Since its inception in 2013, the OpenDaylight Project has been among the earliest proponents of software-defined networking (SDN), and while it enjoyed a banner year in 2016, there are big plans afoot for 2017.
In 2016, OpenDaylight (ODL) saw an influx of high-powered Chinese members, such as Tencent, Alibaba and China Mobile as well the debut of its fifth open SDN release, OpenDaylight Boron. OpenDaylight, which is an open source group in the Linux Foundation
, also hosted its first summit outside of San Jose when it held the event in Seattle this past fall. Prior to the Seattle summit, ODL announced its new "Powered by OpenDaylight" trademark project. (See OpenDaylight Launches Trademark Program and ONOS, OpenDaylight Update Open Source Platforms.)
Telco Transformation recently spoke to Lisa Caywood, director of ecosystem development for OpenDaylight, about last year's milestones and this year's roadmap. In Part I of this Q&A, Caywood spoke about why Boron was important for ODL and its users. In Part II she discusses collaboration among open source groups and standards bodies, and where SDN is headed.
Telco Transformation: Looking back at 2016, what was OpenDaylight's biggest accomplishment?
Lisa Caywood: I have to say that I was away from the OpenDayight community the first six months of last year, but in the second half, one of the biggest accomplishments was the Boron release. The reason that was cool was because it was the release that took the big step forward away from the focus on the basics, and started moving towards integrating the platform into specific use cases and bigger, broader frameworks.
That's really important because the other big development for 2016 was actual large-scale deployments in major organizations around the world. We actually had to limit the number of user keynotes at the summit this past fall. We had more than we could use because there were just so many really fascinating stories of what people were doing with the platform.
I would say that the Boron release was definitely a key part of that transition. It's not that they came on just for Boron, because Boron came out in the fall just before the summit, but many of them had started working directly within the developer community.
They were piloting it and as they got more comfortable with it, as they started working with it, as they started deploying it in production we were starting to see what worked and where we could have more useful functionality. Rather than just saying, "Oh can you guys go do this," they put their own people on it and got involved.
More than half of the new projects in Boron came from user organizations. To me, that was a really critical data point. To me, that was the real mark of this being real, and that it was meaningful to people when they want to put their own resources and mark on the platform.
Want to know more about the companies, people and organizations driving developments in the virtualization sector? Check out Virtuapedia, the most comprehensive online resource covering the virtualization industry.
What has been the feedback on Boron and what's next?
LC: Typically deployment is a release or two behind whatever the current release is. A couple of vendors have brought out things on Boron, but most are on (previous release) Beryllium so that's kind of where most users are at this point.
But Boron is important and those advances are important because they are moving ahead with the things that Boron is focused on. Boron was really focused on integration with OpenStack, support for service function chaining and management planning for programmability. All of that in turn ties in with integration with a higher level framework with things like ECOMP, OPEN-Orchestrator Project (OPEN-O) and OPNFV. Those are all things that have generated a lot of interest this past year independently of OpenDaylight. ECOMP is something that AT&T is very proud of and they've been talking to anyone who will listen about it over the past 19 months. I expect that we'll see that [ECOMP adoption] rippling through other North American carriers as well.
So as that advances, the work that has been done, starting in Boron and flowing through Carbon and future releases, is really focused on making sure that OpenDaylight is an integral component of all of those frameworks, and fully supports what needs to happen in the broader scheme of things.
TT: What's on OpenDaylight's roadmap this year?
LC: If you look at the Wiki page for Carbon, which is our next release in the May timeframe, there's always ongoing work in the core of the platform, the (SDN) controller and so forth. There's a lot of ongoing work with a number of the southbound protocols, BGP-PCEP and things like that. So always ongoing work there, but the new projects are building on the work that was done in Boron. That's more around operational tooling, specifically an analytics engine for monitoring the performance of not just the network but statistics brought back by the controller and making decisions based on that data and the analytics.
There's a second project that's related to machine learning. You can start too see how these things are coming together. You start creating more engines for capturing analytics and then being able to make more predictive type of assessments and using that to drive management plane automation.
TT: What has been the reaction to the "Powered by OpenDaylight" program? How many have earned the seal of approval?
LC: We haven't been pushing it heavily. We announced the program in advance of the trademark being finalized so we've been talking to those that have raised their hands and said, "yes," to them with the caveat that the trademark isn't final. We expect it to be finalized over the next month with the trademark office and then we'll be going full speed ahead, but I think it's about ten-ish right now. The real focus of that program, once the trademark is fully in place, will be coming in the first half of this year.
— Mike Robuck, Editor, Telco Transformation