Networking professionals gathering in Santa Clara, Calif., for the second OpenDaylight Summit July 27 to 31 will get a look at how far the software-defined networking (SDN) platform has grown and matured—and how far it has yet to go.
In terms of growth, OpenDaylight’s third release, Lithium, which came out last month, is comprised of 42 projects, about twice as many as the Helium release and up from 13 in the original Hydrogen release.
Projects are the modules that comprise the OpenDaylight platform and include things like the SDN controller itself, YANG tools for modeling and configuration, the OpenFlow plug-in, and a variety of other interfaces for communicating with the network hardware and orchestration functions.
The number of developers contributing to the project is now more than 400, up from less than 300 for Helium, who have written about 2.2 million lines of code.
As part of the summit, there will be a developers’ conference for the fourth release, dubbed Beryllium (OpenDaylight names releases according to the periodic table of chemical elements). Beryllium, due Feb. 4, 2016, will balloon to more than 60 projects, said Phil Robb, senior director of technical operations for the OpenDaylight project.
One emphasis for Beryllium will be clustering, Robb said. Developers will discuss clustering as part of the core controller, what applications must do to be cluster aware in OpenDaylight and cluster enhancements to the OpenFlow module, he said.
The Beryllium release process will begin its cycle this week, and a sign of the rising maturity of the platform is that OpenDaylight officials will implement a new and more rigorous workflow around individual project lifecycles, Robb said.
That’s not to say OpenDaylight is not stable enough for production deployments. It’s just that current deployments, mostly still based on the Helium release, have less functionality than what is available now in Lithium and eventually Beryllium.
While OpenDaylight was not ready for production deployments a year ago, this week at the Summit, about 20 companies will discuss implementations, notably from AT&T, which is making OpenDaylight part of its Domain 2 initiative; CableLabs, which is using the platform for managing cable and set-top boxes; and Brocade, which bundles an implementation of OpenDaylight with some of its switches.
Despite the growth and maturity, the successes so far are all relative. Only 18 months have passed since the initial Hydrogen release, and like their brethren from the OpenStack world, vendors are working on a version or two behind the current release.
And in these still early days of OpenDaylight, the project finds itself competing with an alphabet soup of standards bodies and release code names—not exactly competing in the marketplace, per se, but for the hearts, minds and business plans of equipment makers and, eventually, customers.
In the past two months, two new SDN frameworks came out: the Open Platform for NFV’s “Arno” framework for carrier-grade network functions virtualization; and the Open Networking Foundation’s (ONF) “Atrium,” an SDN distribution that is supposed to supply the means for connecting or integrating software-defined networks.
OpenDaylight Summit Will Highlight Progress on SDN Projects
A fourth player, Open Network Operating System (ONOS), an SDN controller that is the closest to OpenDaylight in functionality, will release its fourth iteration, “Drake” (named after the bird), at the end of August.
The ONF, which shepherds the OpenFlow protocol that gave rise to SDN, is trying to be a neutral party that attempts to keep everybody working together as best as possible, according networking specialists involved in the process.
“Atrium is the ONF’s attempt to broaden out OpenFlow, very complementary to OpenDaylight and ONOS,” said Shamus McGillicuddy, senior analyst for network management at Enterprise Management Associates. “It’s showing how you can tie together open-source software into a specific stack that services a use case.”
ONOS and OpenDaylight will eventually be viable SDN components, but will settle into specific environments and use cases. Whereas ONOS is targeting service providers, OpenDaylight will likely find a home in the enterprise data center.
OpenDaylight has an architecture “that is certainly broader, and we are looking to cover many more use cases than ONOS is,” Robb said. “Given that, it will take longer for us to get there. ONOS will get there faster. We will get there a little bit behind them, but we will be able to do more when we do.”
Regardless of the individual fates of OpenDaylight, ONOS or OPNFV, the future and benefits of SDN are clear: programmability of network functions via software rather than manual hardware configurations.
This will enable less proprietary code residing on expensive, on-premises networking gear, which means the gear itself can be off-the-shelf and remotely controlled via programmable policies, which will make it cheaper for vendors and end customers.
AT&T Chief Financial Officer John Stephens made that point in an earnings call last week. He sees much less capital spending on networking hardware in the enterprise, in the service provider data center and in homes. “I think there [are] real … opportunities that are happening on a software-defined basis to bring investment down,” Stephens said.
Scot Petersen is a technology analyst at Ziff Brothers Investments, a private investment firm. Prior to joining Ziff Brothers, Scot was the editorial director, Business Applications & Architecture, at TechTarget. Before that, he was the director, Editorial Operations, at Ziff Davis Enterprise. While at Ziff Davis Media, he was a writer and editor at eWEEK. No investment advice is offered in his blog. All duties are disclaimed. Scot works for a private investment firm, which may at any time invest in companies whose products are discussed in this blog, and no disclosure of securities transactions will be made.
