Open RAN Phase 2

Building a common, interoperable infrastructure for radio access networks.


I first wrote about Open RAN in my post Fourth 4G Network Goes Live in Japan. Open RAN is a program driven by a group of European operators to build specifications for common architecture instead of getting “locked into” the closed architectures sold by the big base-station vendors such as Nokia, Samsung, and Ericsson. It is analogous to what servers went through over the last few decades, moving away from proprietary architectures such as Sun/SPARC/Solaris, Digital/VAX, and IBM mainframes, to the systems we have today, which use Linux as the operating system and where hardware is available from a plethora of vendors, not just for the “main” processor but for virtualized network and data storage.

RAN stands for Radio Access Network. There are all sorts of parts that make up a whole mobile network, such as the backhaul that connects the base stations to the internet, to POTS (Plain Old Telephone Service—yes, that really is the name used, although if you want to sound more professional, you can say public switched telephone network or PSTN) to the databases used for billing. There is also the all-important Home Location Register (HLR) which allows anyone making a call or sending a text to know which network the phone is actually located on if you are traveling, and thus how to make the connection. The RAN consists of the base stations that you see by the side of freeways or sometimes disguised as trees (as in the photo here that I took in Chinese Camp on the way to Yosemite). This provides the radio connections from the base stations to the phones (or other devices). In this post, for simplicity, I’m just going to say “phone”, but it has to be understood that it might be a 5G connected car, a smartwatch, or a laptop. The RAN includes the antennas, the radios, and the base stations themselves.

The Open RAN project produced the first list of technical requirements in 2021. It was driven by joint requirements of Deutsche Telekom, Orange, Telefonica, Telecom Italia, and Vodafone Group. There is a digestible executive summary that is 12 pages long. One key paragraph from the start of the summary:

The primary objective of the European operators working under the Open RAN MoU is to guide and foster the development of a non-fragmented Open RAN ecosystem and vendors’ roadmaps while ensuring Open RAN reaches parity with best-in-class traditional RAN solutions. Special attention is paid to encourage a competitive European Open RAN ecosystem of technology providers and system integrators, hence strengthening the European RAN industry.

One key part of the standardization is “Open Fronthaul.” Analogous to the backhaul that connects the base station to the internet, the fronthaul connects the base station to the antenna system. For 5G, these are very complex, with steerable phase-array antennas and other MIMO approaches. If the antenna systems and the base station hardware might come from different vendors, this needs to be a standard (in the same way as PCIe is a standard for connecting processors to much of the rest of the world).

As the executive summary says:

Open Fronthaul is considered as the key interface to implement a disaggregated, multi-vendor RAN, including Massive MIMO, as a baseline for first macro deployments.

In March of this year, an updated list of requirements was produced to move the project closer to fruition. The detailed standardization is managed by the O-RAN Alliance. Like many standardization processes, this is divided up into working groups with catchy names like “WG2: Non-real-time RAN Intelligent Controller and A1 Interface Workgroup”. Note that the O-RAN Alliance has many more members than just the original European operators that kicked off the program. Companies like AT&T, Verizon, China Telecom, SK Telecom, Singtel and many more are members.

In fact, there is a quote by Andre Fuetsch, CTO of AT&T Labs, on the front page of the O-RAN Alliance website:

RAN infrastructure that supports O-RAN ALLIANCE specifications, combined with increasing RAN virtualization and data-driven intelligence will allow carriers to reduce complexity, innovate more quickly and significantly reduce deployment and operational costs

One reason that this should be important to readers is that a lot of the detailed infrastructure will be implemented in silicon. Indeed, as in most markets, the cheaper the parts get, the more gets deployed. Again, a quote from that executive summary:

Other capabilities related to an intelligent and programmable RAN are expected to emerge later, offering the potential for specialist start-ups to emerge and play an active role. All are essential in the long run to build a competitive Open RAN ecosystem.

So it has the potential to be a big opportunity. Base station counts are measured in the millions per country.

As I said above, in March, release 2 of the technical specifications were issued:

Release 1 focused on the main scenarios and technical requirements for each of the building blocks of a multi-vendor RAN. Release 2 builds on those requirements and focuses in particular on intelligence, orchestration, transport, and cloud infrastructure, aiming at defining a fully automated and interoperable Open RAN system.

However, there was no new executive summary. There are plans for a 3rd release covering “full intelligence and automation.”

In 2022, there are also plans to run the standards into ETSI, the European Telecommunication Standards Institute in Sophia Antipolis.

Learn more

Leave a Reply

(Note: This name will be displayed publicly)