Local spectrum: Not just a 4G / 5G monoculture

Monocultures & diversity

In the telecoms industry, it's quite common to invoke the word "ecosystem". I see examples every day - "5G ecosystem", "cloud ecosystem", "WiFi ecosystem" and so on.

Yet there's a bit of an oxymoron here. In nature, few ecosystems are based around a single species. "Monocultures", single forms of life that cover vast areas of territory, are mostly human inventions. They are agriculturally efficient and can yield good returns for farmers, through scale and then automation.

But nature mostly abhors monocultures. They're risky and vulnerable. Most obviously, the impact of any catastrophic risk is magnified hugely. They are susceptible to parasites, infections or natural disasters. They can also monopolise resources, driving out other creatures and limiting the potential for further evolution.

One good natural example of a monoculture is a coral reef, where the "ecosystem" is actually dependent on a central species. But as we sadly see today with climate change and other sources of damage, reefs - and their non-coral residents - are vulnerable.

I'm not going to torture the analogy I'm making here. I'm sure you already know what I'm referring to - a prospective monoculture of 3GPP wireless networks, especially 4G LTE and 5G.

And you've probably also guessed the next word I'm going to borrow from ecology: Biodiversity. Environments such as tropical forests, homes to myriad species, are much more resilient to damage than coral reefs.

I am increasingly of the belief that we need to maintain a strong and rich ecology of "network diversity" - multiple competing and interacting technologies, each with strengths and weaknesses, sometimes overlapping in habitable niches, and sometimes addressing unique domains.

Enabling network-species diversity & dominance

In many ways, unlicensed spectrum has been the great catalyst for diversity, because it allows "permissionless innovation". The various bands have yielded all the versions of #WiFi, #Bluetooth, #ZigBee, #LoRa, unlicensed cellular such as MuLTEfire and numerous niche and proprietary variants. With the recent additions of 60GHz and soon 6GHz, we may get a further Cambrian Explosion of new wireless species.

Yet for licensed (national) spectrum, we've gone the other way - towards a 3GPP monoculture. We started with CDMA, GSM, even WiMAX and TD-SCDMA. But we're now focused down to LTE and 5G NR as almost the only inhabitants of the designated "IMT" or International Mobile Telecoms designated spectrum bands.

In theory, many regulators denote these bands as "technology neutral", but in reality that just means MNOs can upgrade between 3G, 4G or 5G without needing to get a new license. Since the WiMAX era, there has been no serious challenge to 3GPP hegemony, unless you count DECT and Nufront as "official" 5G candidate technologies. (*see brief comment on 5G & IMT at the end of the article)

Put simply - path dependency is in play here. Seemingly minor decisions taken years ago, or external events and influences, have now created "lock-in" for 3GPP wireless that is hard to change. The scale of MNOs, and the power of entrenched assets in both networks and devices, makes it almost impossible for anything but a monoculture to be maintained. Governments have, unwittingly, become agents of this status quo.

New shoots growing?

But in the last few weeks, I've seen a couple of signs of change, right at the fringes of the 3GPP world.

There's a common belief that the world's new "local access" spectrum bands are only intended for #PrivateLTE and #Private5G networks, deploying networks for enterprises or campuses. The main debate is whether it will be telcos, enterprises or new classes of service provider that will benefit. (See my last article).

But actually, the US CBRS band, Germany 3.7-3.8GHz band and UK 3.8-4.2GHz band are not *just* available to 3GPP technologies.

They are also technology-neutral.

The regulations specify power limits, interference parameters and various other rules, such as the need to query a database like a #CBRS SAS system. But they don't say "Only LTE or #5G NR radios are permissible".

And what's interesting is that there's actually quite a bit of innovation going on with non-3GPP technologies here. The lock-in and legacy aspects of national 4G/5G don't scale down to small, single-purpose networks.

I've spoken to several fixed-wireless vendors recently (including Cambium Networks & Doodle Labs) that have repurposed existing proprietary FWA radios to work with CBRS. They access the SAS systems like any other equipment, and can work with general access or protected frequencies.

And on a session on the recent WiFi Now virtual conference (link), someone from Cisco mused about the potential for #WiFi in local-licensed spectrum (LLS) - something I've advocated for years with the WiFi community. They were suggesting it as a possible future evolution path for industrial networks or other use-cases. It wouldn't surprise me if some of the other industrial WLAN vendors such as Siemens are thinking the same way, although they haven't said anything.

Apparently #CBRS isn't ideal for normal Wi-Fi, as it is set up with 10MHz-wide chunks, not 20MHz - but that doesn't stop tweaked non-certified versions of 802.11 for #FWA or other uses. It's worth noting that there are also various other "non-standard" versions of WiFi-esque networks already around that use niche/local bands, such as the US 4.9GHz public safety band.

I'd take a small bet that the future #WiFi7 will support LLS. (If not, it should. Wi-Fi Alliance needs to embrace these innovations)

As well as allowing for "network diversity", using non-3GPP technologies in LLS has an additional advantage - they don't need the heavyweight complexity of EPC or 5G cores, and all the other cellular network paraphernalia like SIMs. For some use-cases, that's a significant benefit, even though 4G/5G cores are becoming simpler and deliverable from the cloud, as many recent vendor moves such as Microsoft/Metaswitch has demonstrated.

Conclusions

Network diversity is important. Not just vendor& supply-chain diversity, as recent geopolitical trends have pushed, but underlying and genuine technology diversity. Look beyond the loaded framing of words like "fragmentation", to realise that single standards can be harmful.

There's a Taleb-style "anti-fragility" to having a well-stocked toolbox of different connectivity options, in licensed spectrum as well as unlicensed.

Now, despite the new emergence of new & diverse options, I still reckon the bulk of LLS usage such as CBRS will be for LTE or 5G, at least in the short term. But that's fine, as long as the regulators stop the farmers from deliberately damaging the precious new species.

If we really do start getting WiFi6e or 7 in industrial spectrum, that's a potential gamechanger, especially if it beats 3GPP Release 17 to maturity. We may see wide-area innovation with CBRS and other national bands too - fixed wireless is less lock-in prone than mobile broadband.

In time, we might see some genuine tropical-forest type ecosystems emerge, without a single vulnerable network platform species or apex predator.

[* for an interesting discussion of 5G NR and ITU's specifications for IMT 2020, have a look at this thread (link) It seems like there's been some shenanigens about 5G not yet having a proper 1-millisecond latency version, but instead marking its own homework to give itself a pass. So maybe 5G NR isn't actually 5G, embarrassingly]