A lot has changed in the meantime, including Gopal’s job. IBM lost some key names in software-defined networking (SDN) last year, and Gopal followed them in October, joining Ericsson as vice president of strategy and technology. He’ll be in charge of exploring new areas for Ericsson, which is strong in wireless networks and wants to parlay that strength into the cloud.
Gopal expects to spend a lot of time in Sweden and Silicon Valley, which respectively represent Ericsson’s home base and its primary outlet for SDN and the cloud. Whether he saves any time by living in New York, kind of the geographical average of the two places, is an exercise left to the reader. We caught him during a recent Silicon Valley trip to reflect on OpenDaylight — and, yeah, to talk about his new job too.
Have they sent you to Sweden yet?
Gopal: I did go there, just before the holidays. I realized the scale of Ericsson — Ericsson’s enormous. Learning how the organization works is something that’s going to take a little time.
You mentioned you had other opportunities to leave IBM. What made this job stand out?
Gopal: This role gave me the scope. It wasn’t, “Come in and you’re going to have to deliver this much revenue,” or “Fix these kinds of issues with the portfolio.” It was open-ended and broad in scope, and that really interested me.
Everyone is connected all the time. There’s an opportunity to complement that with the intersection of data center and cloud and IP. It’s an opportunity to create some significant and hopefully profitable new initiatives.
Are you still involved with OpenDaylight?
Not in a direct way. I resigned my chair when I left IBM, which I was obliged to do. I’ve worked with the Ericsson team that’s working on OpenDaylight, and I still talk to my colleagues, of course. But I’m not directly engaged any more.
What did you learn from your time with OpenDaylight?
Gopal: I learned a lot about how to create an open source initiative. It’s very hard. When we started it, guys like [Linux Foundation Executive Director Jim Zemlin] understood what was involved, but frankly, a lot of us didn’t. We understood what was involved in SDN and the technology, but not in building an open source community.
The entire group has gone through a long process around building a community. Going in, there were a lot of big companies, and a lot of reaction from the press was that it’s going to be dominated by big companies. [Cisco and IBM, specifically.] Slowing down the project so we could get more of a grass-roots effort has really paid off. Others have stepped up, like Ericsson and Brocade. There’s a whole developer community that’s started to build up around it, and creating that took a lot of outreach.
How would you assess OpenDaylight’s progress?
Gopal: April 2013 was when we launched and when we said it’s going to be two years before you can judge it. We’re going to be judged on three different things. One is code, the second thing is community, and the third is user acceptance.
When we launched the Hydrogen release, we publicly started [to have] a critical mass in code. The code base isn’t mature, but we crossed the threshold. On community — by the latter part of 2014, we got to a point where there was a healthy and strong developer community. But I think the commercial use of the product or of the assets within OpenDaylight — that’s sort of still happening.
Vendors are building products using OpenDaylight, but in terms of deployment, that hasn’t happened at scale yet. My guess is, by the middle of this year, which is sort of the deadline we’d set for ourselves, you’re going to see more of that.
What’s your opinion of OpenStack?
Gopal: There’s been significant investment on the part of the vendors to make it consumable. Same thing with OpenDaylight. It’ll get there, but it’s not there. It’s not simply code you can use. So, vendors like IBM, Ericsson, Red Hat, and Mirantis will be delivering something that’s a supported product.
I still wonder if OpenDaylight will try to encompass too many things, if it will keep spreading out like a liquid.
Gopal: What happens with these platform-type projects is, everybody starts throwing everything into it. You need filters to make sure everything’s coherent.
The way to think about these projects, OpenDaylight in particular, is that it isn’t necessarily a platform. It’s a collection of components. You don’t necessarily take the whole thing and use it. You build something based on a subset of it. If you think of it as a collection of open source componentry — so you don’t have an OpenDaylight controller, you have an OpenDaylight-based product — then it doesn’t matter if it spreads like water.
If you think you want an OpenDaylight distro [a ready-to-use distribution provided by a vendor] then it does matter. OpenDaylight will have to decide: Do they want to be this repository of components, which is very useful, or do they want to be something more like Linux, which is a distro?
I know your job at Ericsson hasn’t hit full swing yet, but — any inklings of what you might end up doing?
Gopal: Cloud, clearly — understanding how Ericsson can address cloud service providers, which includes private cloud and various kinds of public cloud operators. And the infrastructure within the data center, which includes IP networks and SDN, which might potentially manage or control those IP networks. Ericsson already plays in these areas, but what I’m doing is looking at perhaps integrating some of those things.