When I’m talking with clients in my day job as an industry analyst about open source core products, price always comes up. The question is the same: “If this is open source, why’s it so expensive?” Every time. And it’s not a fair question. If companies are delivering value, they deserve to get paid for that. But the reason the question is asked is the fault of vendors selling products based on open source.
The central value proposition of open source core vendors has been freedom from vendor lock-in. After all, the core elements of the product are open source, developed by a global community. The core product isn’t owned by a single company, but, in almost every meaningful instance, by the Apache Software Foundation. If the worst happens and we go out of business, the code will live on in the ASF. You’re safe. If you don’t like us, it’s open source. You’re protected. I don’t know what happens next but hey open source.
It’s a good story. If we’re all honest with each other, it’s not true. But still a good story. And a great way to get in the door and quickly close a land and expand deal at the end of the initial 12-month contract.
At the end of the 24- or 48-month period, the vendor needs to ramp up revenues and increases prices, sometimes drastically. That’s a triggering event for clients to get me and my colleagues on the phone. They ask their users about the features they actually use and value. Can they do without the features that aren’t part of the open source component? In some cases, yeah. Who else is in the market that supports the open source bits? Are they cheaper? Are they the same strategic vendors I’m already working with?
Other vendors smell blood and ask us if they should offer support for those components for cheap. Their customers talk to them about other in-house vendors too.
What’s this all about? AWS recently announced support for managed Apache Kafka. Based on my somewhat narrow Twitter feed, there was a lot of disillusionment from open source advocates about AWS’ offering. Apache Kafka is open source, largely developed and stewarded by the folks at Confluent. The prevailing view was that AWS was profiting off of the work of open source developers, many employed by Confluent. The argument is that AWS was profiting from Confluent’s effective development subsidy of Apache Kafka without contributing anything back in the form of code or docs.
However, Apache Kafka is under the Apache Software License. The ASL doesn’t require users to contribute back to the project. AWS is fully within its rights to use the code as long as it maintains the ASF copyright. (I’m not a lawyer but that’s my shorthand for the conditions of the ASL.) Granted, AWS being within the scope of the ASL doesn’t improve the optics for the open source software community. (Note: Other companies, like Redis Labs and MongoDB, have explicitly updated their licenses to prevent cloud providers from capitalizing on their open source work. These changes haven’t come without a backlash.)
I need to get back to the differentiation challenge. Once end users believe they only need the open source elements of an open source core product, the only differentiations are price and, for some forward-thinkers, both platform-wide integration capabilities and PaaS-style management. For those end users, execution excellence and price matter more than community spirit and OSS dogma. Those same end users talk to their strategic vendors about those products and in a few months AWS starts offering managed Kafka. (Note: Azure already offered a Kafka-compatible API as part of its Events Bus service but it didn’t generate the same negative response from the OSS community.)
This is about AWS, and other cloud providers, responding to customer demand, which is what I’d expect any company to do. It doesn’t appear that their primary customers, which are enterprises, are demanding they contribute back to the open source projects they manage. When that happens, I expect the cloud vendors to respond vigorously. And the OSS community may not like that much either.
One last thought
Ultimately, open source software isn’t the utopia it is made out to be. Open source has been weaponized as a competitive tool. In the past, vendors have been quite candid with me about their intent to maintain control over OSS projects they “own.” Blocking developers from competing firms from becoming full committers is the most common. Creating a competing project is another approach. Another method is creating a consortium around an project to drum up interest, but 95% of the commits come from one company.
The current debate about open source “strip mining” isn’t about contributions, community involvement or anything else other than revenue. Open source remains an excellent development methodology. It’s still not a business model.