This is slightly tangential, but it's Atlantic-cable-related, and I just learned about it. Let me set the stage. By the 1920s, analog fax technology was established and there were a variety of experimental and early commercial systems in use. One of the first uses of modems was to transmit faxes over the long-distance telephone network, turning a voltage into a pitch, and vice versa. The electronic marvel of the 1930s in the newsroom, aside from the teletypes, was that you could take a photo in San Francisco and have it published just hours later in New York.
But how do you get a picture across the Atlantic Ocean? Those analog fax systems suffer from noise and in the experimental systems of the 1910s and 1920s long-distance was impractical. Shortwave to get over the Ocean was right out. There were no transoceanic telephone links. But there were submarine telegraph cables. Morse code or Baudot telex. Not very fast with no amplifiers along the lines yet. But you could send a telex across the Atlantic ocean.
Enter the Bartlane transmission system. First, five photographs were developed from the original negative, at different exposure levels, using a conductive developer. So each point of the image is more or less conductive depending on its exposure. Each of these photos is scanned, as if it had pixels, and a 1-bit intensity value collected for for all five photographs, creating a five-level greyscale bitmap image on tape basically. (Five bits worked with existing Baudot code equipment.)
This enormous tape was then sent over the wire. On the receiving end a printer, with intensity controlled by the level indicated, would then selectively expose a spot on a photographic plate to light, as it stepped through the tape. Six hours to send a small image. But digital was the only way to send a photograph from London to New York overnight from its early trials in the 1920s until the 1950s.
Imagine the price of exclusive use of a transatlantic telex line for that length of time! Teleprinters back then commonly ran at 50 baud, so a 400x300 dot photo would take a little more than 40 minutes (120000 dots / 50 baud / 60 seconds)
The linked article mentions that the punched tape could be cut into several pieces then transmitted in parallel over several telex lines at once. At the receiving end the punched tapes would be spliced back together in correct order, and then used to create the photographic negative.
That's an interesting piece of historic technology. Thanks.
Cool. We take it for granted, but the intercontinental data (and voice) connectivity cables are an impressive achievement for humanity -- technical, and for global community. And globe.gl has me "seeing" for the first time a particular cable for which I have a much smaller personal story.
One of those submarine cables being severed accounts for one of only two software/hardware downtimes for a B2B startup's launch MVP, which had been deployed in critical production overseas.
Our station appliances at the facility needed to talk with servers (at the network-closest AWS AZ) multiple times a minute, for cryptographic security reasons. The appliances had a boot-time check for network connectivity (which saw extreme latency and packet loss as network unavailable). For reasons, the facility powered off all equipment at night.
One day, the facility said the appliances wouldn't "turn on", I quickly found they were having network problems, but facility insisted their LAN and Internet connectivity was fine. I was able to carefully SSH in (with terminal responsiveness like a bad dialup to an overloaded timesharing system), and relaxed the timeout on the check that was in a retry loop, and the stations came to life (albeit network requests during operations much slower than normal). The facility manager advised later that day that it turns out a submarine cable had been severed.
Indeed, especially for those few spots seemingly in the middle of the ocean where the cables converge it would be interesting to know what land masses (if any) are nearby. Right now this is very difficult to see.
I actually think there's a real unique and distinctive visualization caused by this decision. We're typically very geo-politically focused when we look to evaluate this sort information, but think about it without differentiating between land and ocean, without countries or borders. Think of nodes and edges in a graph, with the size/volume of edges weighing priority. Totally different way to view "earth".
This might be a stupid question, but is it implicit that these "submarine cables" are all fiber cables? If so, are they of similar or very different bandwidths? I would think there are cables of differing ages that are used for different things, or are all non-fiber cables taken out of service long ago, or are they just not included on the map?
Also, the "Polar Express" cable on that map is fascinating, but it seems like it's not planned to be completed until 2026.[1]
Copper communications cable were in use from 1850 until the 1980s, when they were replaced one by one with fiber.
Almost all the copper cables ever laid are still in the depths of the ocean, as it's not worth salvaging them except near shore. A few of the more recent copper cables are still operational, but they are used only for scientific research.
That’s true for communication cables though, so what about cables for electricity? That’s mostly what my original comment was about, that the title seems to imply any submarine cable is automatically a data cable.
There are high voltage submarine cables for electrical power - these are used to cross rivers and lakes, for example. International cables, like the ones between the UK and the European mainland, are used for load balancing.
As it's difficult technically to transmit electrical power over a subsea cable (and exponentially more difficult at the length increases), these cables have limited application. There's over a million kilometers (~600,000 miles) of fiber optic communications cable around the world right now, and perhaps a thousand or so miles of power cable, so "submarine cable" without any qualifier almost always refers to communications cables.
Electricity is just that much cheaper in some places than others. Bauxite is shipped far across the ocean to places with cheap electricity for the energy-intensive process of refining it into aluminum. Sometimes it makes more sense to ship the electricity instead. Here in Canada, I pay like 20% of what the Japanese pay and 10% what the Danish pay for electricity. If we could figure out a way to export our hydroelectric wealth to regions like Europe and Asia, it could help climate change. High-temperature superconductors don't get enough research, IMO.
they do get enough research, unfortunately they are some of the most complex physical systems and we still haven’t figured out why exactly they superconduct (i used to work on this).
There is (if it has not been scuppered by brexit) Icelink, a project to connect Iceland to the UK. The rationale, as usual, is mismatched supply and demand, where the supply is green geothermal power.
Not all of these lines go across the ocean. If you're trying to move power between São Paolo and Beunos Aries, it might be cheaper to lay a sea cable than to deal with all the terrain, land rights and borders involved in an overland route.
I strongly doubt there are any non-fiber submarine cables. Fiber's main strength is range; trying to run signals over electrical conductors over hundred-km-plus distances is not cost-effective and hasn't been for decades.
(It's been done historically, but with very heavy cables, very high endpoint power levels, lots of repeaters, and really poor throughput.)
> I strongly doubt there are any non-fiber submarine cables.
not for telecom, not anymore, but a number of submarine cable laying ships are equipped specifically for laying high voltage power cables to offshore wind farms. And for things like high voltage cables linking medium sized islands to a nearby mainland (eg: Vancouver Island, BC to the continent).
After submarine telecom, the biggest uses for such vessels are offshore power and the oil/gas industry.
> I strongly doubt there are any non-fiber submarine cables.
I strongly doubt electrical cables are fiber cables for example :-)
On a more serious note, I was thinking that there are perhaps copper cables used for telephone communications that are still used simply because they haven't yet failed, but I assume fiber has perhaps been the standard for backbones long before it became something people have in their homes.
Wouldn't it make sense for some electrical cables to have some fiber in them? When you're already laying them, you could get two services for the price of one.
2) AFAIK most submarine cables are systems — e.g. a ring the connects various places so that there are two paths between points — so trying to plan one to that co-opts existing bits of fiber piggy-backed on power probably wouldn't be easy. Also you'd have the issue of dealing with different generations of fiber, standards, etc.
[1] On reading this it turns out that modern fibers actually carry power as well so that repeaters en route can be powered.
I suppose it would, I guess I'm just wondering for how long fiber cables have been the standard, and if there are any operational cables from the time before that. I'm pretty sure there are plenty of cables dedicated for one purpose (e.g. electricity only) that do not contain fiber simply because there wasn't a need or requirement when it was planned.
edit: Here's an example:
> The world record from 1994, for the longest HVDC cable with the highest capacity, was equalled in 2000 with the successful completion of the Swedish-Polish HVDC transmission link between Sweden and Poland:
> 1. Stretching over 250 km, it is one of the two longest submarine HVDC cables ever laid.
> 2. At 600 MW, the cable is one of the most powerful HVDC cables in commercial operation.
> 3. At 450 kV DC, this cable together with the cable between Sweden and Germany, has the highest voltage rating of all existing submarine HVDC cables.
This cable doesn't seem to have any other purpose than transferring electricity, and it was installed "only" 21 years ago.
Optical fibres are also pretty good sensors so it's not unlikely that some fibres are added just for sensing purposes (they are eput into lots of bridges and other things for example). That said it highly depends on the length of your electrical cables if putting in fibre for communications makes sense. Typically (grossly simplified) you need optical amplifiers when you go significantly over transmission distances of 100km so they need power supplies, take up quite a bit of space etc..
They probably already had some inside but not for general use, but to exchange data between the two ends of the cable. Modern power grid is smart, and it has data connection along side power, they do that also on overhead power lines (the upper cable beside the three phases serves as a protectoin from lightning strikes but also has a fiber inside). In practice the power network also has a completely isolated LAN (for obvious security reasons) to manage the whole infrastructure. At least in my country it's done like that.
I highly doubt it's cost-effective. These things are complicated and have specialized equipment. The endpoints also very different for data and electricity.
In almost all conceivable contexts, no. If you've ever done any work with electrical code, you are familiar with the spacing requirements between electrical conduit and data wires to prevent magnetic interference: at least a foot apart, and perpendicular crossings if crossings are necessary. The voltage on underwater electrical cables are quite high, as is the bandwidth on underwater data cables, so you definitely can't have them mixing if you want anything near optimum results on data transfer.
Some coaxial lines were laid until the 1980s. Seems likely that at least one small island somewhere has a copper submarine cable supplying their data/phone needs still. But it's decidedly obsolete, yes.
Doesn't a lot of these cables pre-date fiber? They had these in WWII. I remember the Gilligan's Island episode where the cable washes up in the lagoon and the professor was able to connect up a phone so they call around the world -- of course Gilligan screws it all up at the end.
But weren't the British Empire's undersea telegraph lines all copper? They lay the first transatlantic cable in 1858, connecting Ireland and Newfoundland. By 1911 their worldwide undersea network was completed, with so much redundancy that 49 cuts would have been necessary to isolate the United Kingdom from its empire.
https://en.m.wikipedia.org/wiki/All_Red_Line
None of that stuff is still in service. (At least not telecom.) They require periodic, expensive maintenance, and fiber outclasses copper for long-distance signatures to a much greater degree than for the last-mile ground traffic for which copper is still used.
Current fiber-containing submarine cables also have electrical lines to power inline amplifiers. Even fiber signals can't travel those kinds of distances without repeaters.
When I lived in the Cayman Islands, the primary submarine cable (I forget its name) was fiber, and I had better net there than I had in the US. It went down for a week or two one time, and the entire island's net traffic was relegated to a much older copper cable that ran to Jamaica. Cayman is a tiny market, but it still crippled our net because the bandwidth was tiny; it was like a couple T1s for the whole population.
Yes operating a non fibre cable is too expensive and does not make sense. Fibre cables on the other hand have remained operational over a long time. AFAIK there are no fibre cables using electro optic repeaters operational anymore. But the earliest fibres which used optical amplification are still operational. This is the beauty about fibre communication. You largely only need to upgrade the endpoints to increase your throughput.
You can find the FLAG cable referenced in the article on this map. The easiest landmark for finding it is the narrow section of the Malay peninsula. Stephenson spends some time in the article covering the overland construction there.
Wow, not sure how useful that is (standard caveat about choropleths and density vs raw numbers) but very impressive! Even the text overlays work nicely.
It’s amazing to see that we are essentially just having one huge LAN party. I wonder if the inventors of the phone or internet could have ever imagined this. Now with low orbit satellites connecting the world too, the innovation is incredible on this planet.
I'm a bit curious about something that I've never thought about before -- how were telegraph signals routed? Was it just a one way street, or did they have a frequency range to work with?
This was long before there were any electronic devices to amplify or otherwise manipulate the signal. The telegraph signals were just short electrical pulses representing Morse code, with a positive voltage for a dot and a negative voltage for a dash. The lines could only be worked in one direction at a time for many years, but eventually circuits were used at each end to "duplex" the cable, allowing messages to be sent in both directions at the same tim.
Little need, it's essentially all government run bases down there and they get their internet through various satellite links. You could maybe run one to McMurdo Station since it's on solid land and by a bay but there's just not enough data to justify the cost.
I wonder if UK can use the fact that most US - EU links go through UK as a leverage.
EDIT: answer is yes:
"LONDON (Reuters) - High-speed sub-Atlantic cables may force banks to keep their armoury of currency trading hardware in London for some time, even if the dealers themselves go elsewhere when Britain leaves the European Union." [1]
A good idea in theory, but I believe phase 2 of our flawless and perfectly thought out Brexit execution plan is to sever all of those cables to prevent unwanted "foreign" bits from coming in and taking up British bandwidth.
We're then going to individually renegotiate brand new communication transport/protocols with each foreign nation separately. On far better terms, over the next 20-25 years.
I think the current plan is to use a different breed of carrier pigeon for each country so that we can tell where the data is coming from and where to send our responses back to.
Actually bit weird how there is not a single cable going from Ireland to France (or elsewhere in the continent). I guess going through UK works, but there are so many cables all around the place that you'd think one of them would cross that gap.
These sorts of projects are always fascinating to me. The hard part of these sorts of visualizations always seems to be acquiring the data in the first place. I would have no idea where to begin looking for a record of exactly how all of the underwater fiber optic cables are positioned. I wonder whether people come across this sort of data and think to do something with it, or whether people start by deciding to make this visualization and then go forth and search out the data.
It's interesting that there are so many cables running through the Suez Canal. I wonder if it's because of geopolitical reasons for avoiding going over land through the Middle East.
There's another one that appears to run through the Gulf of Mexico to the US as well (looks like AL or LA to TX), so I suspect it's surprisingly cheaper and/or easier to get policy approval.
Well, it's also easier to run a cable underwater. As far as I know, the cables are literally just laid down there. You can't just lay a cable on land. You've got to bury it or elevate it on poles. I wonder how they deal with dredging the canal, though.
per another comment[0], it's cheaper to just stay in the water unless you specifically want to attach to a land network. I don't have the knowledge the other commenter seems to have, but that certainly seems believable to me.
Likewise, what's the "benefit" of spreading the cables out by a few kilometers when space allows? If you zoom in you can see all the cables scrunched through the canal, but on more than one occasion they are fanned out to spread evenly across the given space. Likely it is only a visual eyecandy in the way this map was created, and not a true indication of the cable formations... but I tried looking for some answers on https://www.iscpc.org and gave up. I would assume once a suitable cable path is found, most subsequent cables would be laid right alongside for higher chance of success avoiding coral/trenches/etc.
edit: The more I look at the map, the more it becomes clear these lines do not indicate true cable locations. The lines are too perfect... even for fiber optics. The ocean floor is a treacherous beast!
Land almost everywhere comes with ownership issues which can be problematic for installing cables. The ocean is generally unambiguously nationally owned, and so it's much simpler to seek permission from a single source (government) than many sources (owners).
Is there any map of land cables ? I know more about the few submarine cables between south and north america than how the internet actually arrives to inland cities of my state.
working at a tier one telco. we have very precise maps of our ground fibers. physical location is usually down to +-10cm. length of fiber even takes into account any coiled up parts. if a contractor wants to dig anywhere, then they have the responsibility to query a national database before digging. if they break our cable, then we have active monitoring (reflecting a laser in a spare fiber, off the break, calculating length to break) this interfaces with our map software, so we can tell within half a meter where the cable is broken. in most cases we end up suing the contractor for the cost of repair and any SLA payments.
to my knowledge no public maps exist, that will give you a full picture.
As far as I can tell, there aren't that many cable running through the actual canal, instead they cross over Egypt before the canal starts or terminate in Suez: https://ibb.co/NyVXtST
I'm not sure about the Suez, but is the reason why there are almost 0 cables running on the African continent and instead there being several cables laid around the coast.
There’s some interesting military history and tech regarding submarine cables. The first act of WW1 was to sever a cable, for example. And during the Cold War the Americans planted bugs on Russian cables and were eventually exposed by a spy. Nowadays the Russians have some interesting capabilities that can be inferred.
Some years ago I was contacted by the descendants of an English engineer who was working at the Long Island NY cable station of the German Atlantic cable via the Azores at the start of WWI.
Back in the 1950s a family member had recorded him on open reel tape reminiscing about his experiences, and this was later converted to audio cassettes. They sent me the tapes, and I cleaned up the audio for them, and they gave me permission to post the transcription and audio on my site.
Turned out he was actually watching the signals coming in from a German operator when the cable was cut by the Brits, and he talks about this on one of the tapes.
- Cocos Islands has their own cable and they don't even have 600 people;
- Saint Helena has its own cable (googled to find out it is very recent and will start offering 18Gbps in 2022);
- French Polynesia seems well served.
Tristan da Cunha and Pitcairn probably will never get their cables.
I suspect that some of those places host strategically important communications facilities. Perhaps spy satellite downlink facilities, perhaps GNSS ground-segment base stations, perhaps earth stations for other communications satellites.
Oddly, those two specific islands don't. Saint Helena is specifically courting such business, but I don't see any indication that they've attracted anyway. But several other small islands are definitely fibered-up on account of their satellite facilities.
There’s not a lot of content to be exchanged between Siberia and Alaska. And running long stretches of fiber over land is actually more costly and subject to failure than under the sea, so going from Moscow -> Vladivostok (or something) over land, then going subsea, and then Alaska -> West-coast US over land, is not really beneficial.
For me the most fascinating thing about this is vast distances through Pacific covered with the cables. One would assume it would be easier to cover the distance around Pacific and not directly through the middle of it, but it seems those milliseconds are truly precious.
I assume these are all communications cables, but I'm curious if anyone knows what the logistical issues would be to laying high-capacity high-voltage DC lines across oceans? The idea being that as renewable and specifically solar power becomes a larger percentage of electrical generation, it'd be really nice to be able to buy and sell power between the Americas, Europe, Asia, and Africa. It's either that or use a lot of batteries, and power lines seem like they ought to be the more efficient, effective solution. HVDC losses can be reasonably low even for thousands of miles* [1].
I'm thinking of having undersea links, for instance, between Brazil and Western Africa, or Canada to Greenland to Iceland to mainland Europe, or Alaska to Russia.
(Wikipedia shows a number of short-distance undersea HVDC links in Europe [1].)
* "Depending on voltage level and construction details, HVDC transmission losses are quoted at 3.5% per 1,000 km, about 50% less than AC (6.5%) lines at the same voltage."
Huh, surprised me to see how many run along the coast. I would've thought it would be cheaper to run cables by land if they weren't going to cross the ocean. I can see how that might not be the case though, as you can't just lay a cable down along the ground on land like you can in the ocean.
It's interesting that most countries have a single entry point (near the capital city), or only a few. One big exception is Chile that has a cable that goes along the cost and reach many cities. I a similar pattern in the east of Brazil and both sides of New Guinea.
All the way back to the early days of the cable industry in the 1800s, it was easier and safer to lay a cable along the coast of undeveloped countries than to try and run a landline on poles across rugged terrain. Both coasts of Central and South America had telegraph cables running down them, with spurs for inland connections. Same with Africa.
Here's a map on my cable history website showing the network in 1902:
https://atlantic-cable.com/Maps/1901-Berne-Map-BPL-Leventhal...
its interesting that there seem to be cables connecting locations on the North Alaska caost, but nowhere else. I would have thought those locations would be too remote for land based connections, unless there is fiber running from Prudhoe Bay along the oil pipeline south.
Anyone else find it utterly fascinating that we have undersea fiber running to Svalbard, the galapagos and keeling island? These are not highly populated areas, and yet look at how abysmal rural internet penetration is in the usa.
I found a couple that have interesting or creative names - Bifrost, Polar Express, Apollo, Grace Hopper, Amerigo Vespucci. Most of them are pretty dry by comparison, which feels like a missed opportunity.
What happens when cables need to cross an isthmus like the Suez? Looks like they don’t literally go through the canal from Suez to Port Said but they run pretty close by to minimize distance over land. Most connections appear to run from Suez to Alexandria and from Zafarana to Abu Talat. [1]
I would love to be able to plug in two locations and visualize on a map the path my packets would take travelling from here to there. A sort of visual traceroute.
One problem I've had with existing tools is that it's hard to find a server in a given place, even if you know that place has internet. For instance I was facetiming with my partner who is abroad in africa right now, but I'm not sure of a server to [traceroute] against in the country she's in.
One suggestion would be to use a better globe. I want to see where in US the entrypoints are, for example, but it just become a blurry mess when zoomed in.
I'm surprised the Greenland Connect cable connects Newfoundland to Iceland by way of Greenland instead of just terminating at Greenland. I wonder how much non-Greenland through traffic the cable gets. It'd have been cheaper to skip the Greenland<->Iceland leg, right?
It's also interesting how it looks like it's easier to go around the west coast of Africa than to run cables overland across the Sahara.
At ~1 cable per 1M pop they are probably leaving most countries behind them in a cable/person ratio, being topped only by the transit hubs in the pacific.
Protecting the sea lanes of communication is actually a core US Navy mission (and probably high priorities for several other world Navies as well). Particularly the major choke points: Strait of Gibraltar, Strait of Malacca, Strait of Hormuz, the Panama Canal, Suez Canal, Bab el Mandeb, Turkish Straits, and the Cape of Good Hope.
- Guam is a significant data/telecoms hub for the western pacific region
- There are some short cables installed in areas that don't seem to make much sense, Such as the one in the gulf of Mexico, or between Italy and Croatia
- Many extremely sparsely populated islands and archipelagos have their own undersea fiber, such as Svalbard.
Interesting fact if you're in latam/caribe... if you're down here you have situations like DDoS saturation on MAYA-1 (caribbean side) and then you have to make some phonecalls and re-route your BGP over the PAC (Pan-American Crosssing on the pacific side). Fun shit to do getting woken up by your boss at 1am!
The ship only hit the grounds on the sides where it's less deep, but the cables probably go through the middle area where it's the deepest. So there wasn't much of a chance for the ship to damage any cables.
There's land adjacent to many of these cables. It is cheaper to bury cables in water -- there's a lot fewer shovels needed and fewer property disputes.
It would be good if there's an option in which we can increase its brightness. It's too dark, I can't properly distinguish each continent, and some parts of the world are missing.
I was surprised by how many lines seem to just run along the coasts of land. Is it really cheaper to lay and maintain fiber in the ocean then over land along the coast?
I would think so given overland is more likely to hit right-of-ways owned by entrenched telcos - who tirelessly work to screw-over any comm progress that doesn't enrich them as much as they want.
Past that there's endless interests that have to be negotiated with. And any part of the cable has to travel above ground will be more vulnerable.
Along the US west coast, it surprises me that Seattle, Portland, Los Angeles, and San Diego each appear to have more submarine cable connections than San Francisco.
Given the population is always under 10k, and it doesn't serve as an interconnection point for any further points along the way (unlike e.g. Hawaii or Guam), it's just not worth it, apparently.
I also found that interesting, found a comment on Reddit thread linked here:
submarinecablemap.com's FAQ
*Are the submarine cable paths shown the actual route taken by the cables?*
No. The cables routes on our map are stylized and do not reflect the actual path taken by the various systems. This design approach makes it easy to visually follow the different cables and the points at which they land. In real life, cables that cross similar areas of an ocean, take very similar paths. These paths are chosen via comprehensive marine surveys which select routes that avoid hazardous conditions that could potentially damage a cable.
But how do you get a picture across the Atlantic Ocean? Those analog fax systems suffer from noise and in the experimental systems of the 1910s and 1920s long-distance was impractical. Shortwave to get over the Ocean was right out. There were no transoceanic telephone links. But there were submarine telegraph cables. Morse code or Baudot telex. Not very fast with no amplifiers along the lines yet. But you could send a telex across the Atlantic ocean.
Enter the Bartlane transmission system. First, five photographs were developed from the original negative, at different exposure levels, using a conductive developer. So each point of the image is more or less conductive depending on its exposure. Each of these photos is scanned, as if it had pixels, and a 1-bit intensity value collected for for all five photographs, creating a five-level greyscale bitmap image on tape basically. (Five bits worked with existing Baudot code equipment.)
This enormous tape was then sent over the wire. On the receiving end a printer, with intensity controlled by the level indicated, would then selectively expose a spot on a photographic plate to light, as it stepped through the tape. Six hours to send a small image. But digital was the only way to send a photograph from London to New York overnight from its early trials in the 1920s until the 1950s.
http://www.hffax.de/history/html/bartlane.html