SANTA CLARA, Calif. -- Open Networking Summit --
I'd been thinking of OpenDaylight as a way to wrest the SDN conversation out of startups' hands.
"The moment I talk about there's IBM, Cisco, Juniper and all these large players, the question is: What's the angle? ... There's got to be some underhanded, nefarious reason," said OpenDaylight chairman Inder Gopal during a keynote here Tuesday morning. "There is no angle, honestly."
I'm still not fully convinced, and Gopal, who's also vice president of system technology at IBM Corp., did concede that individual companies might have their own hopes for OpenDaylight.
You can see where the speculation came from. It's very hard not to see the "Daylight" name as a foil for Floodlight, the open-source OpenFlow controller that Big Switch Networks developed. Others in the industry see it as a banding-together against VMware Inc., which could combine its dominance in server virtualization with the Nicira acquisition to create something pretty powerful.
As the effort started to snowball, VMware got on board, as did startups -- Big Switch is contributing OpenFlow controller code -- making OpenDaylight closer to a pan-industry effort.
From that perspective, I can believe Gopal when he says OpenDaylight hopes to speed the adoption of SDN. That starts from the vendor side, by creating a common base of software for vendors to work from, Gopal said. The grunt work of undifferentiated code will be taken care of.
One hitch: I don't think people two years ago were thinking of the SDN controller as "undifferentiated code." A lot of other features startups were hoping to mine could get absorbed into that category as well. As one executive outside OpenDaylight was telling me, the coalition could unintentionally smash a few startups regardless of what it wants to do.
A more positive side effect would be interoperability, theoretically, because of that common code base. And all of OpenDaylight's choices will be based on merit rather than politics, he pledged.
"It's not about standards," Gopal said. "The objective here is to create code and to do it at a commercial, carrier-grade level ... something that, hopefully, as-is, can be taken and deployed."
The OpenDaylight steering committee is sifting through the trove of contributed code, deciding which pieces would create the best-of-breed framework that the group envisions. OpenDaylight isn't pledging to have that done before the third quarter.
Wait two years before grading OpenDaylight, Gopal told the audience. Sure. But OpenDaylight will be getting judged continually along the way, and people will be poring for angles in that first release of software.