Is 5G Unlicensed a thing or a contradiction?

This article is from the CW Journal archive.

In the balance between control and empowerment, or perhaps even between stagnation and chaos, where are our telecoms initiatives today? And where do we want them to be with 5G? Today’s mobile technology can be seen as either well controlled or strangling growth. Is Unlicensed 5G empowering or just chaotic?

The journey to 5G really began when the ITU started work on its vision for IMT-2020, now focused on eight capabilities supporting the three use cases of mobile broadband, ultra-reliable low-latency communications and massive densities of interconnected machines. Industry and Enterprise welcomed this vision for its clear use-case focus and the aspiration to enable new economic models.

The first GPRS handsets such as the Blackberry 5820 and Motorola L7089 Timeport  appeared in the shops nearly two decades ago. The point of 5G is that it’s not supposed to be about the technology anymore. The technologies are already there for almost every use-case with LTE as the foundation of wide area voice and data complemented by LPWAN for machines and mmWave for capacity. With IMT-2020 the wireless and network engineering may become distributed and generic but the licensed business model remains centralised. The challenge is in making investment in connectivity more palatable to operators, enterprises and governments with the goal of delivery to consumers without rationing. Enterprises see the value in enabling an unlicensed model alongside traditional MNOs to give better control over the levels of coverage capacity and latency within their organisations.

Mobile networks are highly engineered deployments of sophisticated technology in premium spectrum. Their limits are predictable and well understood. The technology is coordinated, controlled and stabilised making use of powerful computing resources, excellent network radio performance and good metrology. Whether directly through the costs of equipment or indirectly through licencing and regulation, mobile coverage comes at a price and for most users the single biggest frustration is, unsurprisingly, coverage.

In return, a mobile connection makes phenomenally efficient use of the precious resources of our thin mobile devices. Battery life is about screen brightness rather than the quantity of data transferred from base-stations. One would expect the same deterministic quality to be attractive to the enterprise IT team on campus as well as on the road but I believe they have not yet found this part of the 5G value proposition to be compelling.

We think of Wi-Fi as cheap and it generally is to its users, even in carrier-grade form. The performance it delivers also builds on the enormous investment in standards, software, silicon and accessible product design that has been made over its history and this investment continues today. The user and installer can be blissfully unaware of what’s going on beneath the surface and that’s how it should be.

By their nature, Wi-Fi’s unlicensed frequency allocations are shared by others, notably RF heating equipment in the narrow 2.4 GHz band which nevertheless remains the most popular. At 5GHz much more spectrum is available and band sharing is mostly with other Wi-Fi systems.  The license exempt mmWave band at 57-71 GHz (’60 GHz’) has access to at least six non overlapping channels, each 2GHz wide with plenty of capacity for throughput and sharing.

The underlying cleverness of Wi-Fi’s band-sharing technologies, informed by decades of experience, gives a well understood quality of service. Even without the planning and coordination enjoyed by licenced systems, Wi-Fi works very well in our homes and offices where competition for spectrum can be managed automatically. Mitigations against interference rely on distributed protocols so devices, as well as base-stations, have to work harder to manage congestion. Ultimately, if local congestion is uncontrolled then the only way to stay connected is to bring client and access point close enough to connect a cable. I submit that this begs the question: Can Wi-Fi ever be 5G?

EARLY GPRS

The Blackberry 5820 was one of the first GPRS devices. It pioneered mobile email, security  and  user interface design.

Users need guarantees. Investors want guarantees.  But these come at a cost. The licenced model should work well if you want  certainty from the technology and a spectrum licence buys some exclusivity. For regulators, the auction experiment has been effective in allocating spectrum to those most likely to exploit it quickly when growing operators needed to show their assets working on their balance sheets. In this case the cost of the licence is akin to a tax. Where the licence holder delivers service, you get service; where they don’t, you don’t and there’s not much you can do about it.

With an emphasis on improving coverage and quality of service and to make the best use of spectrum the Spectrum Policy Forum is looking at a new model to counter this, whereby if one licence holding mobile network operator isn’t using its spectrum in a particular place, another licenced operator can extend its bandwidth by borrowing the unused frequencies. The plan is to roll this out at 26 GHz. This forms a middle ground between licenced and unlicensed.

Fully unlicensed technology brings no absolute promise of exclusivity, but, with technology and spectrum, it empowers independent investment in infrastructure in ways that licenced technology can’t. How much, depends on the benefit, its longevity and the risk from future overlapping ad-hoc deployments.  The lifetime of the return is set by the bandwidth available for future demand and yet-to-be-invented applications.  The threat from neighbouring deployments depends on spectrum and selectivity. Wi-Fi at 5GHz is already used for infrastructure applications, notably long range fixed wireless access with directional antennas providing both gain in link budget and protection against off-axis systems along with plenty of channel options if needed.  The advent of unlicensed mmWave technology brings even stronger protection from neighbouring systems as a result of the necessarily narrow beams, and 60GHz unlicensed mmWave also brings copious bandwidth for long term growth in demand.

A 5G proposition that carriers, ISPs and enterprises want to use on campus as well as on the road must bring plenty of bandwidth, long-term return and support straightforward installation. A 5G that connects us in trains, towns and villages must support new economic models. With few exceptions worldwide, demand for data is limited by supply,  in turn limiting innovation in products and services.

In the balance between control and empowerment, licensed 5G alone is too inhibiting for enterprises and operators to deploy the cost-effective coverage that helps feed and support our economic growth. With Wi-Fi and mmWave, 5G Unlicensed is a thing.