How the Internet Travels Across Oceans | Hacker News Reader

How the Internet Travels Across Oceans

nytimes.com

365 pointsanuragsoni160 comments

I've rented some of these lines as part of a financial trading infrastructure.

What's immediately obvious is the line is much faster than going over the public internet, and there's no jitter: every time you ping it comes back in the same amount of time, whereas the internet pings will vary by several ms.

You also become acutely aware of cable breaks. Somehow this happens quite often under the English Channel. The network operator will start sending you emails that say

- Cable break detected

- Loading ship (can take ages)

- Found the break

- Patched it

- Back up

Occasionally bad weather would delay it.

How costly is such renting?

You're unlikely to rent the whole thing so it just depends on how much bandwidth you need. If it's only a couple of TB per month it's not that expensive.

But that doesn't answer actually how expensive.

It depends on what you want. Starts at maybe a couple of thousand bucks a month I would guess, I haven't looked for the cheapest in the market. Depends on the link and availability and so on.

Maybe it is expensive. They could be seeing "trading" and charge more because those folks have deeper pockets. Like how the DLSR industry is suddenly charging astronomical prices when the feature set of their devices is for cinematographers.

Edit: Why the downvotes? Adjusting the price to how much the customer is ready to pay instead of just offering the cheapest possible price is usual business practice.

You pay per capacity, not per data transmitted

>You also become acutely aware of cable breaks. Somehow this happens quite often under the English Channel.

You're not lying. :) When the internet is down in the ROI (Republic of Ireland), it's always because something was dragging along the seabed and cut the cable[s].

Given that the ROI doesn't have many IX's for international connection[0,1,2], it's quite hard to miss when it does happen.

[0] - https://en.wikipedia.org/wiki/Internet_Neutral_Exchange

[1] - https://www.inex.ie/technical/network-diagram/

[2] - https://www.submarinecablemap.com/#/country/ireland

I had always had the illusion that Ireland was well connected to US and EU. Given how many tech company have Data Centre in there. Quite Strange its IX isn't as good as it could be.

Off Topic:

Is the short form (Republic of Ireland) ROI used often? Because the first thing that springs to mind reading ROI is Return of Investment.

Isn't Ireland short enough as a word? I understand there is Northern Ireland, but people will actually put Northern In front of it to be specific. Like North Korea instead of Korea.

ROI is often used, yes.

It's a good way of easily differentiating between whether you're talking about Ireland the island, or Ireland the country, which currently excludes the six counties of Northern Ireland (NI) which are still part of the UK.

The term Ireland is often used to refer to the island as a whole, but obviously this can depend on the political leaning of who you're talking to.

Northern Ireland, depending on context and/or who is speaking can otherwise be referred to as Ulster (problematic, as the province of Ulster has 9 counties, only 6 of which are in Northern Ireland). It is also common in some circles to refer to the 6 counties currently occupied by the British simply as "the north".

In short, Ireland is a hot geopolitical mess and when talking about Ireland, it is useful to be as specific and as neutral as is possible.

So ROI/NI are quite useful shorthands.

I strongly suspect most residents of Northern Ireland would take umbrage at the word "occupied"

You're not wrong. There is a loyalist majority in NI. But those that would refer to Northern Ireland as "the north" would generally be among those that definitely would refer to the six counties as occupied.

> The five most recent opinion polls taken in the North show similar results, with support for the North staying in the UK ranging from 45 per cent to 55 per cent, and averaging around the 50 per cent mark. [0]

Way closer than you think. (Using desire for a united Ireland as a proxy here)

[0] https://www.irishtimes.com/opinion/polls-suggest-gradual-shi...

To explain ROI you really wouldn't had to use terms like 'still part of UK' and, as mentioned, 'ocuppied'. This only distracts.

Its a touchy and pedantic subject. 95% of people will understand what you mean when you say "Ireland", the last 5% will want to start a flame war over it because reasons.

I understand what you're saying, but it's not just pedantry.

You do have to be specific about what you mean when you say "Ireland" in any context where it is important that you are understood.

"Ireland" can quite correctly either mean the island of Ireland or the country called Ireland (aka The Republic of Ireland). In a political or business context you are likely to be specifically referencing one or the other.

In the context of Internet connections "Ireland" is precise enough, since either normal meaning works.

Since they're geopolitically separate and have their own infrastructures, the distinction is quite apt for understanding the "who" of who is impacted when the specific cables are cut.

In other words, based on your assumption that "Ireland" is precise enough and since Belfast (NI) has it's own IX, that would mean that you would have dichotomy where "Ireland" could connect to the internet and "Ireland" could not, yeah?

Agree, it's like saying Europe. If you ask a geologist, it will include Russia. If you ask Macron, it will not!

I think in this context (cables landing on an Island called Ireland) we're ok to use simple terms and be understood.

That wasn't the question ksec asked though was it.

"Off Topic: Is the short form (Republic of Ireland) ROI used often?"

But also no. Cables landing in ROI would be under completely different regulatory control to cables landing in NI which would be under United Kingdom regulatory control.

It's easy to remain oblivious and treat any mention as pedantry when you're not among those affected. Even if 95% of people are like that, denying their reality is still a form of soft bigotry. Here's an essay that might help to explain why some people care more than you do, and rightly so.

https://www.nytimes.com/2018/10/18/opinion/england-ireland-b...

Sort of like how people confuse The Republic of China (ROC) with the People's Republic of China (PRC)?

No doubt there are a great many other parallel situations out there. My dental hygienist took issue with the Netherlands being called Holland.

> Is the short form (Republic of Ireland) ROI used often?

Most often usages for me are filling out forms and that's often the shorthand representation next to the soccer team's score during a match.

Ireland is the same length as America, which people often abbreviate to USA or US. This also cuts the confusion of America's multiple potential meanings (North/South continents, etc).

> Is the short form (Republic of Ireland) ROI used often?

It disambiguates the polity from the island, which includes both the Republic and a sometimes problematic piece of the United Kingdom.

>Somehow this happens quite often under the English Channel.

How often? Once or twice a year?

And what were the reason it breaking so often? I thought the English Channel were quite free of Cargo Ships and the like.

> I thought the English Channel were quite free of Cargo Ships and the like.

The English Channel is the busiest shipping lane in the world: https://www.youtube.com/watch?v=pzJwXxUY3MM

Lived near Brighton for years and had never knew that! I stand corrected.

http://www.bognorregisbeach.co.uk/live-shipping-map-english-...

:-)

> I thought the English Channel were quite free of

Very much not:

http://www.guinnessworldrecords.com/world-records/busiest-sh... (1999)

https://www.marineinsight.com/marine-navigation/the-strait-o... (2017)

There is some discussion to be had as to whether there are other sections of water that are more active and have been for some years, but even if the channel isn't the busiest it is certainly very busy.

I think more than twice a year when I was renting it.

-Educated guess is that cargo vessels aren't much of a problem. Anything dropping anchor, dredging or somehow doing stuff on the bottom, on the other hand...

Further, I'd wager that initially buried cable in the Channel often is exposed on the bottom due to strong currents, rendering it more vulnerable - but this is just a guess, mind.

One thing I’ve consistently noticed is that a “network” can tell the tale of natural or man-made disasters in terms of how it first manifests in alarms and breakage in various systems, the human response manifested to respond to it (sometimes across many national boundaries), culminating with the sometimes peculiar “hacks” and temporary solutions conjured up to restore connectivity. The network sees everything and no one ever tells these stories in any meaningful detail.

I'd be very curious what the timelines between each step looks like.. any insight?

Everyone knows immediately when the line is down, and which line it is.

If the ship is ready it seems to take a day or so to get to the site. There's not always a ship though, not sure how it's organized.

Another day at the site and it's done. But any weather blows up the schedule seemingly.

I've never received anything other than emails about this so I'm not actually an expert.

>"I've rented some of these lines as part of a financial trading infrastructure."

Unless you were lighting that fiber yourself you were renting wavelengths and not the physical fiber.

>"What's immediately obvious is the line is much faster than going over the public internet, and there's no jitter: every time you ping it comes back in the same amount of time, whereas the internet pings will vary by several ms"

Jitter is caused by congestion which happens at router interfaces. The absence of jitter is not due to the use of dedicated submarine fiber spans per se as the speed of light in fiber is a constant. What you are likely seeing is the result of a network operator who has optimized for low and consistent latency by having fewer hops in their network before and after the cable landing stations.

Yeah that's right, I guess it could read like them being underwater matters. Point was that being private and optimised made the jitter go away.

General purpose computers networks and software are far from optimal for high speed trading. A few years back ACM Communications published a comprehensive white paper of trading communications bottlenecks including the effect of virtual memory page faults and TCP/IP collisions. Dedicated trade networks want custom software without virtualization features. Or if you dont want to develop that much, use the ancient real time UNIX NASA has in its space probes like VxWorks.

In a few weeks there is annindie film called The Hummingbird Project about avrace between firms to build the fastest dedicated trading line.

The book Flash Boys by Michael Lewis covers HFT in general and a similar fiber optic line built by Spread Networks in 2010. It's definitely a fascinating world.

Apparently HFT's have moved on to microwave (https://www.bloomberg.com/news/features/2019-03-08/the-gazil...) to gain a further edge over fiber optic.

I have a certain admiration and appreciation for this sort of thing, and every article that describes how the internet is physically interconnected.

Having seen internet connectivity in India slowly grow from only satellite-based connectivity, to having private internal interconnecting hubs within the country for bandwidth providers and ISPs to reduce latency, then news of every new submarine line improving bandwidth and latency, thereby reducing pricing. Going from getting 135ms - 200ms via dialup locally, to getting 60ms to Singapore and Hong Kong while playing Counter Strike and Quake 3. It was such a high!

Each development was so critical and it brought me great joy, and excitement. Still does!

If anyone is interested in a really well written travel essay about how one of those cables was installed in the 90s, this one never seems to get old: https://www.wired.com/1996/12/ffglass/

Cryptonomicon/Reamde sized too!

(It's a expression, but I'm sure someone else will get it :) )

~41530 words

And for an older take, I loved this book on the first transatlantic telegraph cable as a child:

Cyrus Field's Big Dream: The Daring Effort to Lay the First Transatlantic Telegraph Cable

For more on the history of the telegraph, I recommend The Victorian Internet by Tom Standage, which also covers in detail the laying of the first trans-Atlantic cables. Fascinating stuff.

I have a quite good book on the shelves on the very same subject, but I am unable to locate it at present.

For a slightly wider take, Arthur C. Clarke's "How the world was one" is a most readable account of how telecommunications has made the world smaller, starting with the first submarine cables, proceeding with satellites (of course he would!) before devoting a section to fibre optics at the end. Brilliant account.

There is a museum on the cape cod massachusetts called the "French Cable Station Museum", that has a lot of the old euqipment used back to transmit telegraph data from Europe to America. Its not open year round apparently.

I did the tour a few year ago, run be retired engineer who was enthusiastic and excellent at explaining the difficulties in getting it all working.

As an added bonus, old school html:

http://www.frenchcablestationmuseum.org/index.htm

Stephenson's "Mother Earth, Mother Board" is still, I believe, the longest piece published in Wired. A tremendous piece that still holds up.

Here's the link for Neal Stephenson's Motherboard Earth: https://www.wired.com/1996/12/ffglass/

Indeed well worth a read even this long after it was published.

Thanks! Somehow I missed this essay, looking forward to reading in later tonight.

Interesting to note that submarine cables played a significant part in the plot of Cryptonomicon by Neal Stephenson.

Cryptonomicon really hammered home how very much we depend on undersea cables to power the internet. Great read for those who have not had the pleasure. Plus you get to learn something.

There is also this book Tubes by Andrew Blum about the cables: https://www.amazon.com/Tubes-Journey-Internet-Andrew-Blum/dp...

Someone gave it to me at a previous job but never left a note to say who so I couldn't thank them.

Came here just to post this link; it's really stuck with me over the years.

This is perfect.

Can't tell you the number of times I've had someone insist that the internet was on satellites.

Using terms like "the cloud" and with everything bring wireless definitely doesnt help with people's misconceptions.

Even after I carefully explain that the internet uses under sea cables and generally either buried wires or above ground stuff I get dubious looks they say "yeah, right. You're probably just confused..."

>Can't tell you the number of times I've had someone insist that the internet was on satellites.

It has to go to space!

Are you referencing the Louis CK thing? I cringe every time I see that.

I'd hear that all the time too but satellites do play a role. I'm just unsure exactly how they fit in.

I’ve blown a few minds explaining that the Internet is delivered via light, sent through very long fiber optic cables under the sea.

I thought this was common knowledge until the first time I explained it, and the disbelief... Everyone had a revelation that day.

I always knew about fiber optics, growing up with dial up and dreaming of one day having a fiber connection. But I guess now that everyone is focused on mobile that piece of infrastructure is abstracted from even reasonably tech savvy people.

How did they think it was done before you explained it to them?

I’ve met quite a few non technical people who assume the Internet is carried by satellite.

that was my own assumption until I learned about the under sea cables.

I distinctly remember "learning" this in K12, before getting deeper into the technology industry. One of our textbooks had a diagram with lines showing signals flowing from continent to satellite to continent. It wasn't until I got deeper into the industry that I learned satellite latency is awful, bandwidth is limited, and everyone uses cables or microwave point-to-point links when they can. (Actually, the role of those point-to-point links is drastically undersold for how ubiquitous they are in our cities. Worked with some Ubiquiti gear professionally once, and now notice them everywhere).

It's actually a very reasonable assumption, as most people don't really know about the limits of fiber, but they do own mobile phones which seem to work almost wherever they are thanks to wireless technology.

It's not reasonable once you consider the bandwidth of satellites.

And the delay

"Normal" people don't know that kinda stuff.. Let's be honest, they make many of their assumptions from TV and Movies - and often those disaster movies focus on the world grinding to a halt because "the satellite network is down" (I do the same with other industries!)

Depends where you are located... e.g. for cook islands this is the reality till they also get their own fibre optics in mid 2020.

I asked this question to a class of smart 18-19 yo college students a few years ago. Almost everyone thought it was done with some kind of wireless transmission (e.g. "by satellite"). The idea that it was physical cables lying in the mud of the sea bed was a revelation.

Did they speculate about how it was done before fiber optics?

I'm sure that Marcono would like their answer as well.

I recently visited friend at his lab. I cannot say much but they work on patent for something he calls “pantone fiber”. Basically within one cable they are able to squeeze about 16,000 different shades of R/G/B “working” next to each other. When I grasp the idea it blew my mind - he looked at me and smiled saying - “yeah about 10,000 times faster than current fiber”. He said two things tho - one state of tech is not ready for such speeds and latency - unlesn you have 16K def TV. Second he said they wont release it until much more cable has been layed out. So that of course they can restart relaying new cable again and make more $ ;)

Did your friend say 10,000 times faster or 10,000 times bandwidth? By using different color simultaneously , you can stuff more data but I can’t imagine it becoming faster than speed of light it already travels at.

For a bit, yes. For N bits, more bandwidth means more effective speed.

Well, no, not unless you define speed as “bandwidth”.

If it takes 1ms for the 1 bit to get to the other side, it’s still going to take 1ms for N bits to get to the other side, even N-multiplexed.

It’s the bandwidth that’s improved with multiplexing.

While you're technically correct, isn't it pretty much an accepted fact that when people refer to the "speed" of internet, they are actually talking about bandwidth?

E.g. "how fast is your internet", etc.

It's of little value to the end user to know how fast a bit travels, all they care about is how fast they can download.

I'm sure he knows that and is just being pedantic for the sake of it

On HN? Never!

Isn't this just an advancement on wavelength division multiplexing?

https://en.m.wikipedia.org/wiki/Wavelength-division_multiple...

https://xkcd.com/1053/

>Inside the ship, workers spool the cable into cavernous tanks. One person walks the cable swiftly in a circle, as if laying out a massive garden hose, while others lie down to hold it in place to ensure it doesn’t snag or knot. Even with teams working around the clock, it takes about four weeks before the ship is loaded up with enough cable to hit the open sea.

How is that not automated? I cant believe that companies would rely on humans doing a boring and repetitive task without error for weeks on end when a single error could cause double digit percentage of time at port. Is this is just such an uncommon task that automating this wasnt worth the time saved at port?

edit: Even in the picture in the article you can see a single worker has removed their shoe covers that all the other employees are wearing[1]. Doesnt that degrade the cable right there?

[1]https://static01.nyt.com/newsgraphics/2019/02/15/undersea-ca...

The cable is going to be sitting in the sea. Shoes are likely not the biggest issue it will face in its lifetime.

The cable will never again have the rest of the cable on top of it. I'd be slightly worried about sand/dirt facilitating chafing/abrasion in the outer layers of cable on its way to its final resting place. Shoes shouldn't be an issue as long as you can keep mud caked work boots out.

No, but it will have the ocean on top of it!

I work for a company which make the full deployment line for submarine cable laying ships; first time I visited one -Incidentally, a sister ship of the CS Durable referenced in this article- I wondered about the precise same thing you did, beginning to sketch solutions in my mind while looking down into the tanks.
A supervisor comes over to me and says something like "I can see what you are thinking. There's no use - it is cheaper this way."
The point was that while the ship was in port taking on cable, it was fully manned anyway. Crews need something to do - so might as well stow cable.
That's our kit in the topside photos. Small world!

) We do all sorts of other things as well - relying on cable laying alone would be too much of a niche market!

  The point was that while the ship was in port taking on
  cable, it was fully manned anyway. Crews need something 
  to do - so might as well stow cable.

Right, but if they could take on cable faster couldn't they get under way faster? Four weeks to load a ship sounds like a long time.

I guess that depends on how fast you can make the cable; I have no idea how that pans out - but the cable is put onto the vessel straight from the production line (which, I guess, may have had its speed optimised to the speed at which crews can reliably put down cable.)

I am now curious; I'll fire off an email with a few questions to my liaison at TE SubCom. With any luck, answers arrive in time to be relevant to this thread.

This is so neat! Few questions though, if anyone can answer:

1) The cable still looks pretty thin, is it being housed in some sort of protective layer/tube before it is laid? Seems like an anchor snag could easily break that line. If that happens, is it possible to 'patch' a fiber cable?

2) Once it nears the coast, how deep is the line being buried?

3) This may be a dumb question (I'm no engineer), but does the light travel from the source all the way to end without any "relay" mechanism?

4) How much bandwidth can travel through one of these lines?

5) Does anyone know how to save that(cabling-world) animated image alone? All I can make out is 'g-map' div ;-)

Make window fullscreen, press Windows-G on a Windows 10 computer with halfway good graphics card. Pump it into VLC and crop away the stuff you don't want ... and profit? :-)

Here you go: https://giant.gfycat.com/UnderstatedLinedBird.gif

That animated image is made up of a canvas background of the world and an SVG foreground of the cables. Both are created using js from downloaded json data, ne_110m.json and fusion-cables-2019-02-27-simplified.json. Quite tough to get that into something that's not using pixel graphics and self-contained (aka only displaying the stuff in the page).

1) The fibre pairs are housed in multiple layers of armor to protect against the elements, supply power to inline amplifiers and -most notably- provide the structural strength needed to survive being deployed. Outside diameter is still less than one inch, though.

2) Depends, but a metre or two (tops!) would be the right ballpark.

3) Not a dumb question; while loss in fibre optic cable is very low, repeaters are installed at regular intervals along the cable. they are powered by high-voltage DC supplied through the cable armor. The ocean acts as the return path.

4) Lots.

3) is wrong. The amplifiers are powered by DC, but one side of the fiber provides positive voltage and the other negative - the ocean is not a part of the circuit.

-You are right; I just looked it up. I guess (not that it makes much of a difference!) I picked this up from a book on earlier days of submarine cables.

Thanks!

I was under the impression that cable systems are best understood as constant-current systems. As a side effect of this, they can withstand a power failure at either end, in which case the ground path is part of the circuit. (Essentially "single rail" operation.)

If nothing else, both ends still share a common ground reference, yeah?

https://pdfs.semanticscholar.org/3c72/998ffe9c39a1aa4a9faba4...

Edit: thinking about this more, it seems quite obvious that you need a current return path regardless. If it's not the ocean/earth - what is it?

>"3) Not a dumb question; while loss in fibre optic cable is very low, repeaters are installed at regular intervals along the cable. they are powered by high-voltage DC supplied through the cable armor. The ocean acts as the return path."

That's incorrect, amplification is done via Erbium doping, it is part of the core of the cable. There are no external repeaters.

...but the erbium doping is only part of the equation, involving additional circuitry powered from the surface? Or have I misunderstood completely?

So to clear this up a bit. EDFAs (which are the in-line amplifiers which use erbium doping) work by having a section of a fiber doped with erbium ions. Only the part which is going to amplify the signal is doped, the rest of the fiber is supposed to be as clean as it gets. Those erbium ions get shot with a laser (that's the part that needs power from the shore) and get excited. The source signal then hits the ions, which then emit more photons of the same exact properties, but just way more of them - therefore the amplification part. Continue 80km down the line, and the same thing has to happen again, as enough light has been attenuated.

I mostly work with land systems, where the EDFAs are pieces of equipment completely separated from the fiber - you get a normal strand of fiber dug into the ground, and once you want to amplify it, you've got to plug it into an EDFA line card with the erbium-doped fiber already contained in it. In the sea, those systems are way more integrated and the fiber is usually spliced right into the amplifier.

I might be able to dig up a card tomorrow if you're curious as to how the system looks.

Yes please! Thank you

Please do; that would be great!

1: I think they don't put that thin cable on the seafloor, they are bunched together, in a much bigger cable with a lot of stuff around it [0] to protect it.

About the patching, they say that in the article: they sometimes cut the cable (in case of a bad weather) and attach it to a buoy. So they can "patch it" ;)

[0] http://2oqz471sa19h3vbwa53m33yj-wpengine.netdna-ssl.com/wp-c...

I'm on an iPad at the moment so I can't give you a direct link to any videos, but search for Hibernia Atlantic on Youtube. They have some short videos showing how the cables are laid, repaired, and so on.

Recent graduate in Optical Communications here :)

1) The fiber is pretty thin, but there are several fibers combined in one transoceanic cable, see here https://en.wikipedia.org/wiki/Submarine_communications_cable. They don't patch them, but splice the fibers. Patching introduces a little loss, so it should be avoided.

2) Not sure how deep they are buried, but I think to remember that the shore end of a submarine cable is better protected than the part in the deep ocean. Due to more ship traffic at the coast etc.

3) Good question. The loss of an optical fiber is roughly 0.2 dB/km. Across the ocean the optical signal must be amplified several times (every 80-100km). Nowadays the amplification is all optical (EDFA or Raman). Before there were electrical regeneration schemes. Check out the history of the field, it's is quite interesting [1].

4) Not sure how much data such a cable can carry, since it depends on how many fibers are deployed within. However, there are multiple interesting things to look into here. In research labs, people are investigating multicore/multimode fibers (space division multiplexing) [2], these fiber have incredible capacity. Personally, I think the most interesting metric is the spectral efficiency, so how much data can be transmitted per second per Herz. Such a metric is independent of multiplexing schemes over space/wavelength/time, and improvements have to come from better devices, signal processing or signal shaping methods [3]. Another mind-blowing area is using the nonlinear Fourier transform for better signaling methods [4].

Feel free to ask more questions :)

and checkout the two biggest conferences for more in depth info.

https://www.ofcconference.org/en-us/home/

https://www.ecocexhibition.com/

[1] https://www.osapublishing.org/oe/abstract.cfm?URI=oe-26-18-2...

[2] Multicore: https://ieeexplore.ieee.org/abstract/document/7341685

Multimode: https://www.osapublishing.org/abstract.cfm?uri=ofc-2018-Th4C...

Multimode+core: https://ieeexplore.ieee.org/document/8535233

[3] https://arxiv.org/abs/1606.04073

[4] https://www.osapublishing.org/optica/abstract.cfm?uri=optica...

Looks like I got some reading material for tonight! Really appreciate your response, this is all very fascinating to me.

Cheers :)

With a degree in Optical Communications, are you trying to go into Optical Engineering or pursue further education in academia? How is the salary for the engineer compared to software?

I received my PhD in December, so I already pursued further education. My masters was in telecommunications, starting my PhD with a background in signal processing/information theory rather than optics/physics. All the master students I supervised either had similar background like me or an optics/physics background. That said, optical communications combines multiple disciplines, which makes it so fascinating, and I think that a PhD is inevitable if you want to pursue a career here. There are so many steps to understand to get from bits to electrical signal to optical signal and all the way back to bits. I don't think a masters degree can cover all of it in depth. Further, a PhD is in particular helpful due to the conferences where all the big players are looking out to hire you, or where one meets senior researchers for potential postdoc positions.

Salary-wise it's difficult for me to answer, as most of the big industry players are in the US or Canada, but I'm in Europe. From hearsay a fresh PhD with reasonable publication list will get around 10k per month in the bay area. However, glassdoor might give you a better idea. Look for companies like: Infinera/ciena/acacia communications/juniper/mellanox/finisar/keysight...

After my PhD, I probably could have pursued a postdoc somewhere in Europe, but I decided to leave academia. I wanted to stay in Copenhagen and therefore had to change my field of work and will be working for a hearing aid company in their signal processing department.

4. Undersea should be the same as under land — 9.6Tbps, 96 100Gbps fibers.

I don't think it's right to say data is in the ocean though. They are stored in datacenters on land. And they travel through the ocean. Unless when we start putting more of these datacenters in the ocean like how some company has been experimenting with.

Aside: Is it just me, or do others find animated charts annoying? It takes me time to understand what's being shown, and before the realization has sunk in, the graphic restarts causing the whole mental process to restart. Rather like a strobe light.

Agreed. I'd love to examine the map as of today, or 2021, say, in some detail and whoops - it's gone. Wait for it, wait for it, wait for it, there - aaaand gone. No controls to stop it or skip to a specific year. Very annoying.

Related map of submarine cables on the globe (with clickable links for the cables): https://www.submarinecablemap.com

Would anyone kindly name a major company that install these mammoth cables underneath oceans?

Surely they should be huge in valuation yet hidden away from the public glare.

They are built by the usual telecom and heavy industry names like Alcatel and NEC.

I wouldn't say they're that hidden. Marea (MS & FB trans atlantic cable) has a nice promotional site with videos and suchlike: https://news.microsoft.com/marea/

Orange is a big player on this market, probably thanks to the fact that (as another user mentioned) Alcatel is one of the largest manufacturer (both historical french companies).

You can follow Jean-Luc Vuillemin [0] on Twitter for very interesting pictures of ships laying cables around the world.

One of the largest undersea cables plant is Alcatel in Calais, France (Dover straits). The factory is linked to the port by huge tunnels under the city. Unfortunately the Wikipedia page is only available in french but contains nice pictures[1]

[0] https://twitter.com/jlvuillemin [1]https://fr.m.wikipedia.org/wiki/Alcatel_Submarine_Networks

Checked out the Twitter feed.

Really great!

Tyco / TE Subcom and Japan's NTT are two of the major players.

KDDI both builds transoceanic cable and has ships to install them.

http://www.k-kcs.co.jp/english/cableShipKOL.html

http://news.kddi.com/kddi/corporate/english/newsrelease/2017...

> The ship will carry enough supplies to last at least 60 days: roughly 200 loaves of bread...

Unless Jesus is one of the crew members, that isn't nearly enough.

Perhaps they don't freeze it, so having more than ~10 days of bread would mean throwing out mouldy bread?

When I went to Fiji a few years back I was astounded by the quality of the internet considering we were in a remote part of the South Pacific. Turns out Fiji is a "layover" for all the underground cables coming from the US West coast to Australia...

Nice visualization at the intro on desktop+. I wonder if Rich Harris and his team were behind it? See recent changelog podcast featuring Rich Harris on being 'a javascript journalist'[0]. Harris is author of the 'magical disappearing UI framework', svelte[1].

[0]http://changelog.com/podcast/332 [1] https://svelte.technology/

Old but good article by Neal Stephenson

https://www.wired.com/1996/12/ffglass

Interesting stuff. Just as important to education as "how to make a car" was 50 years ago.

Things not covered in this article: Expected carrying capacity? How much dark fiber (initially unused capacity)? How much extra cable is required to make sure it stays on the bottom even when going over oceanic mountain ranges? How much of the ocean-floor survey data is open, how much is proprietary, and how much is (military-style) secret?

Is there any sabotage? Do nav charts warn about these things?

I don't have any links handy, but I have definitely read reports of sabotage and wiretapping.

Edit: ah, from another comment down-thread: https://en.wikipedia.org/wiki/Operation_Ivy_Bells

> Last year, Australia stepped in to block the Chinese technology giant Huawei from building a cable connecting Australia to the Solomon Islands, for fear it would give the Chinese government an entry point into its networks.

I don't really understand this move. Surely the way that the Internet is designed, you can connect any network to any network. Why would this be any different with a physical cable in terms of security?

I don't think most politicians understand how the internet works

Related: Arthur C Clarke has a very interesting non-fiction book about the early days of submarine cables. "Voice across the sea".

How the Governments Physically Look at the Internet that Travels Across Oceans. [1]

[1] https://nationalinterest.org/blog/buzz/russian-spy-submarine...

kinda funny that the headline reads "Russian Spy Submarines Are Tampering with Undersea Cables" while the article does not provide actual evidence that Russian Spy Submarines Are Tampering with Undersea Cables. maybe they were just trying to sow discord into the cables.

Might wanna read this as well: https://arstechnica.com/information-technology/2016/05/how-t...

Interestingly I was just thinking about this the other day, and hoping that someone here would have the answer. I'm really curious what is the total bandwidth per second across different continently globally without accounting for intra-continent bandwidth?

Someone who wants to learn about the internet. Check out the KhanAcademy one https://www.khanacademy.org/computing/computer-science/inter...

I read this somewhere that Russia was tying to snoop on US by attaching these cables with electronic surveillance equipments.

IIRC military and defect corps have separate cables installed for themselves, and the Russian’s were trying to read the signals.

Would someone tell more about it or add the Wikipedia page here please?

Huh, it was the other way around: https://en.wikipedia.org/wiki/Operation_Ivy_Bells

Pardon my sieve-like memory.

I should had clarified I am quite dim visaged about it.

Didn't realize Google, Facebook, Microsoft, and Apple outright own some of those cables.

That's exactly what I was thinking as well. And what's more, it seems like they are laying more new cables than what "public" infrastructure is, going forward.

That can literally mean GooFaceAma has better global performance than competitors. Or that competitors will be forced to get hosted on one of those cloud providers to get similar performance.

It makes me feel a bit worried that "the internet" is not as neutral as I would like it to be. (Or that I thought it was before)

Weird, the article always mentions Google, Facebook, Microsoft, and Amazon, but then once at the end in a footnote does mention "Apple, Google, Microsoft or Netflix".

I'm left wondering whether it's Amazon or Apple or both that own some of those cables.

Whoops I meant Amazon (from the map).

What I find fascinating is that they also have many repeaters given the massive distances covered by these cables so they need to provide electricity through the cables on such long distances.

This is exactly the reason why I like DigitalOcean name so much.

I remember discovering them when I was young and thinking how mundane it is. I also thought that satellites are a integral part of the internet just like the TV network.

What happens when they reach the edges of tectonic plates (like in the middle of the Atlantic)? I'd imagine the heat would melt the cables quite easily.

The graph showing the shift of undersea cable shares from public internet backbone to private companies is concerning.

Yes, very much so, in particular in conjunction with attacks on net neutrality. I hope at least some regulators/politicians stay on top of this.

I always wondered if there are not problems in the spot a new cable is laid to cross an older cable.

Unlike data cables which cary information with a voltage, fiber optic cables carry information through light impulses. An advantage being that light is not susceptible to electromagnetic interference or radio frequencies. Having multiple fiber optic cables next to each other doesn't cause any issues.

The electromagnetic interference between cables is the reason why you shouldn't wire a house with data cables and power cables next to each other.

I think his thoughts are more like "what if you need to do something with the cable that lays under the other one?" At least that's what I'm wondering about. ;-)

Yes exactly

I'm not certain but it seems if two wires were crossing and you absolutely needed to move the one under you could cut the wire and fusion splice[1] the two ends together afterwards.

[1] https://en.wikipedia.org/wiki/Fusion_splicing

Can someone just dive and cut cables?

> Lasers propel data down the threads at nearly the speed of light

This is funny. Why “nearly”?

Because no vacuum

Might be weird wording for the fact that speed of light in glass fibre is slower than in air or vacuum.

Fiber optic cabling won't transmit at the speed of light due to the refractive index of the medium. So it's slower than the speed of light in a vacuum or atmosphere.

When people colloquially say "speed of light" it understood to mean the speed of light in a vacuum. Light travels slower the higher the refractive index of the medium its travelling through. Saying "nearly the speed of light" indicates that the light is travelling through a medium with a refractive index very close to 1.

To say that light travels at “nearly the speed of light” in any medium is a bad choice of words. Also with refraction coefficient of 1.40 - 1.45 in fiber, the term “nearly” is not appropriate if the intention was to compare with vacuum.

The first undersea cable was laid in 1858 and what's funny is the two guys on either end couldn't agree on voltages and currents etc. and would destroy each other's equipment. It seems silly until you remember that the only way they could communicate is via post, which took 7 days, or the very cable they were working on.

Reply on news.ycombinator.com