Hacking Deep Sea Cables

There has been an arms race for over a century now between hardened wired systems and encrypted wireless systems. Typically, a closed wired system can be safely assumed to be secure. If your system doesn't connect to the outside network, the outside network cannot connect to you. There is an obvious loss of utility there, but it's not hard to see the upside of that design.

Take the internet for example. It's a big place. Some of it is private and not directly connected to the rest of the internet. For a problem like that the NSA had room 641A. It is a telecommunication interception facility operated by AT&T at  611 Folsom Street, San Francisco. It was exposed in 2006 and the NSA was sued by the EFF. This isn't a conspiracy theory. PBS has covered the story. The EFF has filed two different suits over the obviously illegal operation.

This isn't the first time this has happened.  Decades ago the NSA and CIA began to scheme to tap international telephone lines. They already had access to most of the phone lines that went through the United States. But by the 1970s the planet was criss-crossed with phone cables that didn't route calls through oor two the USA. The U.S. government is just too nosy to leave that alone. Someone might say something about their mom.

So in the Fall of 1971 the U.S. Navy began Operation Ivy Bells. The nuclear submarine, the USS Halibut was piloted into the Sea of Okhotsk, Russian waters. It's mission was to find a marine cable and tap it. The cable was known to connecting a Soviet naval base on the Kamchatka Peninsula with a larger base in Vladivostok. The only problem was that it was under 400 feet of water and they had to overcome that obstacle with 1970s technology.  Undersea cabling had been in use since the 1840, it took over 100 years to learn how to hack it. More here.

But they succeeded. The NSA had developed a waterproof tap, which was 20-feet long . But whether due to limitations or by design the tap didn't transmit data. The NSA had to return monthly to the site to retrieve the recordings and load in new blank tapes. Their fun was ruined in 1981, when NSA employee Ronald Pelton tattled on them to the KGG. Pelton was tried and convicted of espionage in 1986 and sentenced to three concurrent life sentences. The original tap is currently on exhibit at the KGB museum in Moscow. More here.

But the NSA and the CIA weren't done tapping marine cables. In 1979, the USS Parche installed another tap on another  Soviet undersea cable, this time in Barents Sea. That remained undetected until 1992.Other cables were tapped off the coast of Africa and in the Mediterranean. Files leaked by Edward Snowden in 2012 revealed that  intelligence agencies had tapped more than 200 fiber optic cables around the world. There are 277 undersea fiber optic cables in the world today. Obviously they're not done yet. Known projects go by names like Oakstar, Stormbrew, Blanery and Fairview. There are surely others. So the solution isn't isolation, it's encryption.

Sir William Thomson

For his scientific endeavors William Thomson was knighted by Queen Victoria and later became Lord kelvin. He had extensive maritime interests and was probably most noted for his improvements to the mariner's compass. But this does not interest this radioman. Thompson worked on the transatlantic telegraph project and first quantified what we now call bandwidth. This was half a century before Harry Nyquist was even born.

It was known that submarine cables behaved differently than land lines. It was unclear why. Michael Faraday had demonstrated that the properties cable itself limited the rate at message could be sent through it. Faraday realized that the data rate was retarded by induction between the current in the wire and the material surrounding it. It is for this reason that he suggested telegraph cables be insulated with gutta percha. For all of Faraday's cleverness he didn't manage to quantify this observation but Thompson was all over it. He developed what is sometimes called Kelvin's submarine cable theory or Thomson’s "Law of Squares". More here.

Thompson described his results in terms of a data rate. By his math, long marine telegraph cables could be run "constructed of gutta percha containing a core not exceeding 2.72 times it's own diameter that of the conductor..."  Thomson recognized that the speed of a telegraph signal was limited by two measures: capacity and resistance. He calculated that the speed decreased as the square of the cable length increased for any given diameter of the core conductor. In this sense capacity became 'capacitance' aka the amount of stored electric charge for a known electric potential of the cable.

Thompson was not done. In 1855 he published a further analysis, correlating this to profitability. Thomson contended that the speed of a signal through a given core was inversely proportional to the square of the length of the core. Thomson got all kinds of flack for this so it intimated that very long cables woudl transmit messages more slowly, and thus carry less data. Wildman Whitehouse, at the Atlantic Telegraph Company. wrote that Thomson's calculations implied that the cable must be "abandoned as being practically and commercially impossible." Whitehouse was obviously wrong, even though Thomson was proven right.

Submarine Cable Map!

You can see it here.

Greg's Cable Map is consolidates all the available information about the submarine cables, the backbone of our internet. You might think your ISP is the backbone. It's not. As much data as we move around the country, through various ISPs: AT&T, Verizon, Time Warner, Cogent, Level 3, XO, Comcast and whomever... bottlenecks remain between nations. That new Facebook data center is being built in Sweden. the bit.ly root servers are in Libya. half of the Wikipedia severs are at a data center in the Netherlands. You are just a local instance, the internet is inherently international.

So this free map is a nice resource if you're trying to understand how and where your data has to move. He's still tweaking it so don't forget to click the feedback button if you see something awry.

Submarine Telegraphs

I saw a book cover, ornate and gilt that read "Submarine Telegraphs." I was struck immediately with the wrongness of that configuration. You can't have a telegraph on a submarine, wireless yes, but telegraph no. After a bit of research I found three things.
1. Cable terminology has changed
2. Google books scans peoples hands by accident

Today we more often refer to these as sea cables, or marine cable, maybe even submarine cables. Our complete abandonment of the telegraph has slowly led us to refer to the cabel and not the device it connects to.

In 1902 The New York times published the article I clip above. Over 100 years ago we were so reliant on the telegraph that it made business sense to run 200,000 miles of submarine telegraph cable cable in 1,750 separate segments. As the tile says the cost of that was over a quarter of a million dollars. In 2006 moolah that equals $6,423,801. Of course the catastrophic devaluation of the dollar this year reduces that somewhat but I'm sure it's still a very big number. More here.

The construction of such an infrastructure commits you to both it's maintainence and compatibility moving forward. Those cables transmitted 4 billion messages a year. Of course that's less than our modern network move in a minute. Running a cable under the ocean is a big undertaking. At the time there were 1,180,000 miles of cable . Each carried dozens of individual wires, almost 4 million miles of copper.

At the time the aforementioned article was written, there were still cables being laid to criss-cross the Pacific to connect Canada with Australia. These cables weren't laid in straight lines. While that would have saved thousands of miles of copper, it's not practical. In order to access the wire, it needed to "land" at multiple islands. Just negotiating the connections with the multiple municipalities and nations was incredible. More here.

At the time all this was still battery powered. It was only in 1860 that terra voltaism became a possibility. A British civil engineer named Septimus Beardmore worked out a way to operate a telegraph system with only a single voltaic element. it consisted of a pair of dissimilar metals buried in the earth at opposite ends of the line. His early experiments worked over a 300 miles cable between the Cromer and Heligoland Islands.General Oceanic was the first company to attempt a production submarine telegraph. They incorporated in 1845. they later changed the name to the “General Oceanic & Subterranean Electric Printing Telegraph Company.” Mostly they talked a big game. They never built anything. Today we call it vaporware. the real hero of undersea cabling was Nathaniel J Holmes. In 1848 the Electric Company of London had fired Nathaniel J Holmes after the telekouphonon fiasco. He became the principal electrical engineer in domestic and submarine telegraphy. He became manager of the General Telegraph co. A company that lasted until 1951. Within a decade he was overseeing projects for the South-of-Ireland Direct Telegraph Company laying 60+ mile lengths under the sea.

Oh, there be wires

In radio we use wires. The microphone is connected to the XLR cable, the XLR cable connects to the volume pot. the volume pot plugs into the mixer slot etc. here I discuss a couple very early wires. remember them the next time a crappy headphone cable shorts out.

Wires connect anything to everything to everyone. In this most recent stage of our technological evolution we have outgrown what copper can do and are moving toward fiber. I hate fiber, because I can't solder it, can't crimp it and in general I am trying not to be a Luddite. In the end you cant plug fiber into an antenna, so I get my way as well.

The very first interior power wiring systems used conductors that were bare or covered with cloth. This was not tractable outdoors or over long distances so we began to experiment with the first real cables. The above slotted block of wood is an early cable. They were called deal tongues. Into those slots early engineers pounded lengths of metal. Metallurgy was uncertain in the era so alloys were impure and irregular in gauge. All the more important that they slathered it in tar to keep out the water. This one is circa 1065 or so from London.

Another early cable was the first underground wire and was a gutta-percha-covered wire. The brown gutta-percha is cracked like crushed amber or resin. Gutta-percha is a resin from the Isonandra Gutta tree that has some properties as an insulating material. it originates in Malaya and was discovered by Samuel Canning. (who later was involved in laying transatlantic cables) More here.

In India, probably with help from the British, half-pipe (or tile) conduits were used. These were much like the terra cotta tiles we see on roof tops but cupped together to make a pipe. the wires ran through them like conduit. These tiles are filled with what looks like a mortar that appears in some examples to be tarry. Tar or pitch seems to have been used in some pieces for joining a collar into the larger end of the next half-pipe. Some pipes are liberally painted with tar inside and out. More info here and here.