The main issue I see with this system is that there are so many factors that will disrupt the signal. assuming India, How will they handle rain and heavy pollution degrading the quality of the signal? or will it just be a? "Oh, it's monsoon season so no internet?"
I've seen some systems like this that will fall back to 5ghz or 2.4ghz RF connections when the lasers are blocked, so the throughput goes way down but at least the service isn't out completely. Maybe they'll do something like that.
Yeah, was thinking something along those lines too, but it would be a 25 or 50 fold reduction in speed. not to mention stability, but if iẗ́'s just basic web traffic then it shouldn't be a problem.
Fallback to other unlicensed bands like 24GHz or 60GHz would probably work better. The higher frequencies at least retain a compact dish size, smaller fresnel zone, and sharper directionality. That is, imagine hanging several of these optical devices + 5GHz fallback at ~1 meter vertical spacing on a pole. A weather event triggering failover on all optical links could lead to enough 5.8GHz self-interference to kill the fallback links as well. 5.8GHz devices that aren't synchronized in phase with each other often call for 3meter+ separation, eating up tower real estate.
Weather will also degrade 24GHz and 60GHz links. A friend of mine used to work at an ISP that relied on wireless links, and he said that even at 2.4GHz moisture was a significant problem (especially concentrated moisture in living things like trees). In general as the frequency increases attenuation due to water increases; at 60GHz particulates will also become a problem (consider that a selling point of 60GHz is that it does not penetrate walls, so you will have less interference in apartments or large office buildings).
(Edit: Looked it up; 2.4GHz is more readily absorbed by plants than 5.8GHz. Still, attenuation due to moisture seems to be a problem for microwave frequencies generally.)
Indeed the 24GHz and 60GHz gear has its unique rain fade issues not shared by 5.8GHz. Increased atmospheric attenuation of higher bands is the general trade-off for higher throughput and more compact physical isolation requirements. UBNT 24GHz gear is being using for ~10km links, but looks like the ~5km is the preferred limit. My original comment would have been better written to stress need to prove out the signal path over all intended bands, e.g. primary link over (near) optical and fallback over whatever RF band proves viable.
Maybe we're looking at it too much from a "wealthy developed economy" perspective (I don't know where you're located so, that may not be the case, but it's probably a fair judgment for most HN readers). If this is an affordable way for governments to bring some Internet access to places that currently have none, then that's a big jump in the right direction.
Disclaimer: I work for Google but have no knowledge of this program. I'm reading about it for the first time just like you all.
> If this is an affordable way for governments to bring some Internet access to places that currently have none,
But India is big on cellular. There's cell coverage almost everywhere, especially in states like Andhra Pradesh. You can get pretty decent speeds via 4G.
Why is this better than using a fiber optic cable? I guess light travels slightly faster through the air compared to either glass or plastic (the two typical optical media used for cables) which would reduce the transmission time, but I can't imagine that the slight decrease in latency is worth the massive increase in downtime due to atmospheric interference.
Less cable installation. Put it on a pole pointed toward the tower and you're done. It'd be nice in the US where the corporate cartels like Comcast/Charter/Verizon can't as easily blockade access to internet by manipulating control over utility poles through bribery of local and state officials.
Potentially cost of the fiber optic cable and the installation of it.
Also no idea in India but in the US a lot of times although their are existing power poles internet providers can't necessarily get permission to string their fiber on them. Same problem can happen with existing in ground pipes. And any trenching or installing new poles becomes very expense.
The decreased propagation time through air vs glass likely wouldn't yield latency improvement in relevant orders of magnitude. The primary benefit is cost reduction. Erecting a few towers with optical relays over high points in terrain would be cheaper overall to install & maintain than either aerial or trenched fiber. If a 3rd party builds something in your way, just relocate the tower whose line of sight becomes obscured (on the assumption tower sites are solar-powered and self-standing).
An alternative way to imagine networking an area is running fiber into the area then having a mesh network assemble itself building to building in town/city whatever. This has the added advantage of greatly reducing the power costs of the overall network for local communications.
The labor cost, materials cost, maintenance cost, and right of way / easement cost in laying cables is a huge barrier both cost, time, and pain in the ass wise.
The slowdown of fiber vs vacuum is about one third. A 33% reduction in latency is fairly significant for some applications. I've heard of financial trading firms using microwave relay towers for this reason.
The other issue is "long fat pipes." TCP is better tuned for it on machines now than it used to be but when you go over a long distance, the latency can decrease the usable bandwidth significantly. Of course, this seems to be more focused on last mile rather than back-haul. There are very few applications where a several nanosecond latency decrease matters.
I was hoping the article would provide more detail for why this is better or different from existing microwave radio methods. Higher bandwidth appears to the answer. It appears that some laser systems feature a hybrid system with laser and rf technologies, and will automatically downgrade to RF when weather, pollution, or other obstructions come up.
There's a lot of competition for the non-visible spectrum at useful levels. It's only natural we'd start to move up. Also, as your wavelength gets smaller it gets cheaper to start compressing signals in ways that don't interfere over the same physical area.
Be more pragmatic. It's about licensing. The microwave frequencies are licensed. Laser frequencies are not. Want to setup a network without first securing government permission, the obvious choice is laser. This is also a primary reason COPs switch to lasers for speed enforcement.
The bandwidth of individual connections is only half the equation, and largely irrelevant imho. What matters for the network is the lack of interference. Laser connections can spider across a city without real risk of contaminating one signal with another. So even if the point-to-point bandwidth is the same, the total potential bandwidth of the installed network will be much higher.
There are a lot of microwave connections (E.G. 24Ghz, 5.7Ghz) that are in the unlicensed spectrum. We are using these extensively in our network. See https://en.wikipedia.org/wiki/ISM_band
Unlicensed, but not unregulated. There are very detailed limitations on transmitter/antenna powers in those areas, limitations that are very country-specific. The rules for laser, where they exist, are more akin to "don't burn our people's eyes".
I feel like licensing is even less pragmatic than technological capability and limitation. Ubiquiti's 2.4ghz and 5ghz 'air fiber' radios are approved for use in India, license free, and those spectrums are license free under certain requirements. There is no difference between laser and microwave devices for in this use case, except that laser has higher speed but lower persistence in line of site.
https://community.ubnt.com/t5/airFiber/Airfiber-in-India/m-p...
EDIT: maybe a missing detail is that “AirFiber” and similar products are point to point comm devices. So each pair of AirFiber units is a 1.2+ gbps link, same as a laser set up. You may be confusing with an alternate PtMP radio setup.
I was wondering that myself. At least at a first pass higher bandwidth doesn't look like an advantage. The system you link to promises up to 2 Gbps, the same as a Ubiquiti AF24HD microwave system. Perhaps it's easier to put several of these on the same pole, and that's how you get the higher bandwidth?
Yeah, expect laser won’t work any better than RF in real world applicability. We’re doing symmetrical 2.5Gbps to CPE using RF and I can’t, in the least bit, see how and why a laser may make for a better solution fit in any scenario.
There’s a point when a hard wired fiber line just makes sense and it’s all postering to put out press releases about some bs laser solution that supposed to somehow be better than RF but less difficult/expensive than just running a wire.
> If you're like me and immediately asked "how can a line-of-sight laser system work when it rains?" The answer is "it doesn't." Atmospheric disturbances like rain, snow, fog, dust, and heat can all interfere with the light beam, and they have been the major limiting factor in this communication method.
In the past it has been proposed that a dense mesh network of FSO links could deal with interfering weather by dynamically routing around it. But that's a lot of lasers on a lot of moutain tops.
We are using microwave technology transmitting over 30 miles to our satellite ground station, and even in a hellish snow-storm we don't drop any more packets than usual. It actually was pretty damn impressive the first time I witnessed that. We're down in Antarctica by the way.
I'm somewhat surprised that you are not seeing reduced signal quality during heavy snow. Are you using high transmit power, or some extremely directional antenna?
Actually, considering the distances I would assume that high-entropy high-frequency noise near the laser such as people walking nearby, engines roaring, air currents, etc are much harder to deal with than the low-frequency predictable movement of the ship caused by large waves.
This technique works and is in fact quite simple. I suspect it's simply a matter of competing business interests that it hasn't been tried before.
Even near visible laser communication systems can operate at harmless power levels for miles, can be aligned by hand in many cases (you provide a "collection" dish to reflect the light in) and they're dirt cheap.
I actually built one as a project in my required EE classes for my CS degree back in 2002 using a pair of (rather expensive and overpowered) helium neon lasers. In 2002 and with these lasers the speed we had build was quite slow because of the switching time for the lasers, but the point was to build the hardware so we were good.
Given the absurdly low cost of these systems w/ modern components and their phenomenally low power budget, I'm not surprised.
If anyone wants to play around with a system like this, that is basically "open source" - and to see firsthand the difficulties of implementing such a system on a DIY budget, check the following project out:
I used to have a Motorola Canopy dish at home. It was much more reliable than cable or dsl. The idea is to have one omnidirectional tower, and directional dishes at each house. This scales beautifully, because you can simply add more and more towers to increase density (having directional dishes on one side fixes the crosstalk in both directions, up to a point).
It was doing 6mbit with better latencies than cable, back when cable was < 6mbit in my area. I have no idea what similar bandwidth microwave systems can push these days, but it worked fine in rain and fog.
I'm (naively) surprised this works in practice. It seems like the sending laser and receiver would have to be very immovable such that big wind gusts, or a large truck driving by, or large temperature changes in the structure holding the laser, etc, didn't cause enough vibration/movement to cause the laser to miss its target when several miles away. Is this a non-issue? Or maybe the laser spreads out enough at several miles it compensates for the movement?
53 comments
[ 3.2 ms ] story [ 80.5 ms ] threadAnyway, they seem to have a solution to that, but they won't disclose it. So it's anyone's guess, but it seems to have been taken into consideration.
(Edit: Looked it up; 2.4GHz is more readily absorbed by plants than 5.8GHz. Still, attenuation due to moisture seems to be a problem for microwave frequencies generally.)
Disclaimer: I work for Google but have no knowledge of this program. I'm reading about it for the first time just like you all.
But India is big on cellular. There's cell coverage almost everywhere, especially in states like Andhra Pradesh. You can get pretty decent speeds via 4G.
Also no idea in India but in the US a lot of times although their are existing power poles internet providers can't necessarily get permission to string their fiber on them. Same problem can happen with existing in ground pipes. And any trenching or installing new poles becomes very expense.
An alternative way to imagine networking an area is running fiber into the area then having a mesh network assemble itself building to building in town/city whatever. This has the added advantage of greatly reducing the power costs of the overall network for local communications.
The other issue is "long fat pipes." TCP is better tuned for it on machines now than it used to be but when you go over a long distance, the latency can decrease the usable bandwidth significantly. Of course, this seems to be more focused on last mile rather than back-haul. There are very few applications where a several nanosecond latency decrease matters.
http://www.lightpointe.com/airebridge-lx-1.html
There's a lot of competition for the non-visible spectrum at useful levels. It's only natural we'd start to move up. Also, as your wavelength gets smaller it gets cheaper to start compressing signals in ways that don't interfere over the same physical area.
The bandwidth of individual connections is only half the equation, and largely irrelevant imho. What matters for the network is the lack of interference. Laser connections can spider across a city without real risk of contaminating one signal with another. So even if the point-to-point bandwidth is the same, the total potential bandwidth of the installed network will be much higher.
EDIT: maybe a missing detail is that “AirFiber” and similar products are point to point comm devices. So each pair of AirFiber units is a 1.2+ gbps link, same as a laser set up. You may be confusing with an alternate PtMP radio setup.
This is about public ear. We all want to hear about some magic tech saving our internet’s ass. It’s not going to happen.
The solve is in the change of business models of ISPs, not in the tech that brings it to doorsteps.
We have the tech, we need a shift in how these companies make money.
There’s a point when a hard wired fiber line just makes sense and it’s all postering to put out press releases about some bs laser solution that supposed to somehow be better than RF but less difficult/expensive than just running a wire.
> If you're like me and immediately asked "how can a line-of-sight laser system work when it rains?" The answer is "it doesn't." Atmospheric disturbances like rain, snow, fog, dust, and heat can all interfere with the light beam, and they have been the major limiting factor in this communication method.
Well, I was thinking that combined with whatever navigation is going on as well. As a layperson, waves don't seem predictable to me.
Even near visible laser communication systems can operate at harmless power levels for miles, can be aligned by hand in many cases (you provide a "collection" dish to reflect the light in) and they're dirt cheap.
I actually built one as a project in my required EE classes for my CS degree back in 2002 using a pair of (rather expensive and overpowered) helium neon lasers. In 2002 and with these lasers the speed we had build was quite slow because of the switching time for the lasers, but the point was to build the hardware so we were good.
Given the absurdly low cost of these systems w/ modern components and their phenomenally low power budget, I'm not surprised.
http://ronja.twibright.com/
It was doing 6mbit with better latencies than cable, back when cable was < 6mbit in my area. I have no idea what similar bandwidth microwave systems can push these days, but it worked fine in rain and fog.
Implying that google's 20Gbit system isn't eye-safe?