Red Hat CEO Jim Whitehurst was in India recently and gave an informative interview to CyberMedia News. In the interview Whitehurst addresses a range of topics--Microsoft's involvement in ODF, Red Hat's product plans for 2008 and 2009, etc.--but it was his analysis of Red Hat's business model that I found particularly engaging.
I have always believed Red Hat's value resides in its ability to deliver certified, stable binaries out of a morass of complex, ever-moving open-source software. But Whitehurst gives a different perspective:
How to make money (in open source)? A lot of people started with the support model...What Red Hat did was fundamentally different...It's not about trying to monetize the bits or the services. Everybody can do that. The key insight is that the development model of open source is around iteration.
Now if NYSE, for instance, has mission critical software running on (open source), the last thing they want is an iterative change impact. Here, we come in and make sure that (open source) is consumable by the enterprise, and is fully QA-strong, fully tested, performance-tuned, certified, equipped with documentation, SLAs, localization aspects, iterative change development, everything. We are the people who do that and ensure stable tested bits on mission-critical deployments. Besides we commit to support it for seven years. It's not just the support but we make it bulletproof.
It's all about how much can you match the pace of iterative integration and make it consumable for the enterprise. Enterprise-class software is not about functionality alone but about change in tandem in a production environment. It's very hard to dynamically change specs, maintain hardware piles, software piles, compatibility, and certifications. If you talk of us, we have monetized not on the OS but on the value it has.
So, yes, Red Hat Enterprise Linux and other Red Hat products are about stable binaries in a changing sea of open source, but there's something more to it. That "something more" is the ability to not only give enterprises an initially firm and fixed binary to work against, but also to provide an ongoing way to update one's deployment in a noninvasive, production-friendly manner.
The next big challenge for Red Hat will be applications. Whitehurst addresses how enterprises can ensure their applications can be deployed on a stable open-source foundation, but with open source increasingly being used at the application layer, Red Hat will also need to figure out how to apply this concept of iteration there.
I personally believe that the secret to open-source applications is enabling the enterprise to modify its applications (a key benefit from open source), while simultaneously maintaining a stable core to the application, and tracking divergence from that core. When the customer decides to update to a newer version of the open-source application, she should be presented with a view into how her modifications have affected her ability to update to the stable, supported core of the application, and information on how best to update her modifications to make them function on the updated core.
Sound hard? That's because it is, but I'm seeing open-source companies starting to do this already. It's a way to deliver the advantages of open source while holding firm to a scalable business model, one that requires a supported application (or, at least, its core). This is critical as more and more enterprises look to. They need to be able to rely on a stable, supported core but also must be able to extend and modify the application to suit their requirements.