by Chris Rice and Kathy Tse
We’ve talked several times on the blog about rethinking how we manage and transport the huge data loads flowing within a city and from one city to another. These data flows run on our most high-capacity fiber optic lines using laser light. But we need to make these systems more efficient and interoperable as data demand continues to grow.
Over the last several months, we’ve been perfecting that new technology. The next step will be to launch it across our network.
Software-Control and Interoperability
ROADM stands for Reconfigurable Optical Add/Drop Multiplexers. They’re a kind of network switch that manages data traveling over high-capacity fiber optic lines. These are the workhorses of the network. Software-controlled ROADMs can automatically detect and adjust bandwidth. They can move traffic to different lanes as needed.
Starting in our metro network, we’ve been using technology based on the new Open ROADM Multi Source Agreement (MSA). This is the group we helped create to write industry-wide specifications for how software-controlled ROADMs should operate and how hardware from different manufacturers should work together. For the first time, we now have optical interoperability between equipment made by different vendors. We also have common management APIs and a common network model.
That’s huge.
Next Steps
We’re now moving toward scaling Open ROADM deployments using multiple vendors and an optical SDN controller that’s integrated into ECOMP. ECOMP is the operating system for our software-defined network. We’re starting with our Open ROADM deployment in Dallas. Soon, we’ll be deploying Open ROADM technology as the standard design for all metro ROADMs – and eventually all ROADMs.
We’re using the new ROADMs and SDN controller in our own management systems at AT&T. We’re starting to use these ROADMs for production customer traffic.
The interoperability specifications and models were developed by 15 MSA members consisting of network operators and equipment developers. The team is currently working on 3rd generation features including higher rate wavelengths.
Why are we making this change?
Today’s networks require integrated, multi-layer planning and service assurance to provide the best customer experience. We’ve deployed a multi-layer SDN controller in our inter-city network that provides centralized traffic engineering.
This is an improvement on the distributed protocol. Now, there’s a global view of traffic demands and we can see which network resources are available. We’re now enhancing the SDN controller so we can fully reconfigure the optical layer to boost network capacity where and when it’s needed. We’ll do it using intelligent algorithms.
From concept to reality
The biggest takeaway? We’re moving beyond the test and demonstration phase. We’re moving into the deployment and scale phase of our SDN-controlled optical network. This new technology, along with our AI algorithms, will help ensure our customers can continue to do whatever they need to do, from streaming movies to running their businesses.
Starting next Monday at OFC 2018 (Optical Fiber Conference) in San Diego, AT&T network experts will be speaking in a variety of forums about our optical network evolution and experiences with deploying Open ROADM and multi-layer SDN control.
Chris Rice - Senior Vice President – AT&T Labs, Domain 2.0 Architecture and Design
Kathy Tse - Director of the Photonic Platform Development group in the Packet/Optical Network organization of AT&T Labs