Sergey Nivens - Fotolia
Forward-thinking enterprise IT shops like their software open source and their vendors in it for the long haul, but in some cases they can't have both.
Docker data storage software is one example of a market fraught with open source business dilemmas. SearchITOperations recently sat down for a post-mortem conversation with the man at the helm of one early casualty in this market, ClusterHQ. The company ceased operations mid-December, and while its former CEO Mark Davis authored a blog post that broadcast the company's closure, he offered few details about what led to its demise, until now.
Davis: We did have good traction on usage of our software, and I think we had built a nice and strategic role in the Docker ecosystem. But it was very difficult to visualize a profitable business model over time. Not impossible, but it took a fair amount of squinting off into the distance to see when we might be able to do something that was going to be a decent money-making business.
That's one of the essential challenges that everyone's having in containers right now. There's tons of adoption. The adoption is real. But the money isn't there yet. And it may get there, but finding a business model and having deep-pocketed investors willing to be very patient about when that time will come is the challenge.
We basically were out of money, but I had arranged another funding round that was within days of closing. We could have closed, but I looked at myself in the mirror and said, 'If I take this additional money, do I have confidence that we can turn this into a profitable business on this money?' There was a chance. I wouldn't have gone through the motions with investors if I didn't believe it, but I had to take a really hard look and decide, is it really the right thing to do to take additional capital? Essentially what we decided was [that] it was too risky to ask investors to keep putting money in, to ask employees to keep on putting in hours, for an outcome that was going to be really tough.
Is that a fundamental issue with open source in general?
Davis: It's awesome for users, but I think there are very few vendors on the planet who would say they love open source, and even open source companies would probably prefer not to be, but it's the cost of doing business, particularly in infrastructure software these days. It's really, really hard to be an infrastructure software company and have big companies talk to you -- the big Wall Street banks, they won't even talk to you if you're not open source anymore. So you have to play that game, but still you've got to find a way to make money at it.
I don't think the decision that we made indicts the whole industry. There's a lot of momentum around Docker itself; Kubernetes is going crazy right now. There's definitely lots and lots of uptake. But I still think it's going to be tough for people to make significant money with open source software, because the difference between usage and revenue right now for everybody is a huge gap.
Can you say how much usage you had?
Davis: It's hard to say exactly, which is another one of those problems with open source. We had more than 100,000 downloads of [Flocker Docker data storage management software], a couple thousand per week. That's a lot for infrastructure software. It was good usage.
How many of those people were paying?
Davis: Very few. I shouldn't say any more, but it was very few.
So it's not that people weren't interested in the technology.
Davis: The Flocker [data storage management] product is useful, but it is a product that in the best-case scenario was never going to generate much revenue, because what it does will be commoditized by the orchestration systems over time. It largely already is obsoleted by the latest versions of Kubernetes, somewhat with Mesosphere and a little less so with Docker Swarm. There was a different strategy behind the products we launched more recently, but it would have taken a long time to monetize.
So those companies like Google and Red Hat just have the critical mass to sink time and revenue into that orchestration?
Davis: I think they do. Red Hat is certainly investing heavily in containers in a big way with OpenShift and they're very allied with the Kubernetes front. How much revenue they're making, I have no idea. But Google actually has no intention whatsoever of making money on Kubernetes. For them the broad usage of Kubernetes is not about selling Kubernetes or Kubernetes services, it's about getting people to use, in the long run, lots of Google Compute Engine, lots of Google cloud services. In some ways they're the perfect vendor because they don't need to make money on Kubernetes.
Red Hat does need to make money at it, but Red Hat is a company with a very long history of knowing how to patiently build a market on open source. They don't have just OpenShift to sell, they have a large portfolio of products, so as an enterprise software provider they potentially have a great value proposition to enterprises who want to use open source, containers and Kubernetes, and want a vendor they're probably already working with. I would assume they have a great shot at building a great business.
What about Docker?
Davis: Docker has huge momentum, enormous brand recognition, a lot of venture money raised and that gives them some time and some freedom. But I do think they've got an extra challenge with monetization, because the container runtime and a lot of the lower-level container features are clearly not going to be monetizable. But the Docker people are very bright and they've got not only good technologists but good people on the monetization side and I think they'll figure it out. It looks to me like they've done a good job packaging some proprietary products around their open-source platform. I don't know how much revenue they're making, but they raised a huge amount of money early and that gives them time to develop the market.
How is the containerization of IT changing storage, and vice versa? While ClusterHQ closed its doors, the Docker data storage ecosystem has plenty of life.