Status Update: Network AI
Imagine opening the box to a new puzzle. A big one – 1,000+ pieces. A complicated one – shades of colors and blurry details. And as you sift through the pile and start to make sense of it all, you realize there are missing pieces. It’s confusing, frustrating and incomplete. Now, imagine a friend or colleague has those pieces and knows exactly where they fit. They’re able to fill the gaps and complete the puzzle. This is a simple way of explaining AT&T’s success with the open source community and our journey to an autonomous network.
In other words:
We’ve seen first-hand how collaboration is driving the next technology revolution. When you combine open source efforts with technologies like software-defined networking (SDN), 5G, edge computing and artificial intelligence (AI), you get the full picture – Network AI. It is the foundation on which we’re building and integrating software, tools, applications and platforms. And it continues to grow, diversify and improve. Our investment in open source has led us to connect with industry leaders, developers and vendors across the world all trying to solve for similar problems. All trying to piece together the same puzzle faster than doing it alone.
So, what’s new?
In the past 2 years, we’ve contributed more than 10 million lines of code to the open source community. Think of these lines of code as pieces to a much larger, complicated puzzle. Together with dozens of other companies, we’ve helped progress the development of Network-AI by contributing to open source key technologies like artificial intelligence (AI), edge computing, SDN and 5G. Here’s the latest:
Acumos AI, an industry-first AI platform and marketplace, announced its second public code release, named Boreas. This version of Acumos AI provides an enhanced user experience, model training capabilities and support for licensing and commercialization of models. Future releases will continue to build upon each other as we work to make AI accessible to everyone.
A few short weeks ago, Akraino Edge Stack, a software platform for edge computing systems and applications, issued its first release, R1. Akraino is comprised of 11+ blueprints supporting a variety of edge use cases. One of these being the Radio Edge Cloud (REC), the first example of the Telco Appliance Blueprint family which provides a reusable set of modules that will be used to create sibling blueprints for other near real-time appliances. Its main use case is supporting the near real-time RAN Intelligent Controller (RIC), a new network element that enables external applications to control aspects of the 5G radio network. Partnering with Nokia, we contributed the seed code for RIC in April to the Linux Foundation as part of an O-RAN-led project. As a founding member of the O-RAN Alliance, we remain committed to creating an open and intelligent next generation RAN.
And finally, ONAP is poised to announce its next iteration, Dublin, in the coming weeks. ONAP Dublin promises new and enhanced blueprints, standards alignment and the addition of several new LF Networking members.
We’re not just testing ideas in a lab or discussing possibilities. We’re laying out a vision and using open source to expedite the development of Network-AI to make it a reality.
Opening Up for 5G and Beyond: Open Source and White Box Will Support New Data Demands
Software is Our Thing: How We’re Using Software to Continue Our Transformation
AT&T and Nokia Accelerate the Deployment of RAN Open Source
First Public Code Release of the Acumos AI Project