I don't see a future for enterprise software infrastructure components. Applications, though, are a different matter. The key difference between a piece of infrastructure software and an application is that an open source developer is a user of infrastructure and knows exactly what is needed, but developers typically have no clue about business application requirements.

It's no secret that open source has turned into a market force, which is giving enterprise software some tough competition. The same can be said for SaaS businesses, which are steadily eating into the market share of the established on-premise players. While it could easily be assumed that ultimately SaaS and open source will battle each other for dominance of the software business, I don't think that's likely. In fact, open source and SaaS are not competitors, but instead natural complements for different parts of the market. In addition, traditional enterprise software is unlikely to disappear completely. However, a good question is: In what areas will enterprise software remain dominant, if in any?

First, it's worth clarifying terminology, as there is more variety and overlap in the business models than is commonly assumed. By pure open source, I mean software that is free, provided with source code, unrestricted in usage, and installed on the customer's premises. An example of this is Apache's web server. By commercial open source, I mean software that is supplied with source code, but may also contain proprietary components and has maintenance and support services, which while optional, are in practice the de facto standard for most organizations. Here, examples would include JBoss and Alfresco. Traditional enterprise software is usually not provided with source code and comes with a hefty initial license fee, coupled with an obligatory annual maintenance charge. Pretty much any software company pre-2000 fell into the latter category. However, modern enterprise software has moved on. License fees are much smaller and maintenance/support/service fees relatively higher. This broad trend comes into sharper focus when comparing the financial statements of enterprise software firms today with those of 10 years ago. In 1999, the typical split for an enterprise software vendor's revenue was 70% license and 30% services. By 2008, this ratio almost flipped so the license piece now amounts to 40% of revenue while the services are at 60%. Basically so-called enterprise software companies have morphed into services companies with a software component.

The reason I'm a bit of a stickler on terminology is that there is now a range of viable business models with the purest form of open source at one extreme and traditional enterprise software at the other, with a number of variants in between - the boundaries of which are not as clear as is generally assumed, and in fact getting more blurred by the day. Also, since the dot-com bubble burst, VC capital, which is such a powerful driver in the software business, has been allocated disproportionately to models with a higher service component and away from the traditional license-based models.