IP. A business that does not create open source IP–or a large enough proportion of open source IP–is potentially vulnerable to competitors. In turn, that can mean that that business is not itself viable in the longer term, and that can threaten the viability of the project behind it.
There is a sweet spot where creation of IP is rewarded, but does not create lock-in to a particular company. Let's consider two possible scenarios around open source, if this does not happen:
- Open source is purely tied to the creator, with no choice for services.
- Creation of IP does not produce a return. The business of monetizing open source splits from creation of open source, with the likely effect that open source goes back to being for hobbiests.
Both of these scenarios are bad for consumers and bad for technology.
(1) Customers like open source because it allows competition for services (and the choice of not buying services and rolling their own). They do not want to face a situation where they may have the source, but have no realistic chance of switching service provider if the service provider does not perform.
(2) This scenario (where different folk create and monetize) is extremely destructive. It literally kills the goose that lays the golden egg. Unless creation of IP (open source or otherwise) is rewarded, creation of IP will never be of a volume and quality to satisfy the enterprise.
Finally, before we go too far overboard with strategic thinking, we should remember that there is a simple competitive play in Oracle's announcement. It's clearly targeted at Red Hat. Since Red Hat acquired JBoss, a competitor for parts of Oracle's middleware business, Oracle have clearly been looking for a tough response. However, it's pretty clear that this goes farther and has been the result of longer-term strategic thinking at Redwood Shores. It's not just about open source–it makes sense that Oracle want an operating system, to realize Larry's long-term goal of going from #2 software company in the world to #1.
What does this mean for consumers? Well, if it means that Oracle are committed to Linux for the long term, it's good news. If it's a dumping strategy to get rid of a competitor, it's not so good. Time will tell.
- Oracle Adds New Exhibit to Java Technology Museum
- The Power of Adoption: Why no Company is Big Enough to Deny Developers What They Want
- Interface21 / Spring at Oracle OpenWorld 2006
- Is Open Source Dying? Case Not Proven
- Red Hat Reacts to SpringSource's Leadership