I attended the Small Cells World Summit in London this week. The event, hosted by the Small Cell Forum, was back in person after a two-year absence, with a live stream for remote attendees.
I've attended the Summit many times over the years. The event helps set the agenda for small cell technology and product development and is a useful place to identify emerging commercial trends and customer requirements.
Here are some thoughts from the public sessions I attended.
1. Neutral host is taking over in public venues – especially in large multitenant buildings, larger public venues and campuses.
This is because:
a) it is better for nearly everyone, especially building owners and users;
b) mobile operators don't want to fund these builds or operate these networks.
There were lots of neutral host providers at the event and lots of examples of premium real-estate developments that use neutral host – nearly all the biggest, most prestigious venues in the UK, for example. Specialist neutral host providers Wireless Infrastructure Group, Freshwave, BAI Communications, BT Wholesale, Crown Castle and Cellnex were all in attendance.
Granted, this is a UK-centric view (the event, although global, was in London and included a lot of UK people). In other markets, operator-led networks are still prevalent, but it feels inevitable that neutral host will expand globally – it's simply a better model for public access venue networks.
The four UK mobile operators – EE, Three UK, O2 and Vodafone – have collaborated to create the Joint Operators Technical Specifications (JOTS) Forum, which publishes the Neutral Host In-Building specification. This is intended to help neutral host providers build systems that meet operator technical requirements. The head of sales for Commscope EMEA and APAC called this out as very useful and said operators in other markets should create something similar.
2. Private networks are now the big thing in small cells – this is obvious, but there are many interesting aspects to it.
First, there are lots of interesting spectrum choices – too many to go into here. The key is building an ecosystem around the different bands. More broadly, the 5G ecosystem isn't yet in place to support widespread private network deployments, particularly for mid-tier IoT devices. NR Light will be important to address this. Watch this space.
Looking a little further out, mmWave for private networks is very interesting. Ericsson highlighted mmWave with reference to advanced industrial IoT uses cases where ultra-low latency, positioning and reliable high-speed interfaces are all key. This is an argument that Qualcomm has also made strongly – although, of course, you need a very high-quality network with multipoint radio transmission to deliver the required performance.
There were a few audience questions on how public networks (perhaps deployed on a neutral host) could be collocated with a private network at the same venue. This is an area of great interest – for example, at a stadium or airport where you need both operational network access and public access services – but overall, I don't think the program or speakers really had a handle on how this model should work commercially or technically. No doubt the expertise was in the room, but it's a topic I'd like to see more focus on next time.
3. Disaggregated RAN and Open RAN – yes, but … meh
There's a story here, for sure. And it's important, for sure. But customers don't really care, especially enterprises. It's mostly a supply chain and product development story. Open RAN is a useful enabler for smaller companies, sub-systems suppliers and integrators. But enterprises (and operators, for the most part) want integrated, supported solutions.
Think about it this way: if you think that private networks as a service is important, then sub-system suppliers and integration fade into the background. A good example is the specialist integrator/operator Vilicom, which explained how it uses open vRAN to create a supported, integrated solution for the customer. And thinking more broadly, it's easy to see neutral host evolving to, in effect, a "RAN-as-a-service" model. In summary, open RAN will be there but buried pretty deep in the small cell stack.
4. Flexible architectures
The interminable RAN split discussion shows no signs of resolving itself. There are pros and cons. The Small Cell Forum is obviously attached to nFAPI/Split 6 because it wrote the spec. And to be fair, it has a lot of merit in many small cell scenarios. But after n years and a lot of slides, I think we can say this probably won't ever be resolved. In the meantime, it is best to think in terms of multiple architectures, deployed according to the use case, and for vendors to create products that can adapt. This was basically how Airspan explained its approach: support all deployment architectures on a common technology base.
Start-up silicon companies Picocom and EdgeQ were on hand to say, 'don't worry too much, our silicon is flexible enough you can design multiple systems, with different splits, on our products.' They're not promising software-defined RAN splits, but the idea is that a system vendor can have multiple product SKUs on the same chip and flex in software.
For further reading…
@KeithDyer from The Mobile Network and @DisruptiveDean from Disruptive Analysis were in attendance burning up the Twitter with live tweets (#SCWS2022). There's also the usual LinkedIn coverage – although it's pretty vanilla fare, as you'd expect on that platform.
Thanks again to the Small Cell Forum for a useful and productive event.
— Gabriel Brown, Senior Principal Analyst – Mobile Networks, Heavy Reading