Thanks for posting the 1901 map. I'm amazed at the extent of the network back then.
Of course, in 1901, each of those lines was measured in (low double-digit) words per minute. Now the metric is more like feature films per second. And the word "cable" once again means a communication line, not a message transferred over one.
I do find it impressive that Australia was more or less connected to the rest of the world by telecommunications in 1872 once the overland telegraph line was completed. The US was only connected reliably connected to Europe about six years earlier in 1866. So not a bad effort given the vaster geographic distances to span between Europe and Australia.
What's amazing to me is that these cables were laid without a lot of fanfare. Seems to me if I were Level1 or whoever else and had just completed a 10,000 km cable across an ocean, I'd make sure the press release saw some circulation.
The same thing with satellite launches. I feel like these are fairly major feats of engineering that go basically unheard of unless you're in certain small circles.
Modern telecoms cable is some amazing stuff. These days there's no metal in the entire cable. Instead, it's all lasers through glass fiber, and when the signal attenuates, it gets boosted by passing through sections of doped fiber that lase using power from another beam, adding power to the signal. And we can do this just as easily as the people back then dropped telegraph cable. For me, at least, there's always something to be impressed by.
First, most EDFA (Erbium Doped Fiber Amplifiers) need power to operate since the lasers that perform the pumping action are local to the doped sections of fiber.
Second, most submarine cables make extensive use of Raman amplification which leverages the transmission fiber as a gain medium via forward or backward pumping of lasers with frequencies chosen to produce Raman gain.
In both cases, you need an amplifier every 70-100km (150km in some cases) in order to have a usable signal on a long span.
Raman amps are pretty cool but tricky. For one thing, the power required is pretty high. +1W dBm isn't uncommon for certain Raman amps. At those power levels, you need to be careful you dont cause burns to personel when they are working on the equipment.
The power to run the lasers is usually carried on conductors that are part of the cable itself. The voltages are somewhere in the several kilovolt range so laying submarine cable is also kinda like laying a high-volage transmission line across the ocean.
It's so weird because I think it's such an impressive feat and one of those things that I feel should be technically really challenging yet it was done so long ago. Sometimes I learn things and am shocked because I either thought accomplishing the thing would be easy and it turns out to be insanely hard or vice versa. This is the latter.
> I'd make sure the press release saw some circulation.
A bunch of it was laid by ATT, which then got bought by Tyco.
The entry of that NJ HQ building was wild: It had pics of the ships used to maintain the cables. (When the cable needs to be pulled up in the middle of the atlantic, it's a bit of a challenge.)
IIRC, Tyco had 9 ships. If you glanced quickly, they looked like battleships with all the equipment on them. It was quite the operation.
It's getting kind of dated now, but Neal Stephenson's "Mother Earth Mother Board" is great reading if you're interested in both the history and the process of laying submarine cable: http://www.wired.com/wired/archive/4.12/ffglass.html
It's not exactly what you outline, but The Geek Atlas by John Graham-Cumming lists many places that would fit the description "hacker tourism." It's a travel guide... but it's published by O'Reilly!
Wow, there are two cables running from northern Norway to Longyearbyen, on the island of Svalbard! According to Wikipedia, "Longyearbyen has approximately 2,060 inhabitants (at the end of 2007..." The total population of Svalbard is not much more.
The reason for that is there's a an important satellite ground station on Svalbard - it's in an ideal place to communicate with polar-orbiting satellites:
Photographer Christian Hougue produced a really beautiful series on this ground stations called Arctic Technology. He mentioned that the air is some of cleanest in the world, which is a real boon to the equipment they're running.
While reading a bit about these cables I found this unusual fact on Wikipedia:
> Because the [TAT-8] cable was the first fiber optic cable and not coaxial cable, the electrical interference shielding for the high voltage supply lines was removed. This removal did not affect the fiber, but it did cause feeding frenzies in sharks that swam nearby. The sharks would then attack the cable until the voltage lines killed them. This caused numerous, prolonged outages. Eventually, a shark shielding was developed for the cable.
Interesting factoid, Sharks have a fairly unique sense that most animals don't, they can sense fairly faint electrical fields that living creatures produce (electroreception) that they use for "targeting" their amazing jaws, finding prey hidden in seafloor sand, and possibly even navigating in Earth's magnetic field.
This is especially important when they are about to attack some prey. Due to the structure of their jaws that actually unhinge and push forward during an attack, they lose sight of their prey and use this sense to hone in their bite.
I remember watching video from an experiment where Sharks were fed live fish and raw meat, and the sharks kept missing the meat but almost never missed a live fish.
I cannot help but salivate at the dense links connecting US/Europe and US/Asia.
Here in Kenya we only have 3 cables. Since last week the internet has been very slow after the main cable was severed by a ship anchor. The second cable seemed to have problems upstream so all the traffic has been routed to the third cable.
I found it absolutely fascinating how some continents use underwater cables to connect different locations within a continent (Africa, notably, seems virtually surrounded), while others seem to only have significant outbound connections (North America, Australia).
I can only guess it has more to do with whether you cross national boundaries (since even the US has underwater cables bridging the mainland to Alaska) than land features.
I would assume it has more to do with how many intermediate places within the continent you want to connect to. If you're connecting two cities on the African coast, and there's really no infrastructure in-between that needs (will pay for) connection, and the roads in between are poorly maintained, it's probably more economical to go by sea. Someplace like the US, if you're running cable from say Boston to NY, first of all, you also want to wire up all the towns and cities in between, and you've got interstate highways so your service trucks can easily get to all parts of the cable, you've got existing utility tunnels you can use, etc. In that case, going by sea is an unnecessary expense.
If you take a look at an Internet penetration map[1] you will see that most of the cables around Africa have only a few big "customers": Morocco, Egypt, Nigeria and South Africa. If you compare that map with a GDP map you will also see why: poor countries, without terrestrial infrastructure and some with unstable political situations.
I noticed quite a cluster on the South American coastline - Brazil-Argentina coastline looks like it's got 4 or 5 that go nowhere but that coastline. You'd think Brazil was stable and prosperous enough that they'd go with landlines.
The population is strongly clustered on some big cities in the coast, the roads aren't super safe, and the terrain around the coastline is not flat, so it's probably just cheaper to lay them out in the ocean.
I see no cables actually get routed through the Suez canal, even though many cables come up the Red Sea. The same is true for the Panama canal. Is it because they are too shallow? Maybe too concentrated?
I think it's due to accessibility and the way the canals work. The Panama canal itself is actually a series of waterways connected via gates. Allowing ships to be raised and lowered as they need to change elevation. I doubt that'd be very suitable conditions for laying cable. Far easier to make connections to land based routes on either side of the isthmus.
As the Stephenson article linked elsewhere in this thread says, cablelayers hate and fear shipping lanes. Anchors kill underwater cables dead. Additionally, canals have to be dredged fairly often, which are similarly incompatible with seafloor communications cables.
I've always wondered where they are. Can you imagine just how long it took to wrap a wire around the whole world a few times over?
The thing that shocks me is how recently these were added. I was checking out the cables in Australia, expecting them to have been added in the 60's or something....however most are around the year 2000-2010.
Hmm, then what happens to outdated cables? Do we pull them up, or do we just turn off a switch at the shore and leave the rest in the ocean? Do the cables contain any substance that might pollute the seabed as the casing degrades away?
Probably, but we dump orders of magnitude more crap (both in volume and in hazard) in the ocean every day.
Plus, the impact from burning the fuel it would take to have a ship pull up a cable would be thousands of times worse than anything the cable might do.
In certain cases, e.g., the cables AT&T laid between Hawaii and the mainland US, they were decommissioned from active service and then used to power a network of undersea sensors used for research by the University of Hawaii:
I can't imagine what materials they would be made of that would be particularly harmful. My sense is they're mostly constructed of things like rubber, plastics, other polymers, steel, copper, and of course a little glass fiber.
The APNG-2 cable (http://en.wikipedia.org/wiki/APNG-2_(cable_system) ) reused part of the PacRimWest cable. They literally pulled some of it up, respliced it and moved the end-point to Papua New Guinea.
Given the cost of such an operation, I don't think you can get some profit even if they were made of gold.
As time passes, most of the cable is buried deeper and deeper under debris that falls on the ocean floor. This means that you cannot just pull it up, you would have to dig to get it out.
10-20 years...gez there must be a lot of cables down there then. I am sure that some kind of salvage operation would be worthwhile if well thought out...it wouldnt be easy to do it at a profit but i could sure go for a few thousand kg of copper!
I want to know what optics they are using. On segments that are 7,000 miles long - are there active optic splice boxes along the length of the cable? Or are they using lasers that actually push 7K miles?
Submarine cable systems often employ repeaters (hardened EDFAs). I remember one vendor bragging that not one of their units had EVER failed in service.
That PDF is great, thanks. Jesus - they are huge. Earlier in the PDF, they appear small.
Is that a single strand they are amplifying with these? I'd like to find cost figures for how much it is to run a single fiber trans-oceanically (if thats a word).
They talk about it in the PDF. Apparently they use a 1400nm laser along the same fiber that excites some receptor, which provides the power for the amplification.
The diagram of the Polar King cable-laying ship that gk1 posted here shows a "repeater stack" near the holds full of cable. I assume the repeaters are spliced into the cable as it's being laid. http://en.wikipedia.org/wiki/Submarine_communications_cable#... describes the repeaters in a modern submarine telephone cable, and says: "The optic fiber used in undersea cables is chosen for its exceptional clarity, permitting runs of more than 100 kilometers between repeaters to minimize the number of amplifiers and the distortion they cause."
They are "powered" by lasers themselves. No electronics involved; excess energy from a pumping laser boosts the intensity of the data laser, using sections of special doped fiber.
That's very cool, but the "submarine communications cable" article I linked to above claims that the repeaters are powered by DC fed through a single conductor in the middle of the cable, though it doesn't cite a source for that information.
I wouldn't be surprised if both methods have been used in different situations or at different times.
Are you saying that there's an extra "pumping fiber" in the bundle, fed by a pumping laser on shore? Or are the pumping lasers in the repeaters, in which case there would still need to be electrical power supplied down the cable to drive the pumping lasers?
The DC may be likely for old-style repeaters, as Er-doped fiber amplifiers are (relatively) new.
It's my understanding that the pumping laser would be a different wavelength carried by the same fiber (just as a fiber can carry multiple data channels using different wavelengths in DWDM). The boost happens because of the special doped section.
The pumping laser essentially puts the doped medium in an energy condition where the data transmission can stimulate additional emission (similar to how lasers work in the first place). Few photons in, many out. The energy required to do so comes from the pumping laser.
But seriously, that is damn cool - while I have been in IT/networking for a long time - I was not aware of "pumping lasers" as everything I have done, even in designing massive fiber cable plants have all been for relatively short distances.
1. I'm always amazed that anchors can seemingly hit a cable at the bottom of the ocean floor. Given how hard it usually is for us to find anything on the ocean floor with any precision, we seem to be able to drop anchors and cut cables with surprising accuracy.
2. Do the cables have slack in them to account for continental drift? I don't know how much we drift, but I'd imagine over time the cables would need some slack in them to make up for it.
Alaska has the combination of having a lot of coastline (more than the contiguous US) and lots of cities that aren't easily accessible by land. Most of southeast Alaska's cities (including the capital) have no roads in or out at all, but are on the coast, so undersea cables are the only practical means of providing connectivity.
It's unfortunate that TeleGeography, while on the one hand demonstrating beautifully how closely connected many of us have become, seems at the same time to collaborate with BlueCoat, a company that helps censor the content coming down these tubes on a national scale.
I remember working on the application that was used for administering the SAT-3/WASC cable. At the time I was excited because I knew it meant South Africa would be getting faster internet pretty soon after that. 5 years later I was still stuck on a 56k modem. ADSL launched about two years before that, but it was just way too expensive. At the moment I'm paying about $185p/m for a shaped (20kbps between 7am and 7pm during the week) uncapped 8mbps ADSL line.
Last night I was curious about physical wire connections across oceans for communications. So I started researching and learned about the first ocean line implemented in 1858 for telegraph communication. Seeing this link on the top of HN today was pretty much exactly what I was trying to conclude my research with last night. What a strange coincidence.
There's only 2 connections to the USA at the moment, and SCC charges monopoly prices. There is another AU<>NZ<>USA connection being built in 2014 and another AU<>NZ connection built in 2013 (the 2 grey lines). Southern Cross already dropped their prices 44% in response.
I was talking to an astronomer the other day who mentioned that they're trying to set up a big distributed radio-astronomy rig out in the desert. One of the consequences is that this thing is going to be producing a ton of data - I think he mentioned on the order of petabytes-per-second. So they're trying to work with the Australian government and Telcos to defray the costs by offering the connection as a service, as well. So the cost could drop, soon.
Or it could just be offered as a $300 MEGA-SKYLINK-INTERTRON.
Before a shooting war, like during te USA USSR cold war, militaries and intelligence agencies put a lot of effort into tapping one another's undersea cables. The trick is doing it without being detected, especially on a fiber cable. Operation Ivy Bells was one of the most documented.
That's so awesome, I remember seeing a special on the Discovery Channel years ago (10+) on how they laid cable for the phone lines I would love to see how it's done now and how it's connected to backbones etc..
For North Korea specifically, an undersea cable doesn't really make sense anyway. Geographically, you're better off serving as a conduit between mainland and the tip of the Korean peninsula rather than getting your own underseas. Geopolitically, your current ally shares a land border with you, and their side of the border is an area strategically fairly important to them and so it is likely fairly well connected.
Well, strictly speaking, it's not the usual Mercator projection, but the "Google Mercator" or "Spherical Mercator" projection, EPSG 3857 or EPSG 900913.
This is a Mercator-like projection that assumes the Earth is a sphere instead of an ellipsoid. Many online mapping systems use this projection because it is a much simpler calculation than than a more correct Mercator projection would require.
I know this is incredibly childish and probably impossible given the elasticity of the cables laid out BUT I would really like to tug on a submarine cable or even just a thread that goes from Euroland over to the USA and have it move and wiggle over there. And then next up: a submarine miniature tunnel/hose/pipe to send a small little vehicle across. I can't help it.
http://upload.wikimedia.org/wikipedia/commons/a/a5/1901_East...
And somewhat related, a wonderful book called The Victorian Internet:
http://www.amazon.com/dp/0802716040/
If you ever wondered whether the Internet really is a "series of tubes", here are the tubes:
http://www.popsci.com/files/imagecache/article_image_large/f...
That photo is from an interesting article (by my namesake James Geary) about undersea cables:
http://www.popsci.com/scitech/article/2009-03/who-protects-i...
Enjoy! :-)