Learning About 5G Cellular: The Technologies
In this series, we explore 5G cellular technology – how it isn’t a threat to Wi-Fi, and why Wi-Fi will continue to prosper even as 5G becomes the norm. Last week I defined what is 5G Cellular. Today we dig into the technologies involved.
No matter whether we’re talking wide-area (cellular) or local-area (Wi-Fi) or any other class of wireless, the basics – driven by the laws of physics – remain the same. The objective is to get a signal from point (a) to point (b) reliably, efficiently, and with throughput and latency sufficient for the application at hand.
When we get right down to it, the air is a hostile environment – by comparison, wire is a land of peace and tranquility. Fortunately, a remarkable number of very smart people have devoted countless hours to solving the technical challenges of wireless, which is why today’s gigabit-class speeds no longer seem so amazing – even though, of course, they really are. Remember the days when a 9.6 Kbps modem seemed speedy?
Most of the advances in broadband wireless over the past 20 years have in fact been driven by the wireless-LAN industry, and the wide-area guys have steadily adopted the innovations that are at the heart of Wi-Fi today. 5G, as we’ll explore below, borrows liberally from Wi-Fi in improving the various dimensions of performance (noted last time) required to eliminate the historical gap between wire and wireless. This was a key goal of 5G in the first place, and the justification for the hundreds of billions of dollars (globally) that will need to be committed to make the 5G vision a reality.
Here’s a rundown of how 5G will become more like Wi-Fi. Most of this, I trust, will look familiar to anyone operating a Wi-Fi infrastructure today. There are two keys to wireless success, and they are making better use of spectrum, and clever engineering, obviously with a lot of overlap between the two. Expanding a bit:
- More Spectrum – Spectrum is the highway that wireless drives on, and just as eight lanes are better than two in the automotive world, more spectrum is key to higher performance in radio. Unfortunately, like real estate, they’re not making any more spectrum, so we have to be clever in terms of how we allocate and use spectrum.
While Wi-Fi operates (and very well, I might add) in the wild west of the unlicensed bands, cellular has historically traveled the reserved but very-limited bandwidth allocated via the monumentally-expensive and wasteful auction process. While such works very well indeed in generating revenue for governments, it’s not so good in assuring optimal use of spectrum.
5G will require wider channels and we’re pretty much out of appropriate spectrum that can be auctioned. That’s why 5G proponents are looking at the unlicensed bands as well as the millimeter-wave bands above 60 GHz. Since both of these are suboptimal (5 GHz because Wi-Fi is already there, and 60 GHz because propagation is problematic).
The key will be in re-allocating existing carrier-licensed spectrum to 5G, which really only works well if carriers can rapidly migrate the installed base to 5G handsets as well. With essentially everyone in every industrialized society now wielding an expensive mobile device or three, such won’t be easy.
- Small Cells – Since assuming the availability of more spectrum isn’t a very good strategy, making better use of the spectrum we already have is a much better option. “Small cells” is the cellular term for the dense deployments that have been a mainstay in wireless LANs for over a decade; in other words, less range means improved throughput and capacity and easier reuse of that limited spectrum over and again – the essence of cellular itself.
- Cognitive Radio – This is the ability of a radio system to dynamically optimize its use of radio spectrum – specific channels, channel bandwidths, and even varying modulation as appropriate. Guess what? This set of capabilities has also been in Wi-Fi products for decades. It would also be nice to be able to re-allocate spectrum among carriers on a demand basis, and such sometimes occurs, even as the auction process increases costs here.
- Massive MIMO – Switching to clever engineering, MIMO (multiple input, multiple output) of course forms the backbone of both 802.11n and 802.11ac, so it obviously works. There so much of a boost in performance that it appears to violate the basic laws that govern communications systems themselves.
I’ve never met anyone, even those with very sophisticated backgrounds in the technologies of communications (to say nothing of the underlying mathematics) who didn’t scratch their head at least a little when first presented with the concept, which is as follows: apply multiple transmitters (the inputs), each with a unique encoding of the information to be transmitted, along with multiple receivers (the outputs), in a multipath-rich environment. Multipath, as a refresher, are the echoes and reflections of radio waves as these bounce around in a typical three-dimensional spatial environment. Normally multipath is destructive, but, in the case of MIMO, the more multipath, the better the performance. Go figure.
We typically see Wi-Fi MIMO implementations with two or three (and, with Wave to of 802.11ac, four) transmitters and receivers, and the .11ac standard specifies up to eight. Physical limitations on most mobile devices make two or perhaps three a good upper bound, but there’s nothing to prevent the infrastructure side from having a very large number – tens or more – of transmitters and receivers.
More MIMO, then, and we see more capacity, throughput, and reliability, albeit with at a higher cost, of course. Since 5G is all about capacity, reliability, and throughput, then, massive MIMO is a key area of research at present. Results of 145 bits/second/Hz. in a 20 MHz channel have been reported in the lab – truly remarkable, and perhaps indicative of what we can expect in subsequent generations of Wi-Fi as well.
- Meshes – While you’ll more commonly hear the term multi-hop used when wide-area wireless networks are discussed, meshes are, like everything else on this list, old hat in the WLAN world. Meshes are an easy way of improving coverage, albeit usually with at least a slight penalty to overall capacity. Still imagine all those connected, driverless cars (and they’re on the way, really!), and you can see just how valuable this strategy really is.
- Network architecture – We expect 5G to take liberal advantage of major advances in network architecture, including Cloud, network functions virtualization (NFV), and software defined networking (SDN) – just like, you know, Wi-Fi.
- A Single Common Technology – And, finally, no one should underestimate the value of deploying only the minimal number of technologies required to get the job done. Ideally, like 4G today, this number will be one.
There were over a dozen 1G technologies, at least four 2G, three 3G, and three 4G technologies at given moments in time, with corresponding marketing battles back and forth. Remember WiMAX, for example? It was an official 4G technology, but died a slow, painful death, with its proponents not realizing that it had no real market or technology advantage until it was simply too late. Billions were wasted.
5G has a real shot at provisioning, with backwards compatibility to 4G and perhaps 3G as well as a single technology base for the future. Such offers enormous engineering, manufacturing, marketing, and cost advantages. The less friction in each of these domains, the greater the likelihood of success.
On the other hand, I and a few others have, with straight faces, proposed that 802.11n and .11ac Wi-Fi be considered as official 5G technologies (and while I haven’t gone as far as to suggest that Wi-Fi’s protocols be modified to work in the licensed bands, well, you get the idea). As you can see, the differences between Wi-Fi and proposed 5G technologies are minimal at best.
Sounds pretty good, right? Well, it is, what with (surprise!) 5G borrowing so much from the proven technologies of Wi-Fi. And all of these advances have led some to conclude that Wi-Fi will eventually be subsumed into 5G. And that 5G technology will be all we ever need to address essentially every application, from IoT to real-time 4K (and more) video.
As seductive as this vision might be, though, it’s still dead wrong. More on that next time.
All Posts In This Series: