The emphasis on secrecy is rather overblown IMHO: there's an entire company (not mentioned in the article!) whose sole business is building/maintaining these networks and they make plenty of information public on their website http://www.mckay-brothers.com/product-page/
I remember reading in a book once that during the cold war, the Soviets used microwave communication partially to get around the Allies snooping on their RF communications. So, the Allies would position satellites such that they were in the line of sight of the microwave links so they could listen in on the extra transmissions that overshot the receiving end.
I wouldn't think it would be too hard to use a gyro-stabilized gimbal to keep the antennae pointing in the right direction. Most drone quadcopters used for aerial filming already do this.
I think the hard part is the salt-water environment. Any bearing wants lubricant and putting anything that complicated in salt water sounds like a bad idea .
Gimbals can only do so much. In the mid-Atlantic ocean wave heights are between 6 feet and 20 feet. That's not going to feel like much if you're on an oil tanker, but a relatively small barge is going to move up and down a lot.
They don't need to be pointed at each other. The antennas aren't very high gain which means they will still hit around the target. But yes it is difficult.
Gain, even though this isn't a correct analogy, is like the difference between a laser and a flash light. The laser is a very narrow beam, which is a high gain antenna, and the flash light is very low gain.
What I'd be worried about are waves and storms knocking out the network. I would assume that microwaves would go through a wave since they are so nigh energy but they are used by the navy for wave-detection (radar is basically microwave). Moving around isn't too bad because they only need to be on at about 50-60deg. Heavy waves would knock out the network like you are worried as well since I think microwaves would be able to punch through the ionosphere most of the time (UHF is usually used for sat-coms by the military).
Microwaves have a penetration depth in water of just a couple centimeters. The higher the frequency the less. I guess K-band is probably already below one centimeter.
It would probably be more like a large vertical tower with most of its mass floating underwater than a flat barge with an antenna mast on the deck. This would mean even large waves would hardly affect it.
The bigger problem is you'd need at least 70 of them to cross the atlantic.
The microwave transmissions stay in the analog realm to get their signal boosted at the intermediate towers, so you don't lose time in D/A-A/D sampling conversions like in the typical fiber signal booster (there are optical signal boosters for fiber under development).
> The most you'd be able to push through the air is a few Gigs
Total capacity isn't the key, here. It's latency. And you're really exaggerating the impact air has on the speed of radio propagation. When you take into account the physical latency of the distance to satellites in geo stationary orbit, it's easy to see that a microwave network can be faster. If you're using non-geostationary satellites for communication, you have to negotiate the infrastructure changing out from under you constantly. Reliability is key for these networks. Waiting milliseconds for your physical layer to switch to a new satellite means you lose. QoS also adds significant latency. It's not a viable choice for extremely low-latency networks.
As an aside, you'll probably get a more positive response around here if you endeavour to keep your comments from getting so emotionally laden ... saying bullshit multiple times doesn't do anything to strengthen your argument.
A quick bit of research shows that the speed of light in air only 0.3% less than the speed of light in a vaccuum. So that works out to 298,896km/s. This is significantly faster than the speed of light in an fibre optic cable (around 200,000km/s). Combine this with the ability for a microwave link to have a shorter path than optical fibre, and it's easy to see why a microwave link can be lower latency.
Nobody is disputing the fact that fibre optic can handle much higher bandwidths, but that is irrelevant in this application. The amount of data required to make a trade is probably in the order of kiloBytes or MegaBytes, easily handled by the few Gbps limit of microwave links.
The reason why satellite internet is so terrible is because a geostationary satellite sits 22,236 miles above the surface of the earth. So for a roundtrip journey, a packet of data has to travel nearly 45,000 miles.
I so very much enjoy these kinds of stories. Even if there is debate over the actual speeds, etc., the general info is great to learn about! This is the kind of stuff that I show to my pre-teen daughter, so that she can see that there are wondrous, technical things going on all around us; that to some kids might seem fantastical and bordering on science fiction. And hey, if she ends up going into a STEM-related career because of exposure to these cool types of articles, that's not so bad either. ;-)
To answer the question "why use your own microwave network?" the author provides 2 reasons:
> The first reason is somewhat obvious; if you have your own network connection, it's usually easier to guarantee things like security, quality of service, bandwidth, and other factors that businesses value highly. The second reason, as we've already alluded to, is that microwave networks—somewhat surprisingly—can have lower latency than fibre. With some advanced networks, that latency is only a few microseconds slower than the speed of light. Fibre can be pretty quick over short stretches, but it soon starts lagging over longer distances, such as between two stock exchanges or a multinational's offices.
The 1st reason is really off the mark. We know how to securely transmit data over insecure networks, and quality of service and bandwidth can be better handled by service contracts with the network provider. He states "other factors that businesses value highly" which is vague but could possibly be interpreted to include competitive advantages - yet cannot in light of the fact that the 2nd point addresses competitive advantages. By owning the fastest microwave network, you can execute faster than your competitors and directly use that for monetary advantage. So his 1st reason is completely waste of space.
So back in the day, perhaps over a decade ago, I remember stumbling upon a textfile that was along the lines of "interesting IP addresses on the net" (does anyone know what I'm talking about? I can't find it since.) and one of them was a very curious telnet console that had all sorts of settings. Googling the terms, it looked like it was the control panel for a microwave antenna. I was too scared to mess with it though. I only remember exploring through the menu options. I never figured out where it was or what it did, but I remember it was surprising to me at the time that standalone hardware would have a telnet terminal.
Fast forward to today and we have a huge multitude of IoT devices with telnet interfaces (a ton of webcams) and default (and hardcoded) passwords... Now we have the Mirai botnet...
> "One proposed method of stopping incidents like the Knightmare, and generally to limit HFT from wading too far into ethically grey areas, is by leveraging blockchain tech. For example, front-running—sneaking in trades a few microseconds early based on advance knowledge of pending changes to the market—could be stymied if the market adopted a slightly slower, transactional, blockchain-style ledger."
Adding another layer does not stymie anything - it might add latency but this will be exploited by those with the most resources resulting in [additional] advantages (e.g. on top of the microwave network advantage, a race to write to the blockchain).
The refractive index of optic fiber is in the 1.4 or 1.5 range which means light travels at 0.66-0.71c. Said refractive index for radio waves is practically 1 in the troposphere so microwaves outrace light.
With recent work in millimeter wave radios I am sure you could get the capacity of this network much higher than it is now, although rain fading might be a problem.
35 comments
[ 5.3 ms ] story [ 92.7 ms ] thread[1] https://sniperinmahwah.wordpress.com/2014/09/22/hft-in-my-ba...
[2] https://news.ycombinator.com/item?id=8354278
How would that work when the dishes need to be pointed at each other and the barge will be bobbing back and forth?
Does the signal spread enough to cope with that?
Gain, even though this isn't a correct analogy, is like the difference between a laser and a flash light. The laser is a very narrow beam, which is a high gain antenna, and the flash light is very low gain.
What I'd be worried about are waves and storms knocking out the network. I would assume that microwaves would go through a wave since they are so nigh energy but they are used by the navy for wave-detection (radar is basically microwave). Moving around isn't too bad because they only need to be on at about 50-60deg. Heavy waves would knock out the network like you are worried as well since I think microwaves would be able to punch through the ionosphere most of the time (UHF is usually used for sat-coms by the military).
The bigger problem is you'd need at least 70 of them to cross the atlantic.
That actually doesn't sound too bad. I would've expected that number to be much higher.
I wonder how that cost would compare to the thousands of miles of cabling we currently use.
Total capacity isn't the key, here. It's latency. And you're really exaggerating the impact air has on the speed of radio propagation. When you take into account the physical latency of the distance to satellites in geo stationary orbit, it's easy to see that a microwave network can be faster. If you're using non-geostationary satellites for communication, you have to negotiate the infrastructure changing out from under you constantly. Reliability is key for these networks. Waiting milliseconds for your physical layer to switch to a new satellite means you lose. QoS also adds significant latency. It's not a viable choice for extremely low-latency networks.
As an aside, you'll probably get a more positive response around here if you endeavour to keep your comments from getting so emotionally laden ... saying bullshit multiple times doesn't do anything to strengthen your argument.
Satellite internet has high latency not because "pushing through air" but because geostationary satellites are very far.
Nobody is disputing the fact that fibre optic can handle much higher bandwidths, but that is irrelevant in this application. The amount of data required to make a trade is probably in the order of kiloBytes or MegaBytes, easily handled by the few Gbps limit of microwave links.
The reason why satellite internet is so terrible is because a geostationary satellite sits 22,236 miles above the surface of the earth. So for a roundtrip journey, a packet of data has to travel nearly 45,000 miles.
> The first reason is somewhat obvious; if you have your own network connection, it's usually easier to guarantee things like security, quality of service, bandwidth, and other factors that businesses value highly. The second reason, as we've already alluded to, is that microwave networks—somewhat surprisingly—can have lower latency than fibre. With some advanced networks, that latency is only a few microseconds slower than the speed of light. Fibre can be pretty quick over short stretches, but it soon starts lagging over longer distances, such as between two stock exchanges or a multinational's offices.
The 1st reason is really off the mark. We know how to securely transmit data over insecure networks, and quality of service and bandwidth can be better handled by service contracts with the network provider. He states "other factors that businesses value highly" which is vague but could possibly be interpreted to include competitive advantages - yet cannot in light of the fact that the 2nd point addresses competitive advantages. By owning the fastest microwave network, you can execute faster than your competitors and directly use that for monetary advantage. So his 1st reason is completely waste of space.
Adding another layer does not stymie anything - it might add latency but this will be exploited by those with the most resources resulting in [additional] advantages (e.g. on top of the microwave network advantage, a race to write to the blockchain).
With recent work in millimeter wave radios I am sure you could get the capacity of this network much higher than it is now, although rain fading might be a problem.