4 ms·

I see some major issues here. 1. Weather - The higher the frequency, the more things like rain/fog/sleet/snow start to affect you. For example, 24ghz in Florid

I see some major issues here. 1. Weather - The higher the frequency, the more things like rain/fog/sleet/snow start to affect you. For example, 24ghz in Florida is useless for more than about 1 mile, unless you're okay with it going out in the rain. In Arizona, it's great for up to 5 miles. 39ghz will be more limited than this. 2. Spectrum - The most efficient radios I know of achieve about 10bps/hz -- that means you need 100mhz of spectrum for 1gb/s. That's total speed, a 900x100 or a 500x500 connection. I don't know the details of what technology is being used or planned here but it's going to be at least half of that for a practical multi-point system, even worse since these will be user-installed where someone will settle on a weak signal that burns more spectrum for the same speed. My guess is around 2bps/hz, but I'll stick with 5 for now, giving the benefit of the doubt. Assuming the entire 39ghz(39000mhz to 39999mhz) band is available for use, at a likely best-case scenario of 5bps/hz that's about 5gb/s of traffic for any given signal area. If you're selling 1gb connections, even at a 50 to 1 oversubscription rate, that's only 250 connections. That will barely put a dent in an urban environment with 100,000+ people. 3. Spectrum re-use. Cell companies are able to re-use the spectrum they have because they can control with a good degree of accuracy where their signal goes. At 39ghz, this signal is going to reflect off anything and everything, making spectrum planning very difficult. You might need two access points on either side of a large building for coverage, but thanks to reflections, you won't be able to re-use any of that spectrum because the signal will be bouncing all over. You have very little control over where your signal goes in an urban environment. A truck driving down the road may send your signal 2 miles in the direction you don't want it to go. 4. Bandwidth -- How many rooftops have 10+gb/s of internet available? Probably not many unless you happen to get lucky and have a carrier hotel in the same place you need signal. I'm a feet-on-the-ground(or sometimes tower) guy, so these are napkin numbers. I've seen a couple of real wireless engineers around here, maybe one of them can pitch in with some more concrete numbers. When Aereo first started, it was a sound idea with very little in the way on the technical side for implementation. As we know, Aereo's problems weren't technical. This project has many apparent technical issues, and probably some legal ones as well. I don't want to unnecessarily poo-poo his project. If it works I will be glad -- it will be another tool I can use to do my job -- but it seems like he's promising rainbows and unicorns right now, like so many other investor-bait news articles I read. Source: 20 years of experience working on/for/by/with many wireless internet providers and cell tower companies.

acjohnson55 11y ago
As far as the spectrum reuse question, they talk about active phased array tech, would that possibly allow them to get a lot more reuse, assuming they can somehow deal with reflections (with CDMA, or signal processing that integrates multipath signals, or something)? For weather, what if the base stations only need to work over some distance less than a mile? For bandwidth, if you consider this to be essentially a last mile tech, and furthermore, if the base stations can mesh network to reach backbone, might that solve the issue? Just some thoughts. I'm a total non expert.
- sathackr 11y ago
  I believe they are referring to a beam-forming antenna when they say active phased array. Basically, timing of the RF signal to various portions of the antenna is altered such that it restricts and focuses the beam towards an intended target. The technology has only recently come to the point that it is practical for consumer level applications. I don't see it helping with reflections, with endpoints all over, you'll have a bunch of very strong beams scattering everywhere, but it can help push the signal further than a simple omni-directional or even sector antenna. I don't believe it helps much with the received signal. A real engineer will be able to answer the signal processing question much better than I, but to my knowledge, multipath can only be effectively dealt with when all versions of the signal originated from the same device. When you have effectively a stray signal, it's difficult to tell even where the carrier is. There is some mitigation possible with timing of transmit/receive signals(usually with GPS), but this also requires some control over the propagation of the signal. As far as weather, for most areas, 39ghz would probably be practical out to 2 miles, with some brief interruptions for very heavy rain. But with the bandwidth limitations, this isn't in line with the stated goal of gigabit level service to high-density urban environments. This would certainly be a last-mile tech, but the described target is high-density urban areas, where population densities far exceed 1000/sq mile. I can't see this technology ever being able to function at gigabit speeds with those densities. Mesh networking has it's own issues. If you're going to run it on the same frequency as your endpoints, then you are sharing spectrum with them and your overall bandwidth is significantly reduced. You can backhaul it on a different frequency, but now you have more equipment and expense, licensing, etc... I have to be careful saying it because I haven't kept up with the lastest in mesh developments, but over the last 10 years, I have never seen a mesh network system work reliably with any sort of appreciable speed. Most have failed miserably. The idea looks great on paper, but rarely works out in the real world. There are tons of last-mile tech available that work very well. FSO(free space optics)[1] is one of them. It has nearly unlimited bandwidth. It's limitation is, it is, as the name implies, optical in nature, so line of sight is an absolute must. Currently available equipment is overpriced as well, but a newcomer could change that. The FCC just opened up about 150mhz of spectrum in the 3.5ghz area also for last-mile delivery of wireless broadband - Citizens Broadband Radio Service[2] This band is aimed at rural deployments, and isn't concerned with pushing 1gb to everyone, but will go a long ways towards getting a lot of people on the internet at decent(20mb/s+) speeds. The 5ghz unlicensed band is also very effective at distances under 2 miles, and can be pushed up to 100 miles in certain situations. There is about 600mhz of spectrum available here, some of it ideal for urban situations as it shares frequencies with weather radar and is under strict power restrictions that make it difficult to use for longer and rural connections. On a tangent, I deal with people nearly every day that can't even get 3mb/s DSL service. The only high-speed that's available to them is satellite, with data caps and horrendous latency, or 3g/4g with spotty speeds and availability, data caps, and impractical per GB billing. When they are able to get even 10mb/s of real internet, they are happier than a 5 year old on Christmas morning. It's by far the most rewarding thing I get to do. It would be nice to see more money flowing towards things like this. But not like FRBA [3] -- Essentially $24 million wasted. [1] https://en.wikipedia.org/wiki/Free-space_optical_communication https://en.wikipedia.org/wiki/Free-space_optical_communicati... [2] https://www.fcc.gov/rulemaking/12-354 https://www.fcc.gov/rulemaking/12-354 [3] http://blog.tmcnet.com/on-rads-radar/2013/04/florida-broadband-litigation-woes.html http://blog.tmcnet.com/on-rads-radar/2013/04/florida-broadba... edit: fix footnotes

acjohnson55 11y ago

As far as the spectrum reuse question, they talk about active phased array tech, would that possibly allow them to get a lot more reuse, assuming they can somehow deal with reflections (with CDMA, or signal processing that integrates multipath signals, or something)? For weather, what if the base stations only need to work over some distance less than a mile? For bandwidth, if you consider this to be essentially a last mile tech, and furthermore, if the base stations can mesh network to reach backbone, might that solve the issue? Just some thoughts. I'm a total non expert.

sathackr 11y ago
I believe they are referring to a beam-forming antenna when they say active phased array. Basically, timing of the RF signal to various portions of the antenna is altered such that it restricts and focuses the beam towards an intended target. The technology has only recently come to the point that it is practical for consumer level applications. I don't see it helping with reflections, with endpoints all over, you'll have a bunch of very strong beams scattering everywhere, but it can help push the signal further than a simple omni-directional or even sector antenna. I don't believe it helps much with the received signal. A real engineer will be able to answer the signal processing question much better than I, but to my knowledge, multipath can only be effectively dealt with when all versions of the signal originated from the same device. When you have effectively a stray signal, it's difficult to tell even where the carrier is. There is some mitigation possible with timing of transmit/receive signals(usually with GPS), but this also requires some control over the propagation of the signal. As far as weather, for most areas, 39ghz would probably be practical out to 2 miles, with some brief interruptions for very heavy rain. But with the bandwidth limitations, this isn't in line with the stated goal of gigabit level service to high-density urban environments. This would certainly be a last-mile tech, but the described target is high-density urban areas, where population densities far exceed 1000/sq mile. I can't see this technology ever being able to function at gigabit speeds with those densities. Mesh networking has it's own issues. If you're going to run it on the same frequency as your endpoints, then you are sharing spectrum with them and your overall bandwidth is significantly reduced. You can backhaul it on a different frequency, but now you have more equipment and expense, licensing, etc... I have to be careful saying it because I haven't kept up with the lastest in mesh developments, but over the last 10 years, I have never seen a mesh network system work reliably with any sort of appreciable speed. Most have failed miserably. The idea looks great on paper, but rarely works out in the real world. There are tons of last-mile tech available that work very well. FSO(free space optics)[1] is one of them. It has nearly unlimited bandwidth. It's limitation is, it is, as the name implies, optical in nature, so line of sight is an absolute must. Currently available equipment is overpriced as well, but a newcomer could change that. The FCC just opened up about 150mhz of spectrum in the 3.5ghz area also for last-mile delivery of wireless broadband - Citizens Broadband Radio Service[2] This band is aimed at rural deployments, and isn't concerned with pushing 1gb to everyone, but will go a long ways towards getting a lot of people on the internet at decent(20mb/s+) speeds. The 5ghz unlicensed band is also very effective at distances under 2 miles, and can be pushed up to 100 miles in certain situations. There is about 600mhz of spectrum available here, some of it ideal for urban situations as it shares frequencies with weather radar and is under strict power restrictions that make it difficult to use for longer and rural connections. On a tangent, I deal with people nearly every day that can't even get 3mb/s DSL service. The only high-speed that's available to them is satellite, with data caps and horrendous latency, or 3g/4g with spotty speeds and availability, data caps, and impractical per GB billing. When they are able to get even 10mb/s of real internet, they are happier than a 5 year old on Christmas morning. It's by far the most rewarding thing I get to do. It would be nice to see more money flowing towards things like this. But not like FRBA [3] -- Essentially $24 million wasted. [1] https://en.wikipedia.org/wiki/Free-space_optical_communication https://en.wikipedia.org/wiki/Free-space_optical_communicati... [2] https://www.fcc.gov/rulemaking/12-354 https://www.fcc.gov/rulemaking/12-354 [3] http://blog.tmcnet.com/on-rads-radar/2013/04/florida-broadband-litigation-woes.html http://blog.tmcnet.com/on-rads-radar/2013/04/florida-broadba... edit: fix footnotes