Importance of Natural Resources

Mike Olson: Open Source Business Models


Back in the early 19th century businesses actually had strategies about electricity, where were you going to get it or were you going to produce your own, now you know what you don’t have an electricity strategy, it’s just part of the landscape. I believe open source software is the same way. It was worth thinking about as a really discrete confusing complicated thing in the ’90s. These days most businesses touch open source in fundamental ways. So if you want to be a business and you want to adjust open source software, heck if you want to write and give away open source software, how do you get paid? A bunch of different things have been tried. One is take an open source project that is not very good and make it your own secret internal project and make it better, but don’t share any of your enhancements and really that’s what we did at Illustra. We took the Postgres code out of Berkeley and made it way better, but made it a proprietary product and didn’t share it with anyone. Sold it for cash money, just as if we were Oracle or Ingres or other proprietary database vendors. And the problem with that is you surrender the contributions of a global community. Your company has to all of a sudden be better than the rest of the planet in the aggregate. Other models that have been tried include, well, you know the last version is freely available in open source, but the current version isn’t, so version lagging. You pay for the current version and the old ones are free. You can do what MySQL and others like JBoss and Red Hat actually gets some credit for doing this. I think Red Hat’s strategy is more nuanced, but just be a services company. The software is free, but if you want support, if you want consulting, you’re going to come to me, you’re going to pay me money. The problem with that business model is that services companies don’t drive real high margins and it doesn’t leave that much money to invest in making the core product better. So in general, those projects tend to languish. Red Hat’s strategy was actually a little more nuanced than that. Red Hat offered a cloud-based management service, Red Hat Network, in order to operate those systems back in the day. That was a proprietary software of theirs, it just happened to run on their data center machinery, it didn’t get shipped out. The operating systems they ran for their customers all of that was freely available, but part of their commercial value was in proprietary software. Sleepycat; last example and then I will talk about what we do at Cloudera. Sleepycat used a technique called dual licensing. So the idea here was we have a good piece of software, but it’s only good if you combine it with software that you have. So if you’re building a mail server or web server you’ll want to put our software in it because then you get good data management services. Well you can do that for free as long as all your stuff is also open source and free. You got to give your IP away as well. If you don’t like to do that, well you can come and pay us money and we will sell you a different license. We called it dual licensing, kind of viral open source for anybody that wanted to do that and if you didn’t want to give your IP away you paid us money. Another way to think about this is, it’s little more jaundiced now that I have the experience of doing this, it’s kind of like distributing the poison and selling the antidote. The good business if you can get it, but your relationship with your customer begins based on a threat and that’s not a really healthy place to start out. So here is what I believe and here is what we believe at Cloudera. Open source innovates faster, spreads faster, does great work faster than any single company can and embrace of the Hadoop ecosystem, giving away software as part of that ecosystem is really important for us. But we have to have reasons for companies to buy our product uniquely and it can’t just be that our people are the most awesome, because it turns out there are way too many awesome people in the world. We’re not going to be able to control that precious resource forever. We can, we do build proprietary software of our own that lets our customers get more value out of their data, out of their infrastructure, operate it better, manage it better, secure it better, that’s allowed to be proprietary from us and it creates differentiation from the pure play open source distributions that are available from some of our competitors. It gives customers a reason to come to us; it gives us a recurring revenue stream that we can then invest back in the open source. We call that really the Cloudera model. It is I think the wave of the future. I think it’s a sustainable way to build an open source business and the best example I can give by the way is this is in some sense what IBM does with Linux. It contributes substantially to the Linux ecosystem, but it builds and delivers database software and middleware and other stuff that runs on top. That allows it to invest back in that infrastructure.


Leave a Reply

Your email address will not be published. Required fields are marked *