One of the things I’m most excited about with our transition to a software-centric network is our involvement with the open source community. Open source will speed our innovation, lower costs, and help us build the foundation for our 2020 goal.

One example I’m discussing today at the Open Networking Summit in Silicon Valley is GigaPower. That’s our super-fast gigabit home and enterprise Internet service. Customers love it. But we have to install a lot of complex and expensive equipment in our central offices to deliver it to each neighborhood. Stuff like “gigabit passive optical network open line terminals,” or GPON OLTs, if you’re a glutton for acronyms.

Our engineers have figured out how to turn those complex appliances into software running on commodity servers and other hardware. That’s the basis of what’s called “network function virtualization,” or NFV. But we’re taking NFV a step further.

Enter “disaggregation.”

Disaggregation is a big deal. It means we don’t just clone a hardware device completely in software and continue running it as before. Instead, we break out the different subsystems in each device. We then optimize each of those subsystems. We upgrade some and discard others. That’s what we’re doing with the GPON OLT, as well as other pieces like the Broadband Network Gateway and the Ethernet Aggregation Switch.

And while we’re breaking down those components, we’re also releasing into open source the specifications for the commodity hardware that will run those components. In tech speak, we’re virtualizing the individual line cards in each OLT, and turning them into a single PON MAC (media access control) card.

This vOLT, or virtual OLT, will become open hardware that reduces power consumption, scales faster and costs less. And by making the technical specs available in open source at the Open Compute Project, we’re inviting any white box hardware maker to build and sell them to us and also allowing others to build on the concept and design.

We’re the first telecom company to do this. Being first is never easy. But we hope to see prototype devices by the end of 2015, with trials and deployments in 2016.


One of the tenets of the open source community is that you don’t just take code. You contribute it, as well. This is a big deal for us. Here are some of the ways we’re doing that:

  • A lot of the software to enable vOLT came out of our AT&T Foundry innovation centers and the AT&T Labs. Those engineers are now working with the open source experts at ON.Lab. The result is an open source system to support broadband access, just like GigaPower. It’s called CORD, short for Central Office Re-architected as Datacenter. CORD will not only benefit AT&T, but other carriers and cable and satellite operators alike. It definitely accelerates the overall innovation ecosystem. And that’s good news for everybody.
  • We configure devices in our software-based network using a tool built on a data modeling language called YANG. We’ll submit our customized YANG design tool into open source through the OpenDaylight Community. Innovators will be able to create services that plug into our software-defined framework.
  • We’re active contributors to OPNFV, which just recently released ARNO, the industry’s first open source platform for NFV. It shows a lot of promise in allowing users to customize their platform to test different virtualized functions.

As I’ve said before, this year is a key step towards our goal of virtualizing 75 percent of the network by 2020. We plan to hit five percent this year. That’s the critical part, the foundation. The progress outlined above is just the latest proof that we’re well on our way.

Network on Demand is one of our first SDN-enabled services. Thanks to the collaboration of the open source community, more are coming soon.

John Donovan
John Donovan CEO – AT&T Communications