You have to keep in mind that with enough antennas (and enough VERY high bandwidth connections between the antennas - but only has to exist on the service provider side, plus the antennas can be pretty close together and still work) there is, in theory, 2^n interference patterns, each of which is localized. Each of which can transmit data.
If you use the interference patterns to transmit data instead of the signals directly, and you have enough antennas, you have essentially unlimited wireless bandwidth, at least at relatively short range (ie. cells). Granted, doing this for more than 4 or so antennas is currently beyond the capabilities of deployed equipment, but we'll get there.
I don't understand why people keep saying that we won't be able to satisfy bandwidth demand on wireless. We can, and it may be a lot cheaper (esp. outside of the cities).
I don't understand why people keep saying that we
won't be able to satisfy bandwidth demand on wireless.
Is it possible, from an engineering perspective, to build a wireless network with very high bandwidth and generous transfer limits? Absolutely! You just need to increase last mile bandwidth every time you get a new subscriber, so cells never get overloaded.
Is it possible, from a market incentive perspective, to build a wireless ISP business where it's more profitable to increase last mile bandwidth or turn away customers, rather than overload a cell? I'm not sure.
> Is it possible, from a market incentive perspective, to build a wireless ISP business where it's more profitable to increase last mile bandwidth or turn away customers, rather than overload a cell? I'm not sure.
As someone who's now lived >5 years in 3 locations, I've always found that there are things that oppose last mile bandwidth buildouts and there are things that don't matter.
What doesn't prevent capacity buildouts:
large costs (but within an order of magnitude or so of reasonable, I'm not talking "FTTH to a fishing village on a remote island" large costs)
technological limitations (again within an order of magnitude of what's possible)
What prevents capacity buildouts:
monopolies
cartels
absurd (and required) levels of capital investment (like buried fiber/copper capacity)
And you know what I find another huge positive about wireless networks ? There's 15 in every place I've lived. 2 "really" independent ones, 1 "more or less"/partially independent one, and a dozen or so that piggyback on the others' networks, providing for more price differentiation. Physically, there's room for something like up to 5 truly independent ones, using different frequency ranges. If the military were to get off half the spectrum, we can easily make that 20, so if the situation gets bad, there is an actual reserve. This means that governments can always threaten the existing providers with allowing another one in, in a way that they can't with wired networks.
Wired providers, there was never more than 2. Mostly there was just 1. Even when there were 2, they were absolutely not comparable (I mean, I've done telecommunications consulting for huge companies, I know there are really 50 or so, with 49 having extremely limited reach, unusable for consumers). This lead to monopoly, which lead to all sorts of shit situations, including bandwidth limits and turning away of customers. Even then, in the one place were 2 were available, the key to good service was to switch wired provider every 2 years.
Am I right then that to get super fast wireless everywhere we need lots of fibre to back-haul it? That's my understanding. All the 4G super fast masts springing up are all piping that hot traffic down fibre to the next hop.
All wireless internet connections meet copper and fiber eventually.
Wireless is fairly useful to provide basic coverage to a huge area at relatively low cost. The quality is lacking though. Low Cost, Coverage and Quality, pick two.
Wouldn't there be too many noise problems to reliably determine the right interference pattern?
I've heard this suggested before, but it seems to be slow in the making.
If you use the interference patterns to transmit data instead of the signals directly, and you have enough antennas, you have essentially unlimited wireless bandwidth, at least at relatively short range (ie. cells). Granted, doing this for more than 4 or so antennas is currently beyond the capabilities of deployed equipment, but we'll get there.
I don't understand why people keep saying that we won't be able to satisfy bandwidth demand on wireless. We can, and it may be a lot cheaper (esp. outside of the cities).