Cloud Native ComputingDevelopersFeaturedOpen SourceTFIR InsightsVideo

Open Source Is About People & Collaboration Not Licenses – Rob Hirshfeld


In his recent column, open-source expert Matt Asay wrote how RackN took the conventional open core model and turned it into an ecosystem centric model. I sat down with Rob Hirschfeld, Founder, and CEO of RackN to talk about not only RackN’s model but also how open source is changing. Contrary to a popular belief, he says, open source is not about the license at all, it’s about people and collaboration. I hope you will enjoy our discussion.

Here is the rush transcript of the interview:

Swapnil Bhartiya: So let’s talk about Digital Rebar, Rob. This was the project that RackN is sponsoring and the foundation of RackN so talk about the evolution of the project because you took an Open Source project, and based on the community feedback and commercial interests, you made a lot of changes to it. So, talk about the history and evolution of Digital Rebar.

Rob Hirschfeld: Digital Rebar literally goes back 10 years to our very early days with the Crowbar Project. That was generation one of what Digital Rebar is today, which is a fourth generation. Probably, it doesn’t have a single line of code. It’s a completely different language, but there’s this evolution where we started with what we’d written with Crowbar, which was an open core project, because we had to be able to get it outside of Dell and be part of the OpenStack community.

So we were steeped in this very open code, leave the code out there. Open-core, by the way, is a model that says, let the core of the code be open and then monetize all these pieces around it. I have come to see it as freemium software, where you’re like, “Oh yeah, use the free part of our software all you want, but as soon as you cross into this threshold, you’re going to be into pay software.”

It’s a well-established commercial model, like many. That’s not about Open Source or closed source. It’s about, “I need a product that lets somebody try it and use it.” But then all of a sudden with open-core, there’s a ton of these buried tripwires, where it’s like, “Oh, you wanted to be able to do that. Yeah, no, that’s going to cost you money.”

It can create very unnatural acts in the open-core communities because that means you’re withholding really essential pieces of the technology from the open. We’re seeing this a little bit with the HashiCorp tools, where they open the core of the tools but control it with a single vendor. Then if somebody says, “Well, I want something in the core,” that they are then monetizing, then it creates this fight between the community of people who are using the product and want to add features and the rest of the community.

Our experience had been… We’d actually been working on Digital Rebar as an open-core project. We got it to a point where we had people spinning up thousands of servers and data centers and building businesses around the core product, which you would think would be a fantastic success. But they weren’t turning to us and then paying for support or using the commercial enterprise or even collaborating us. They were just taking the core software and not even saying thank you, just taking the core software and using it.

That was never our objective, from that perspective. Our objective has always been that we would collaborate with our operators. We were building software that had to be part of this shared experience. That’s what operations is. When we look at how we want to license the product, it really comes back to shared experience.

Swapnil Bhartiya: When we look at any Open Source project, Linux is a very good example or Kubernetes is a very good example. There is a core of communities. There is a core of the external. Now, there is a lot of contribution coming to the core, but if you look at the whole businesses that is built around it, it’s not built around that foundation. It’s built around what people are building on top of that, and that’s where most… You are the user. If you look at it, your distribution, the contribution is not going to the Linux Kernel. They are working on GNOME. They are working on some audio. They are working on networking stack, not to the core. So similarly, when you look at a project like Digital Rebar, because you have better perspective than anyone else from the outside, what kind of contribution were coming to the product that actually kind of is encouraged. What inspired you or drove you to adopt this model?

Rob Hirschfeld: You’re exactly right. I do think we have a tendency to use Open Source as a marketing term. Not we, the industry we. RackN actually has come a long way with this. And we’ve pulled back from using open as a marketing value in itself. It really is about collaboration and our focus is on collaborating with vendors, partners, and with our community. What you asked was “How do we know that we had it wrong?” And what we knew is our partners, customers, and prospects wanted to see the code that we were not sharing. They said, “I want to see the IPMI drivers. I want to see Radeon Bios configuration. I want to see how you integrated with Terraform. How do you connect the dots into my Switch or Ansible?”

Those were, operating system integrations. Those were absolutely essential for our operators to collaborate with us on or our partners to collaborate with us on. And what we’d done because of the model was all of that we had to guard and, nobody ever said, “Oh, I really want to understand how you store JSON objects behind the API.” That was not a place where somebody wanted to collaborate. They were like “Bracken, you better make sure that works really, really well. It performs really well.” And we have to invest a significant amount of time and engineering expertise and ensuring that we have those capabilities, but nobody wants to collaborate with us on doing it because it only benefits Digital Rebar. If we improve our IPMI drivers or improve our HP, Dell, Lenovo, Supermicro support, our whole community benefits from that.

And if it needs to be adjusted or changed, then the person who has that model, that we didn’t support, they have immediate benefit for adding it. A lot of times they can do it very quickly themselves. And then the whole community gets the benefit on the other side for that and it keeps RackN frankly, from having to buy every model of possible gear, which is not feasible under any circumstances. And that’s where we really look at. When we looked back at the code contributions that we’d had over a eight year history, only our team had ever been contributing to the core of the product. It’s just very, very, very clear data from that. And we had a lot of requests for people who want to jump in and see how we’ve done the integrations into these different vendors’ infrastructure.

Swapnil Bhartiya: True. And if you look at any industry, look at farming, chips, fries, we do need farmers who are green potatoes, but you don’t see in this input. Other people are investing in transportation, restaurant. Same if you look at automobiles, nobody is going inside the engine, people are building accessories for cars. Any industry, or Apple, Windows, whatever it is, nobody actually worked in Core. I did want to talk about this because sometime people confuse that core. Yes, it is critical. It is important. It has to be maintained, but maintained by the stakeholders. I like what Greg, creator of Linux Rose, say that we want selfish user. We don’t want users who are contributing to a project out of charity. We also need people who have a vested interest, selfish interest in contributed code, but they will not be thousands of people. There will be five or six core players, but the community is going to be much bigger than that.
Earlier, when I was talking about licensing is at the heart and soul of Open Source, you said, you kind of disagree with that. So I want to touch upon that also because you’re not talking about how you double up the code. You’re talking about how you license that code, because that dictates how you’re developing the code. So, I do want to talk about that a bit. Why do you disagree that licensing is not the heart and soul of Open Source?

Rob Hirschfeld: It’s not. People and collaboration is the heart and soul of Open Source. That is it, just period, full stop. The purpose of Open Source software is to allow people to collaborate. The licenses can be distractions, from that perspective. The question becomes, “does the license match the way people want you to collaborate in the product?” It’s that simple. So if you have a license that discourages people from being able to collaborate with you, then you’ve created a real problem with this. If you have, a license that encourages people to take the software without participating, then you have a real problem with this. And that fundamentally is what you need to look at. You have to find a balance and it’s not just license. It’s how you deliver the software. It’s how your commercials are structured, that says, I want people to interact with me to become part of that community.

You mentioned this farming analogy, which I think is a great analogy for us to talk about. It’s about supply chains, right? So yes, you do want to just consume the fries. There was a really good example, not long ago, where people were talking about Platform as a service infrastructure, as a service and do it yourself. And they were talking about pizza shops and how much of the pizza you had to make or build or deliver. But those analogies are a little bit flawed. This is your point about the french fries. If you’re consuming the french fries, you actually need to make sure that the transit systems, and we’re learning this with the pandemic, you better make sure that the delivery systems are good, that the workers who are cooking the fries are good, that the healthcare for the workers who are delivering the fries are good, that the farmers are good, that all of those pieces and parts are interconnected as part of the system.

And so the idea that we have an Open Source software ecosystem in which people can just say, “I’m not worried about how that software is being maintained, or how the people who are writing that software are doing it.” We need to not think about this. This is a flag I’ve been raising for a couple of years. It’s a whole model. And so the idea that you’re using software, that’s being supported by charity, that was your word. That’s not a sustainable model for you to build anything on top of. Feeding equipment, commercial, whatever your app is. We need to make sure those things are fed. I’ve heard even the same thing is true with journalism and newspapers and things like that. We need to think through where the money is coming from for that.

And Open Source software is no different. For RackN, when we made this change, people were like, “Oh, but now you’re not going to have this huge community of people just downloading and using your software for free.” And that’s the Open Source model, give it away for free and you’ll make it up in the profit, in the margin, the volume. And when we looked at what it takes to actually support a running data center and the amount of operational interactions and collaboration that are necessary for us to do our job well, we needed a licensing model that actually encouraged our communities to participate with us and had high expectations for what we delivered and had a way to participate in the places where they had a lot of value to add.

And so the open ecosystem was a much better model for us. And like you said, this is not new. Amazon, Google, Microsoft, all the major clouds, VMware, Oracle they’re all ecosystem plays. They don’t monetize people using their product and adding extensions to those products. That’s a wonderful economy around those core products. And then that works because those products are rock solid. It wouldn’t work if Amazon’s cloud was a painful, bad, unstable cloud. It works because Amazon’s core infrastructure is rock solid. And then they’ve left room for people to innovate around them.