I can’t think of anyone who has had a more epic impact on the history of open source software than Brian Behlendorf. He’s the co-founder and/or board member of a raft of open source communities, including:
Apache Group, which became the Apache Foundation
Electronic Frontier Foundation
He’s worked at the White House Office of Science and Technology Policy on open source issues. Now, he’s looking for companies “out to significantly change the world over the next ten years.” (his words)
He’s also the newest member of the Cloudscaling Advisory Board.
Brian and I caught up at OSCON a few weeks ago. We talked about his journey in open source, the birth of Apache, interoperability, and advice on how to be a great open source foundation board member.
Here’s the play-by-play:
Randy Bias: Brian, how do you describe yourself?
Brian Behlendorf: I like to describe myself as an open technology maven. Currently, I’m on the board of the Mozilla Foundation and the board of a non-profit called Benetech that does software for the human rights movement, environmental monitors, that sort of thing. I also just recently joined the board of the Electronic Frontier Foundation, where open source plays a role in helping defend civil liberties.
My day job these days is as a senior technologist with an investment company called Mithril, which is trying to figure out what the world will look like in ten years, then help it be as awesome as it could be.
What started your journey in open source?
It started as a college student at UC Berkeley, learning about UNIX. Berkeley was the home of UNIX those days. It was really an education, not just in a technology but in a culture, a culture that permeated both the UNIX operating system and the nascent Internet at the time. It was a culture of collaboration, one where you never had all the smartest people right next to you, so you’d naturally turn to email, and online mediums to share what you’re working on and try to see if it’s any good or not.
And that evolved into working on Apache?
I was trying to not be a dropout at Berkeley – and failing at that – I was also working at Wired magazine, where I’d set up their first website, and we started asking ourselves if we could make money on this by putting ads in place and original content. And so the free web server that I was using was the NCSA server, and I just started fixing bugs, adding features and sharing all those improvements with other NCSA web server users.
After the original developers all left to this new company called Netscape, we said, “Well, why don’t we just continue doing this work ourselves and continue this under a new name?”
And that was the birth of Apache.
So that was the birth of Apache. What about the birth of the Apache Software Foundation?
Apache got started around 1995. Three years in, it was much more successful than we ever thought. It was running on two-thirds of the web. The Vatican was running it, the CIA was running it. It was kind of crazy to us. We started to get a lot of people giving us advice such as, “Gee, if you implemented somebody’s patent and they came after you they’d sue you and they would take your home, take whatever assets you had,” so we thought that would suck, and that we needed some sort of shell to protect that.
But then we also realized, hey, it’d be nice to be able to move on from this project at some point, you know, and have a structure that could outlast any one of the individuals. None of us wanted to be a tyrant forever on this project, right?
Now there are many, tens, or maybe even hundreds of projects under ASF, right?
There are more than 150 different projects, counting the Incubator projects, official projects, and each one of those has active communities, so a dozen or a couple of dozen – I think there’s over 1,500 people – who’ve committed or contributed a patch. There’s about 400 members. I may have those numbers a little bit off, but it’s still – it’s a pretty broad stroke. It’s not just web servers anymore.
Right. So I’m on the OpenStack Foundation board of directors, and the OpenStack Foundation is like the Apache Software Foundation, but it was actually designed – it was more purpose-built, like it was really to serve the interest of infrastructure as service clouds. So when you look at something like that and you compare it to the Apache Software Foundation, which has a much broader mission, how do you think about that, having been in open source for a very long time?
There are a number of different organizations in this space. ASF is one kind of model. It’s very member-driven, very kind of chaotic. There’s not really a clear road map, all these individual projects do what they want and it’s kind of like let a thousand flowers bloom. There’s other projects like Eclipse, which are actually much more purposeful in each of their individual efforts and they do a synchronized release once a year.
Then there’s the Linux Foundation, which feels to me like perhaps [most like OpenStack] where you’ve got a funded core that helps make sure you’ve got some persistent developers and architects and things like that, and you have this ecosystem of companies, large and small, kind of orbiting around that, providing services and that sort of thing. And it’s that neutral core helps preserve the kind of longevity of the project, the stewardship of the brand, that sort of thing. There’s a huge role for that. From what I see, OpenStack’s doing that incredibly well.
There’s one point of contention: how important is interoperability and compatibility. Folks take different stances on this. The Linux Foundation folks told me not to worry about that. They said that they were simply there to promote Linux. It’s a hands-off stance. But from your perspective, how many times have you had to deal with the interop bugaboo?
There’s kind of a resonance there between open standards and open source, right? At least that’s the way I view it. The Internet Engineering Task Force has existed as long as the Internet itself has in different pieces, and every successful open standard has had an open source reference implementation that was also production quality. So SMTP and Sendmail… DNS and BIND… HTTP and Apache, right?
When you say there is a need for a standard in this space, having a good open source implementation of that is important. But implementing every standard doesn’t have to be the end goal of any software project. It should do that so far as it gets it further on its mission, right? That allows it to bring more users in, allows it to live up to being able to migrate people off an old system to a new system, or be a good citizen with other software in the same ecosystem.
The goal of any one project shouldn’t be just interoperability at the end of the day. That should be a means to an end.
I’m new at being in a board of directors for an open source software foundation. Give me one piece of advice for what I should do to help the community succeed?
Be transparent. Tweet a lot. Post a lot. Be very clear when you’re speaking as a leader of your company versus speaking for the foundation. There’s always a natural tension there, and the only way to answer that is to be transparent and honest and say, “we all want to build a bigger ecosystem.” The companies have a strong role to play in that. So, just be upfront and transparent.
Awesome. Thank you, Brian.