Researchers from New York University have offered a surprising demonstration of millimeter-wave wireless communications.
While this largely unused and untested frequency band—usually assumed to be a key component of 5G—is characterized by its poor performance across long distances and among even low-density intervening objects (like bushes), engineering professor Ted Rappaport and colleagues found in experiments spread across the wooded, rural hills of southern Virginia that it performs much better than expected. Even where signal paths were interrupted by trees or hills, millimeter-wave signals could in some cases be detected over 10 kilometers from their source, a base station on the porch of Rappaport's cabin.
It seems like just yesterday that we were rocking 3G connections and feeling pretty good about it. This is the technology, after all, that normalized mobile video, pocket web surfing, location services, and the cloud itself. The next thing, 4G—including LTE and WiMAX—seemed to come almost at once, offering data rates not too far off from those expected from regular old piped-in broadband. By contrast, 5G is still quite a ways off, with estimates suggesting that we'll be waiting until 2020 at least.
Read more: Why It's Taking So Long to Get to 5G
What, exactly, are we waiting for? More than speed, what 5G aims to provide is capacity. That's sort of the riddle. We can provide internet across cellular networks at the speeds required by cloud services and video streaming and all of the rest of it, but scale is less assured. To be clear, faster wireless internet of the sort that we might imagine seamlessly integrating with the piped-in kind is part of the still rather loosely-defined 5G picture, but the volume of connections required by the Internet of Things, in particular, is poised to change the demand side of the equation enormously.
To make this work, 5G will need to use millimeter-wave signals. These are extremely high-frequency radio waves capable of transmitting large amounts of data in relatively small, dense chunks of signal. The catch with millimeter-wave signals, however, is that they don't do so well out in the wild—a bit of rain or some foliage is enough to really screw things up. As such, it seems likely that 5G cells will need to be fairly small, even on the scale of individual buildings. Millimeter-wave 5G wouldn't seem to have much to offer rural areas then, given its apparent poor performance over long distances.
And this brings us back to Rappaport's porch, which was stacked with radio equipment this past August. The ensuing experiment wasn't quite the exercise in contrarianism that it might sound. As Rappaport explains in a paper posted recently to the arXiv preprint server, the models generally used to predict millimeter-wave performance are based on an outdated and poorly understood device known as the rural macrocell (RMa) scenario. Crucially, experiments validating the RMa scenario have been limited to relatively low frequencies (below 8 GHz), while actual millimeter-band wireless might be between 30 and 300 GHz. This should invalidate RMa at higher frequencies, according to Rappaport, leaving rural 5G in need of a new model.
That new model, according to the arXiv paper, is the close-in (CI) free-space reference distance model. This is what actually fits Rappaport's measurements. "A first-of-its-kind RMa propagation measurement campaign at 73 GHz was conducted in a rural area to confirm the accuracy and validity of the proposed CI RMa model," he reports, "while demonstrating the remarkable distances and coverage that may be obtained using [millimeter-wave] communication beyond 10 km in an RMa scenario."
Speaking as someone that's suffered through it for the past few years, rural internet suuuuuccckks. Even good rural internet is bad. And, at first glance, it would seem that future wireless technologies in the form of 5G are poised to make the rural-everywhere else internet gap even worse, or at least offer the status quo. Maybe not.
Get six of our favorite Motherboard stories every day by signing up for our newsletter.