How does a technology that enables uber-speeds but only spans 200 metres accomplish anything meaningful?
Frequent readers of Test Cafe know that I have recently focused on the rollout of 5G, the fifth generation wireless communication system. I’ve chosen to focus much attention to 5G in a test and measurement blog because of this basic fact: Wireless communication is the most consequential segment of the test and measurement industry, and each generation enables massive shifts in market share and instrument architectures. Wireless communication testing is both, mission critical and difficult. So, when a new technology wave begins to roll across the horizon, I’m on the case.
In my previous column I expressed some scepticism about the timing of 5G. After all, when so many current wireless users complain that coverage is their major issue, not peak speed, how does a technology that enables uber-speeds but only spans 200 metres accomplish anything meaningful? Answer: it doesn’t, at least for that use case. For the sceptic’s argument against 5G being deployed uber-quickly, just read my most recent column "Not So Fast." Once reading that, you may agree with me that it is critical to understand the use cases that will drive the applications. However, on the flip side, this column will now propose a use case that could drive the 5G rollout.
Hint: It’s not mobile communications.
Why do I say that? Well, I’m defining 5G to be mmWave communication. Given that, we know how chaotic propagation is at 30GHz and above. Actually, we don’t even know that well, because it is really chaotic. Small wavelengths require beamforming to compensate for attenuation at those frequencies. Well, the beams are really ill-behaved. Sure, they work well in theoretical free space, but put them in an environment full of bizarre reflections and absorption (e.g. planet Earth) and the beamforming becomes very complex. Fortunately, there are algorithms where you can servo a good beam to a client- but what if that client is moving?
While at the International Microwave Symposium (IMS), I attended a Keysight seminar keynoted by Mark Pierpoint, VP and General Manager of Keysight’s Internet Infrastructure Solutions Group. One of many interesting charts Mark displayed was one that showed feasibility of servoing a beam onto a moving object. If you consider interference patterns, a quarter wave movement can change the combined amplitude of the signal greatly. Coupled with practical closed loop servo timing and latency, Keysight generated what velocity/frequency combinations were viable, (anything within the 1ms latency objective of 5G):
Figure 1: Green good, Red bad. The summary of this is that automotive speeds are problematic for mmWave communication. None come close to being servo’d within the 1ms 5G latency objective. (Source: Keysight Technologies)
Summary: real time beamforming for a high-speed moving object just made 5G communication exponentially more difficult.
So, self-driving cars aren’t the killer app. At least at the get-go.
However, there are some non-mobile applications that may just be the trick. I’ll choose one: being the final link of an ISP (Internet Service Provider). Why do we need fiber to the home (or apartment) if we could just beam the data to the recipient?
Figure 2: Bird's eye view of a city. Starry shows their coverage footprint from a single base station as concentric blue-grey circles, while clients are shown as orange dots. Is this an example of the first 5G application? (Source: Starry)
In fact, this application is tailour-made for cities. A 200 metre circle from an access point covers a large number of users, particularly high-rise apartment dwellers. No need to tear up streets and walls to lay fibre. Indeed, there is a start-up that is already doing this: Starry. The first trial is occurring now, in Boston. Place their access point near a window in a served area, and get data rates up to 1 Gigabit/sec. For many wireless providers this is an entirely new service with new customers. It is not merely satisfying current users. Think about China as they migrate population from the rural regions to cities—what will be the telecom infrastructure to service them? Fibre to the intersection and then mmWave to the apartment seems like a perfect architecture.
New services. New revenue. Easier to design and achieve. If I had to place my bet on the first 5G mmWave killer app, I don’t think it will be mobile at all. I’d bet fixed internet access.