LF Networking (LFN) is going deeper and wider with the Open Network Automation Platform (ONAP), lining up its eighth release today with what it calls a “5G Super Blueprint.”

The goal is to provide a reference design, hosted by LF Networking, that works across many open source frameworks for the 5G stack. This includes Magma Core Foundation, LF Edge Akraino, Cloud Native Computing Foundation (CNCF), and O-RAN Software Community, a project the Linux Foundation and O-RAN Alliance initiated about two years ago to develop open source software for 5G radio access networks (RAN).

ONAP Honolulu introduces more cloud native functionality with container network functions (CNFs) and Kubernetes resources, early support for end-to-end 5G network slicing, and deeper integration with O-RAN and 3GPP standards. The group is also expanding ONAP for use among enterprises and seeking contributions from organizations in multiple industries to make the platform more applicable to various business use cases.

“Originally, ONAP’s focus was really to support the network automation, which was required by carriers,” but the group eventually realized that some of the ONAP components could be reused for virtualization requirements in many vertical markets, explained Catherine Lefèvre, assistant VP of network cloud at AT&T and chair of the ONAP Technical Steering Committee. 

A major addition in the latest release is network slicing, which now supports RAN, the network core, and transport domains, Lefèvre said, adding that enterprises will also benefit from these improvements to network slicing.

One of the key initiatives for the Linux Foundation in the telecom industry is to harmonize standards in open source, “and I think we’re starting to see that full circle,” said Arpit Joshipura, GM of networking, edge, and IoT at the Linux Foundation.

“We have some good collaboration, formal collaboration that we wanted to set up so that we can continue this harmonization across these groups in the industry,” he explained. “That way we’re not bifurcating the market and the industry. It’s one standard, one group of requirements, one specification.”

The collection of documents now integrated into the ONAP stack runs millions of lines of code. “It’s not trivial, but at least we’re not saying ‘you can either do it the open source way or the specification way.’ That’s what we’re trying to avoid,” Joshipura said. 

ONAP’s key area of focus with the 5G blueprint is to integrate Magma for core and edge, Anuket, Kubernetes, and localized mobile edge computing.  The community intends to soon include RAN suppliers to continue on its journey of assembling an end-to-end design that is fully cloud native built in coordination with ETSI, TM Forum, MEF, 3GPP, and the O-RAN Alliance, he said.

Magma, which was initially open sourced by Facebook and brought under a neutral governance model at the Linux Foundation earlier this year, also provides support for end-to-end network slicing, Joshipura explained. That is scheduled to be in production this summer with a fully disaggregated open source 5G core running edge computing, and then gaining O-RAN compliance and support before the end of 2021.

“Network slicing is an end-to-end construct,” and it still requires growth to fulfill that vision, he said. Every layer, including RAN, the edge, network core, network automation, orchestration, and the physical layer needs to support network slicing. 

“The work on defining what a true slice is is complete. The work on implementing slicing is, at least from an ONAP perspective and from an orchestration and automation perspective, very mature,” Joshipura said. However, work is still needed in the packet or network core and the RAN, he added. 

The Linux Foundation expects several early deployments and trials of end-to-end network slicing to be underway by the end of this year. Some operators may have already deployed sub-network slicing, but that’s a more narrow framework and independent of the more complete definition and vision of network slicing, Joshipura said.

“People will continue to build on top of it and commercialize it, so early 2022 will probably be where all the pieces of the puzzle” are developed and ready for deployment, Lefèvre said. The next ONAP release is slated for the second half of this year.