AT&T: SW-Defined Access Gets Real in 2018

Software-defined access will come of age in 2018, ready to support multiple media -- fiber, copper and wireless -- and the full range of residential, business and other services, an AT&T executive is predicting.

Fresh off announcing a successful 10-Gig XGS-PON virtualized network trial in Atlanta and Dallas, Eddy Barker, assistant vice president of Access Architecture and Design, tells Broadband World News that AT&T Inc. (NYSE: T) expects to develop a scalable, open-source solution ready for commercial deployment. That latter step will likely happen in volume in 2019.

Key to that major step forward is the progress made in the XGS-PON trial, including creating a software ecosystem robust enough for deployment, that ties together multiple open-source efforts from Open Network Automation Platform (ONAP) at the orchestration and management layer, the Open Networking Foundation and its CORD project and hardware from the Open Compute Project.

The centerpiece of the XGS-PON trial was the Open Source Access Manager Hardware Abstraction (OSAM-HA) software, the renamed version of Virtual Optical Line Termination Hardware Abstraction (VOLTHA), which AT&T developed and contributed to ONF. The renaming is not incidental, Barker says.

"VOLTHA is more of a generic name and it mostly applied to a single technology (PON) and we wanted it to be more open for multiple types of access -- virtual RAN, PON and G.fast," Barker says. "The other aspect of it is, this year is really all about getting the software ecosystem that we have built to be robust enough to go to deployment."

That involves tying ONAP down into the virtual applications running in ONF, which was one of the major existing gaps, he adds. OSAM will now be an open source operational suite for managing consumer and business broadband service elements -- once controlled by vendor-specific element management systems -- and will be incorporated into ONAP's Beijing release, due out in late May.

"We are trying to bring together different open source communities," Barker says. Many are now under the Linux Foundation auspices, he notes, but there is still a need for coordination among their efforts and that is one thing AT&T is hoping to aid.

"With the trial, we were trying to take white box designs from OCP, tie that with the latest releases from ONF for PON physical and virtual functions with ONAP open source code, spinning up VMs through controllers," Barker says. "On top of all that, we are also using some other virtual functions like a virtual broadband network gateway, which used to run on high-end routers, for layer 3 gateways, and running gigabit and multi-gigabit services for residential and business" markets.

Then there was the need to prove that GPON and XGS-PON could coexist on the same fiber infrastructure, as intended.

One of the hardest things about the trial was selecting vendor participants, as AT&T couldn't accommodate everyone. That is a key reason why the operator isn't identifying who was involved in the trial, he explains. Barker says the trial did point up operational gaps, but that was an intended outcome, so AT&T can begin to address those.

"There are still some gaps in the software, around operational management things," he says. "That was a big focus as we work with operations to do the trial, that helps them figure out what does it not have that I really need. It brought all of that to the surface. That is why in 2018, we have to get out of proofs-of-concept and into something we can deploy, so we will focus on the gaps and commercialize and target getting it ready to go prime-time out in the field. That is the real test this year."

— Carol Wilson, Editor-at-Large, Light Reading