Disruptive technologies meet staid businesses. Stuart Cohen is bringing the open source development model together with big business, and finding it to be a perfect fit. Joe Brockmeier talks to Cohen about the Collaborative Software Initiative's first year, and where it's going from here.
Some ideas seem obvious, as soon as someone else has them. Stuart Cohen, formerly the CEO of the Open Source Development Labs (now the Linux Foundation) is the man behind the Collaborative Software Initiative (CSI), a company that’s bringing the open source development model to bear on the problems of big IT.
Specifically, CSI is helping companies build essential but non-competitive software that will eventually be released under open source licenses, as well as building communities around that software. Cohen announced CSI in April of 2007. With nearly a year under his belt with CSI as its CEO, we spent some time checking in with Cohen to see how the venture is going.
Linux Magazine: First time we spoke, I believe you were associated with OSDL, want to talk about your background with open source in general and how you got started with OSDL?
Stuart Cohen: I got started with OSDL, kind of an interesting relationship, I was with a company called RadiSis, an embedded hardware company in Hillsboro Oregon, supplier of embedded hardware for the telecommunications equipment companies, big customers were companies like Nokia and Alcatel and NEC. We got to doing some work with Alcatel and Siemens… and in both cases they asked about what we were doing with Carrier Grade Linux, and the IBM guys and the HP guys all talked about what they were doing with Carrier Grade Linux… this was going back six years ago, seven years ago…
We talked about it a little bit and literally a week later, I get a phone call from a headhunter asking if I’d be interested in being the CEO of OSDL. And, so it’s kind of interesting, we had these conversations with Siemens, Alcatel, IBM and a week later I get a call from a headhunter. People say that they’re unrelated, but I guess the truth will never be known…
I spent a little over four years at the Open Source Development Lab, had a great time with it. Obviously we grew that organization and its mission, very broadly and that’s what led to the Collaborative Software Initiative as much as anything.
Linux Magazine: And how did OSDL lead to the Collaborative Software Initiative?
Stuart Cohen: When I got to OSDL, there was a small number of vendors that were members, there were about 20 members. There were the eight big guys, and a few smaller companies, that made up the group. One thing the board wanted to do was to expand the number of members of OSDL. We always had an office in Tokyo, and we had and office in Oregon where the developers were.
We did a number of things, we went on a big recruiting mission to recruit a number of members to participate in what we were doing in the workgroups and expanded the number of workgroups we were doing.
Now, with the SCO activities at the time, in 2003 and 2004, a lot of the big end user companies that were using Linux in a big way, as customers if you will, were not terribly interested in getting involved in using Linux in a very visible way. They were worried about whether they were going to sue people and so forth. So we formed these customer advisory councils, one in the U.S., one in Europe, and one in Japan, and they moved up to about 60 companies between the three, and they gave us kind of advice, input and thoughts on what was needed as an operating system standpoint and it was a good way for us to feed information and have dialogs and discussions with kernel developers about what customers were needing on a large scale.
One of the things now about two years ago, some of those customers were big banks and insurance companies, mostly located in the New York area, they came to us and said that this model works great with the developers to develop this operating system, and it’s been a great benefit for everybody — you know the story, I mean, lower costs, greater productivity, better flexibility on hardware platforms, et cetera, et cetera.
So they said, can we apply this same model to some software areas where we really need to look at, and they were specifically around some compliance and regulatory issues that they were dealing with as a group, if you will. We said, will you tell us a bit more about it, and they said [new laws] have created an environment where we all have to assess our vendors that have access to confidential information through, whether it’s outsourcing agreements or the like.
So they said, “is there a way to work together with you like the vendors work with you and we pay you pennies on the dollar and you develop software we can all use?” One thing led to the other, they had two requirements, one we all want to develop this together and it has to be easily deployed in our environments — and some of them are Windows, some are Linux, some are HP-UX, AIX.. and you need to provide commercial support for it…
So that’s really how it got started. We talked to the folks at OSDL, the board specifically, and I was on the board at that time, and talked about whether we wanted to focus on the OS and a set of these applications, because this is a new area for open source to go, right? Because, open source had great success in the OS and great success on the infrastructure, and now these companies were looking at very specific applications where they’d like to form a community to form a project to do development, but there aren’t open source developers, on their own, going to focus on compliance and regulatory issues.
So it very much came out of the customer advisory council, not only from the financial services industry, but from others as well.
Linux Magazine: Let me circle back, I want to ask you about something you said there. Having been involved with a lot of customers, not just from a vendor standpoint, and having the benefit of several years of hindsight. How would you characterize the impact of the SCO lawsuit now? I’ve heard people suggest that in the long run it made Linux stronger because it was tested and they weren’t successful, what are your thoughts there?
Stuart Cohen: The first 90 days were a bit of “what does this mean?” But once you get through the “what’s this mean” and “what’s this all about?”, it’s been a huge acceleration of the adoption of Linux. It forced everybody to take a look at not only the operating system and the code, but the model. Right? By that, you know, that light getting shined on it and very aggressively from every angle, whether it’s a technical angle, a business angle, a legal angle, a support angle… whatever it was, the operating system came through with flying colors. I think it was a huge catalyst for success for the operating system… I think it’s led to a real rapid deployment of open source software, not only within the server market, but the desktop market, the phone market, the supercomputer market, the appliance market, and then out into the infrastructure and now into applications.
Because it’s the model, the model of open collaboration is what’s really driving it. I think it’s been a huge, huge success. I think in a lot of ways that SCO was what led to that.
Linux Magazine: If you had to explain CSI to somebody in the proverbial “elevator speech,” how would you explain it?
Stuart Cohen: It is an opportunity for a number of companies to come together for basically pennies on the dollar to have a project developed and then delivered back to them that runs easily in their environments, whatever environment it might be, and is either available as a community built open source software, or as commercially supported software.
And then, ideally, that software then gets rolled out to their extended, whether it’s supply chain, partners, health organizations, banks, vendors, where there’s a tail for longer term support opportunities.
Linux Magazine: Now, obviously, you are focusing on vertical apps and industries like banking, you’re not focusing on the desktop, and consumer type software. Why not go after the larger market that’s out there?
Stuart Cohen: Well, a couple of reasons. One, the model of whether it’s an industry standard or it’s a government requirement that requires companies to come together to work on a certain problem they have to solve, is very interesting and appealing to us. It also works out economically. If it costs a million, million and a half to do it, and risk mitigation and leverage for doing it together, it accelerates community developed applications.
Second of all… they [desktop apps] are areas where a lot of different companies are all trying to do it themselves. And for us to try and bring a group of people together to work on something where they think they can do it themselves, they can do it individually, flies in the face of what we’re interested in.
Linux Magazine: So, this model isn’t what we can count on to deliver desktop apps that a lot of people are looking for, then?
Stuart Cohen: Not when you think of very, very horizontal consumer apps, because I don’t know who would want to come together on that. Let me just expand on that for a second. If it’s a number of, hospitals that want to work together on a certain problem that runs on the desktop, or a number of states that want to work on a certain department of transportation problem, or a number of banks that have the same regulatory problem, or same insurance companies… we will then work on those, and they will run “on the desktop,” in a Windows environment or Linux environment.
We are working on desktop applications. The first two projects we’re working on, have a PC/laptop version and server version. We’re developing apps that are run in that way, but not normally in the way you think of consumer apps.
Linux Magazine: Tell me a little bit about the first year of CSI. What’s happened since the inception of CSI? How do you feel it’s going?
Stuart Cohen: I’m very pleased at the start we’ve had. We’re just about to have an ’08 kickoff meeting and planning session we’re going to have in a couple of weeks that we’re going to have with everyone. Our intention was to start the first year, to get two projects underway. One that you’ve seen us announce in the banking industry with the shared assessment program as part of the BITS trade association and a number of the big Wall Street banks, coming out of the customer advisory council I told you about.
We also have a second project underway, haven’t announced it yet, not in financial industry, still around compliance and regulatory issues. Think about a government-regulated environment where all the states have to meet and be in compliance with that. A number of states and a government industry want to come together and collaborate on that, similar to the way that banks want to come together and collaborate on outsourcing and banking activity.
So, from a project standpoint, I feel good about where we are, we’ve hired some great people to work on those projects as developers and as program managers…
We met the revenue milestones that we put in place from the beginning of the year, so we feel good about where we are from a revenue standpoint, we obviously raised $1 million as seed money… we haven’t spent all that money, we’ve used that money wisely, and now we’ve grown additional revenue on top of that. We feel good about that, we have some big plans for ’08 on the development and support side. We are going to release software, software as a service, and we may do a an appliance as well. Working with one of the big hardware manufacturers on an appliance as well where we’ll preload it … and ship a true hardware appliance. We feel good about that, we’ve got some goals about the number of projects we want to put in place, the amount of developers, the amount of revenue, etc… we think we’re off to a good start.
Linux Magazine: Have you announced what license you’re going to be using for this project when it’s released?
Stuart Cohen: No, but we anticipate using the GPLv3 license.
Linux Magazine: Do you want to talk about why that license is appealing?
Stuart Cohen: We like the credit it gives us as the originator, and we like the flexibility it gives the users. Obviously the giveback capabilities, that they have to contribute that back to the general project, we’re very pleased with it [the GPLv3].
Linux Magazine: As an outside observer, how do you feel about the progression of OSDL since you’ve left?
Stuart Cohen: You know, I think they’ve done a good job. Obviously, as I said, we had this idea that we were going to do not only the operating system but look at applications at the same time, but for a variety of reasons we decided that the Linux Foundation would become OS-focused and operating system only. I think it’s worked out great. The safe haven for kernel developers works out well. It’s worked out well for what they’ve set out to accomplished, and it’s good for Linus, then it’s good for the developers…
I’ll also tell you on our advisory council are IBM, HP, and Novell, as well as Eben Moglen and Brian Behlendorf, I feel like I got the best of both worlds. Linus and I still live a quarter mile away from each other, and still see each other. I think it’s working out well for him and what he’s trying to do, I feel really good about it.
Linux Magazine: If you talk to people and read the financial news right now, there’s a lot of gloom and doom, [thoughts that] we’re heading into a recession in the United States… do you think that’s good or bad for you and the CSI?
Stuart Cohen: Personally, I think the time is great for open source. It’s proven to be a much lower cost implementation, you’re not paying the big upfront licensing prices. It’s proven to be lower development costs, because you’re sharing the costs with others… and people have seen the efficiencies, right?
Whether it’s been Linux, Apache, Firefox, whatever it might be, they’re seeing great benefits from open source software. I think that, as IT budgets get squeezed and dollars get squeezed, people are looking for more and more cost savings, but they still want high quality code, they want high quality support, the collaborative model, the community model is great for that. I think it only helps.