EPISODE 1795

[INTRODUCTION]

[0:00:00] ANNOUNCER: Subsea Cables are high-capacity fiber optic lines laid along the ocean floor to enable global communication by transmitting data between continents. Spanning thousands of miles. They carry an estimated 95% of international Internet, phone, and data transmissions. Critically, these cables are vulnerable to sabotage by state actors as they form critical infrastructure for global communication and economic stability. Indeed, Russia and China have been implicated in activities targeting subsea cables as recently as November 2024, and experts warn that these networks are likely to be focal points in future conflicts, heightening geopolitical tensions.

Josh Dzieza is a reporter for The Verge and has covered the subsea cable industry and the strategic importance of subsea cables. He joins the podcast alongside Gregor Vand to discuss this visible and increasingly important network infrastructure. Gregor Vand is a security-focused technologist and is the founder and CTO of Mailpass. Previously, Gregor was a CTO across cyber security, cyber insurance, and general software engineering companies. He has been based in Asia Pacific for almost a decade and can be found via his profile at vand.hk.

[INTERVIEW]

[0:01:26] GV: Hi, welcome to Software Engineering Daily. Today, we've got Josh Dzieza. Very exciting to have you here today, Josh.

[0:01:33] JD: Thanks for having me.

[0:01:34] GV: Yes. Josh, you are a writer at The Verge, and I think a lot of our listeners will certainly know The Verge, and you've done a lot of things before that as well. So, yes, maybe, today's topic, we're really all interested about subsea cables, which is a pretty esoteric topic in many ways. But as we'll probably get into is incredibly interesting, especially for software developers, maybe not appreciating just how much cables that drive their daily life. But yes, I mean, maybe if you could just give us a quick background on how did you get into tech writing generally, and what led to this - you wrote a specific article for The Verge on subsea cables. Curious to know how you kind of got into that.

[0:02:15] JD: Yes, I've always been interested in technology and particularly the overlooked infrastructures behind technology, whether that's sort of the people working behind the scenes, or working on various platforms, kind of making them run. Then, I stumbled upon cables as a topic a couple of years ago, actually quite a while ago at this point. It was pre-COVID and I was talking to somebody who mentioned the repair ships behind cables. So, we're talking about kind of the cable system, the global cable system, and what happens when cables break. That was something that I've been kind of familiar with. Every once in a while, you'll see a story, some cable goes down. Someone says, sharks or something, and everyone makes jokes, and then everyone kind of moves on.

My outsider impression was that, this was an infrequent phenomenon, kind of a freak event. This person was saying, who worked in the cable industry, was like, "No, it actually happens all the time, a couple of times a week, and you just never hear about it because they get repaired so quickly. But we're having this issue with the companies that do repair, is that they're sort of getting squeezed, the ships are old, and there just aren't that many. There's like two dozen ships basically in the world that do this work." I thought that was really intriguing. I had no idea that it was kind of that small of a crew kind of making the global Internet work. So, I wanted to start looking into it. This was pre-COVID, and then it's a secret of industry, and then COVID happens. It just took a really long time to actually get to reporting the story.

[0:03:59] GV: Yes, super interesting. I think the title on it is like, the cloud under the sea. I think that's a good way to frame it. For the listener base today, I assume the majority of the listener base, don't have any real insights into subsea cables, maybe some have read even your article, and there's a couple others out there. But generally speaking, I'm assuming that the listener base, like myself, are very uneducated on subsea cables, and especially sort of in the context of software and what that all means.

I mean, just to maybe give some background, obviously, you wrote this article. It seems like you had some quite sort of deep access. Could you maybe just sort of outline who were you speaking to, to get all this knowledge? A lot of the article focuses around a ship. Were you on that ship or how did that kind of look?

[0:04:48] JD: Yes. It's a pretty small industry, the kind of cable laying and maintenance industry. There's fewer than a dozen companies, maybe a thousand people worldwide that do repair work specifically. So, everyone kind of knows each other. They've been in the industry for decades. So, a lot of the reporting was kind of talking to people who had retired recently and could speak more freely, or just people on the background or off the record to kind of learn a bit about its history. You have a few major fault events, like a cable failure events where you have a bunch of cables go down and it's sort of not routine, and you have to take these emergency measures.

One of them was the 2011 earthquake and tsunami in Japan, which did just incredible damage, killed lots of people, triggered a nuclear meltdown in Fukushima. A lot of that was well known, and what was less well known is at the same time, it wiped out more than half of Japan's trans-specific cables. So, you had network engineers who were scrambling to kind of reroute traffic, and you had all of Japan's international traffic on. A cable or two is very precarious. There was not a lot of redundancy. So, you had to basically get the broken cables back online as quickly as possible.

Typically, what would happen in a sort of cataclysm like that is, lots of ships would come in from the area to assist. But because of the nuclear meltdown, all the cable faults were sort of in the zone of possible radiation. So, no one wanted to go or were prohibited to go by their safety procedures. So, you had one ship for the first stretch that was tasked with fixing all the cables, and it's a ship that's based in Yokohama, Japan, run by KCS, which is a subsidy area for the telecom there.

I have been speaking with the crew, many of whom are still working, about their experience and they gave me access. Because it was an old incident, because some of the cables were sort of nearing retirement or already retired, they could speak a little more freely about what happened. So, I spoke with them about the event. I went to Yokohama and went on board the ship, the Ocean Link while it was in port. They wouldn't let me go on an actual repair, unfortunately. Mostly because, when they go on a repair, they have no idea how long it would take. So, like, it could be two weeks or it could be three months. We don't know. So, yes, I sort of toured the ship and interviewed the crew while on board there for a week or so. 

[0:07:28] GV: Yes, awesome. I mean, I think that really comes out in the article. Obviously, we'll have that linked in the show notes. So, let's maybe just sort of set the scene. I mean, can you describe why are cables so significant to the world? And then, maybe sort of just take that one step further, like, I guess to software generally.

[0:07:46] JD: Yes. So, the Internet is cables, basically. There's that Ted Stevens quote that he got a lot of flak for about the Internet being a series of tubes, but it's basically true, especially what makes it global are these cables that run. There's about 500, 600 systems globally that just run across all the oceans. They're quite thin, they're basically like a garden hose. At their core are hair thin strands of fiber optic glass that transmit just a tremendous amount of data. That's basically what makes the web worldwide, it's how - if you're sending an email anywhere across the ocean, if we're talking right now across several oceans, and that data is going on these cables under the sea.

They're laid by - I mean, they're quite literally on the bottom of the ocean, like a ship, just with a big sort of thimble on the back, goes across the ocean, laying these cables. For the most part, they sit there and it's generally quite safe there, three miles under the sea. Then, every once in a while, they'll snap because they're actually quite fragile, they're very thin. But anyway, going back to sort of the internet infrastructure component of this. Previously, a lot of it was kind of point-to-point, person-to-person, city-to-city communication like we're doing now. I think in recent years, and we're talking relatively because this industry's roots go back to the telegraph era, like a lot of these routes and a lot of these processes. Even the company's genealogies go back to the Victorian Era.

But more recently, a lot of the routes and data being transmitted are kind of data center to data center. You have giant user-generated content platforms that are syncing their libraries around the world, why you can get TikTok and YouTube videos from anywhere, globally, cloud services, making sure they're always online. When you land in a new country, and you can get your Microsoft Office, or Google Docs, or whatever, it's because all this data being synchronized globally across these cables.

[0:10:12] GV: Yes. I think you brought it out quite well. There's a sort of misconception that satellites play a huge part in this. Actually, that's just not the case.

[0:10:21] JD: Yes. Satellites play basically no role, the stats vary, but it's at the low end. Ninety-five percent of global trans-oceanic traffic is on these cables. So, the high end, it's 99.5% or something. Satellites, they just don't have the bandwidth. There's so much data going across these cables. So, you think about all the video content, all this sort of AI requires a ton of data, and satellites, they rely on the cable system too. Like at some point, the base station is plugged into the cable network. It's able to access the rest of the web.

[0:10:57] GV: Yes. I mean, this episode isn't about satellites or the infrastructure that's using that. But yes, sort of, I think it's good to point out that sort of even the satellites ultimately are sort of relying on cables somewhere, when someone is trying to get access. It's just that if they can't get access from a cable that's sort of near them for whatever reason, then, maybe the last mile can sort of be handed off to a satellite or something along those lines effectively.

[0:11:19] JD: Yes. I mean, we're seeing that they could be very good for last mile stuff, especially kind of remote islands, remote areas. But yes, the cable system is the heart of the Internet.

[0:11:28] GV: Yes. I think one big kind of theme of your article and is interesting just to touch on is the idea that the maintenance is the bit that kind of nobody really wants to put their name to or their money to. It seems like there has been a lot of building that's been done. Again, that gets done over the years. So yes, could you maybe just speak a bit to like, who has actually laid these cables in the first place? Then, I guess who's maintaining them? Then, we'll get on to the future stuff as well. But I'm also curious, how has the landscape changed in terms of who is laying them now say versus, I don't know, 20 years ago.

[0:12:09] JD: Yes, it helps to sort of get a sense of the industry's origins when it comes to repair. And sort of going back even farther, you have - I know it was originally kind of national telecom monopolies, and they did everything, and maintenance was just part of the cost of doing business. It wasn't kind of its own industry. They laid cables, they also fixed them when they broke. Then, around the 1950s, when you started having coaxial cables, the national monopolies realized it didn't make a ton of sense to just have a repair ship standing idle to fix their own cables, just because it happened relatively infrequently. So, they kind of pulled their resources, divided the oceans into zones into, AT&T will have a ship in this region, and it will repair everyone's cables that break here. 

Then, kind of around the turn of the century when the monopoly started to get broken up and went apart, you have the ship operator's kind of spit off into their own businesses. So, you get companies like SubCom in the U.S. which used to be part of AT&T, and now, they lay and repair cables. They're their own business. It's governed by private contracts. You lay cable, you sign a contract with them to lay the cable, and then repair works slightly differently. It's a little bit more of kind of an insurance like arrangement where you have - so it gets a bit complicated and unusual. But there are two ways of sort of organizing repairs now.

One of them is, you have a zone maintenance agreement, and they're kind of like the zones the old telecoms used to have. But you'll have like a handful of ship operators that are part of the zone, just governed by like a nonprofit organization that various cable owners will pay into and say like, we'll pay an annual premium, plus like a day rate if a cable goes down, we call in, and you're supposed to have a ship on standby that will go out in 24 hours. And sort of we pay whatever the day rate is. Now, you have other sort of more private arrangements too, where you'll have a cable owner that will just say like, "Hey, SubCom, fix our cables. We'll work this out privately."

That's just a very different kind of structure than you used to have. Now, you have ship operators that like, they're running a business, but it's not a great business. No one really wants to pay for maintenance they want to pay as little as possible. And they'll pay kind of enough right now to keep things being fixed, but not enough to sort of invest in new ships, or sort of grow, and expand. Particularly, we've seen - so maybe, like back track a little bit. You have this sort of privatization moment and you have, at that point, a cable owner could be a dozen different telecoms that will chip in to fund a cable being laid because they're just very, very expensive. 

But then, in recent, maybe the last eight years or so, you have the big content companies, and then, you call them hyperscalers. But you have particularly Meta and Google to say, "Well, we want our own cables." They need to keep their libraries synced and cloud services up, and so, they will just pay for their own basically or maybe with one or two partners to run their own cables. They're by far the biggest investor at this point.

The maintenance industry, they have mixed feelings about it. We can get more into the future of it, but they just feel squeezed. I think that's hard to negotiate against a tech giant like that, and they want to keep their costs low.

[0:15:54] GV: Yes. It fair to say that there - well, I guess that these hyperscalers, yeah, I think you're alluding to, they're less interested in maintenance and more interested in just laying new cables. But. I mean, I'm curious, would they rather lay a new cable over fixing one? I mean, from a cost perspective, that doesn't seem to make sense to me, but I'm curious like what you've seen.

[0:16:16] JD: Yes. I mean, I'm not totally sure what the global strategy is either - I think, I mean, and they haven't spoken to this directly, but talking to the cable industry, their perception of how the hyperscalers are thinking about it. Is just if you build enough cables, you can wait a while to fix one. You'll have enough redundancy. But I think also, there's just sort of kind of corporate incentives and mechanics. That like, whoever's job it is to negotiate the maintenance agreement with to keep costs low and they're not necessarily thinking about, is this the best long-term value? Do we need to be supporting the industry more beyond just kind of the couple year horizon of our maintenance contract?

[0:17:00] GV: Yes. So, we've talked obviously a bit about the fact we've got cables full stop and why we even have them, data running around the world, and it basically all runs on these cables. That sort of has led into why it would make any sense for people like Meta and Google to get involved because, yes, if a cable goes down, then - it's not like an electricity where sort of someone says, "Well, my country's - our city's electricity is down and it will be back in a couple of hours, and nothing to do with Google." But sort of, well, if it says, "Gmail is down right now," then actually, that does sit on Google having to explain what happened.

I think that reflects more on them than any other infrastructure. I think the thing that your article brought out was the fact that the repair process is both very interesting because it's actually not that different from how it's ever been done, but it's also quite an interesting process. I think that maybe also just helps the listener base maybe get an understanding of when their service goes down, whatever service is that they're annoyed about is not responding right now. Could you maybe just walk us through roughly what is actually happening when it's known that a cable has been damaged or even almost fully severed?

[0:18:17] JD: Yes. It's a sort of strikingly analog and physical process. When a cable gets cut, alarm will go off in the network operation center. Someone at the NOC will call their maintenance provider and say like, "Hey, we think we've got a cable down" and they'll start getting a ship ready. Also, someone at the landing station sort of where the physical cable comes on shore on either side of the ocean will try to assess where the break is. And they'll do that basically by shining a light down the cable, and timing how long it takes for the reflection to come back. That'll pinpoint the fault within a couple of meters. So, they'll start to get an idea of where in the ocean did this happen, what might have been the reason, was there a ship in the area, was there just an earthquake.

That'll give them an idea of kind of what they're looking at. The ship, meanwhile, is loading up whatever type of cable from their depot is the cable that broke because each one is different. And they're just sort of spooling it to the back of their ship, and getting the crew ready, and fueling. They're supposed to sail out in 24 hours. Depending where in the ocean it is, it could take a couple hours to a couple days to get there. They survey the terrain, try to get a sense of kind of where in the ocean, what depth, what the ground is like. Generally, I think one of the first misconceptions is that, it's all sort of robots and submarines, but basically, anything past - I forget what it is exactly, but not super deep after - you can't really use an ROV. That's really just sort of for near shore work, but a lot of the deep water work you have to just sort of use big hooks that you lower into the ocean basically.

So, the first step, and this is a little bit tricky to visualize, but imagine the cable is sort of on the bottom of the ocean, it's laid with very little slack. So, you can't actually just pull it up to the surface to fix. The first thing they do is they lower like a big knife into the bottom of the ocean and drag it across perpendicular to where they - like the cable path. They're watching the tension meter, and they can see it kind of increase, and then drop off. That means that they've sliced the cables, and now they have a clean cut. So, they reel back the big knife, they swap it out for like a hook of some sort. They have a bunch of different types depending on the ocean floor.

They take another pass perpendicular to the cable, and they hook it, and then they draw it up to the surface. They tie it off to a buoy, and then they hook the other side. They, then splice it, and we can go into more about that process, but it's extremely minute and precise. They splice some of this cable onto that end, they sail back to the buoyed off end, they splice that, they lower back down to the ocean.

The thing to remember is that, this is happening at just a tremendous scale. Like in Japan, they were dealing with about three miles of water depth. So, it took half a day to lower the first hook down to the bottom of the ocean, half a day to reel it back up. Like, they're going dozens of miles on each pass. So, this can take just like week, in the best-case scenario. It's a lot of work, a lot of very sort of slow painstaking work, and dealing with miles of high-tension cable, and various sort of high-stakes processes.

[0:22:12] GV: Yes, I mean, I think that was what was sort of from a pure - it's not quite software engineering, but it's just engineering full stop perspective, was the fact that, when a cable gets laid for cost, and then natural. So, utility reasons, it's pretty taut. There's not a little slack running around. When they cut it, I believe the effect they're adding more, and slack does get added. Then, they have to actually lay it in a way that makes it taut again. Is that -

[0:22:37] JD: Yes. I mean, it's kind of a remarkable engineering feat that on a scale of thousands of miles of ocean, you can get the cable length to such an exact length that you're able to lay it across this sort of variable ocean floor with so little slack. The reason you want so little slack is, one, the cable's expensive. But also, if you have a lot of slack, currents and things are going to move your cable around. It's going to snag a rock, and it'll break. So, when every time you do a repair, you've added some amount of length, sort of proportional to the depth you're working in, in order to bring it to the surface to do the repair. So, you end up with this big loop of excess cable, that then you have to tow to the side so that you can lay it back down in a way that there isn't a lot of slack.

When you look at a cable map on telegeography or something, they look like these straight lines point to point. There's a lot of variation in there anyway, because they're sort of routing around obstacles and things like that. But then when you have repairs, it's really quite a bit more of a zig-zaggy, loopy thing than the sort of clean line.

[0:23:42] GV: Yes. I think it was interesting you mentioned the time frame there, like a week is not unusual to fix a cable. There has - I mean, so maybe just to sort of start to move towards the players involved, and sort of the interests involved. I believe there was, in Vietnam, they actually had like a fairly major sort of internet slowdown for a while because of some severed cable or cables. I mean, that sounds like there was probably more than a week maybe for that one potentially. But in general, we don't hear of say, the U.S. having a slowdown because of cables. So, it sounds like there's sort of the stakeholders of who owns the cable or who's paying to repair the cable has a fairly sort of significant impact on things.

[0:24:23] JD: Yes. I mean, you have a couple of things happening there. Vietnam is sort of a notorious problem area, because the water is shallow, and there's a lot of fishing, and so cables get cut a lot. And yes, you periodically will have multiple ones go down and they won't get fixed for months just because ships are busy elsewhere. That happened fairly recently. The other times that you hear about it are when an island gets cut off. You had Tonga a couple years ago with a volcano and cut its one cable. They're totally offline and sort of more poorly connected regions. This happens somewhat frequently in Africa, you have the Congo River Delta that has a lot of submarine landslides and will periodically knock out whole regions.

The U.S. is in a privileged position for a couple of reasons. One, it's just big, so a lot of the cables sort of go over land, which also break all the time, but you have some more sort of options to route around. A lot of Internet is based in the U.S. You have all the data centers. All the big tech companies are here. It would be a problem if all of our cables got cut. It would be a problem for just the entire planet. There are so many sort of routing, either through the Atlantic or the Pacific that it's not a huge risk short of kind of external aggression, like some sort of deliberate attack.

But it's not to say that we're immune, like there's not actually as much redundancy as you would really like a lot of the cables just sort of go into New Jersey, and New York, or kind of LA, San Francisco, Oregon. So, they all sort of go to the same place, but compared to an island or a region that's sort of getting more connected, there's a lot of redundancy. But I think the big thing is that, the Internet being located here is a big factor. It's hard to say sort of what all the different software dependencies are and what would happen if a place were to get cut off. There was one interesting study that I found that was looking at a couple EU countries while I was doing this story.

It was just illustrative, that even in kind of a big data center heavy place like Dublin, if you're loading a mobile app, you're probably pulling from somewhere over the ocean, you're pulling from data centers in Europe and the U.S. Sort of nowhere is totally immune, and so, you would expect that something would happen if the US were to be totally cut off. But that's kind of the reason you don't hear about it as much in the US, these sort of big slowdowns when you have a cable cut

[0:26:59] GV: Continuing this a bit, I'm trying to make some analogies again, sort of thinking, listener base, like what analogies can we draw here? I'm thinking, one here is data centers and obviously, today, GPU time that can be sort of rented effectively. So, you have these very expensive GPUs that people can book time on effectively. Is it kind of the same with the cable infrastructure? Just not from a pure, maybe time basis, but who are the stakeholders who are saying, well, you get this much bandwidth, and who are the ones buying the bandwidth, and divvying it out. How does that work?

[0:27:33] JD: Yes. This is a bit sort of more outside of my reporting zone. It gets very complicated, kind of depending on who owns the cables. I think it has changed somewhat or is in the process of changing. Like you just look at kind of the big content providers, and they used to just buy bandwidth from the cable owners. Now, they have their own cables. They can sell it to other people if they want, but they kind of use it for themselves. Then, you have all sort of intermediaries, various ISPs that then have their own customers. You have several levels of it. I'm not totally up on kind of all the different permutations it can take.

[0:28:18] GV: That makes sense. The only thing I knew at subsea cables before, sort of looking into it a bit more before this episode was just knowing that Chicago Stock Exchange was actually one of the sort of quant trading hubs to begin with because they had decided they were going to fund the cable between like London and Chicago, basically. So, it was actually not New York, it was Chicago. So, it's that sort of interesting thing where you've got a very private operator saying, "Well, for us to have an advantage, we're going to fund this cable and this cable can get us like millisecond trades as opposed to seconds" and that sort of changes things.

[0:28:50] JD: Yes, and then that's the whole other thing, is that you have what's called dark fiber, which is sort of non-public cable that entities will privately run high-frequency traders, governments. The U.S. military has their own sort of secret cables. But yeah, there's been a couple sort of cables that have been proposed or laid that are just slightly, slightly more direct route between trading centers just to get that little bit of edge.

[0:29:18] GV: Yes. Let's sort of move to like looking ahead, so to speak, or at least anything, also, post. When you wrote that article, I think that article came out in April. I believe like off the bat, there was a sort of major event that can fairly, reliably be attributed to like a Chinese ship cutting cables. Sort of, what impact did that have? I mean, in practical terms, but also like in sort of, how do you see that affecting how the cable landscape like is looked at from now on? Because that's quite a sort of - that's a pretty big move for a country. I mean, obviously, maybe you want to just describe maybe what happened and why it's so significant.

[0:29:56] JD: Yes. It's this is a super weird incident, and it's still kind of developing. I think as we're recording, the freighter is still just sort of stuck out in international waters surrounded by NATO ships. But you had this Chinese flagged freighter that kind of - the last report I read switched off its ship locator, and then dragged its anchor. Could have entered it to four NATO countries. I think it was Sweden responded pretty, and this is one of the more interesting aspects of it to me, is sort of how much the sensitivity to cables has changed in the last year or so.

I think it was Sweden, one of the countries sort of dispatched a patrol vessel pretty immediately to intercept the ship. Something similar to this happened, you know, I think it was a year or two ago and the ship just sort of left and people said, "Oh, maybe it was sabotage. We don't really know, and it was kind of too late to really investigate." But here, they responded very, very quickly. There's been conflicting reports about whether it was deliberate or not. It's sort of a strange situation. There wasn't any bad weather or anything that would cause a ship to maybe just lose its anchor like that for a long stretch of time. But also, it's just kind of a strange provocation.

If it were deliberate, I think it's previous port of calls in Russia, and so there's questions about that, but it's all sort of under investigation. I think in terms of disruption, it was pretty minimal. I saw some announcement that there were some ISP had like a thousand customers or something, and these are big customers. So, I think, traffic was rerouted pretty seamlessly, and I know a ship got dispatched almost immediately to fix at least one of the cables. But I think it's a symptom of the increased geopolitical sensitivity around cables, and this have been building for a while. But particularly in the last couple of years, there's just a lot of talk about the possibility of sabotage this kind of gray zone warfare around data infrastructure.

[0:32:02] GV: Continuing the just general geopolitical, but not specific to any sort of exact incidents. But the locations themselves, I mean, I think sort of I'm aware that places like Marseille have become like a hub for whether it's like laying the cable, or providers, and something along the lines of like because of Brexit, that Oman has become a strategic point as well. I mean, even if you can't speak to these exact examples, like what other kind of places come to mind when it's sort of these maybe slightly odd places have popped up actually as being very critical to sort of smoothing out the infrastructure based on now what we're seeing is like these cables are quite a, they're strategic asset more than anything.

[0:32:47] JD: The geography of this stuff is really interesting. So, you have major landing points, which are - I think it was Cornwall, or there's a point in New Jersey, there's some places in Northern California where it's just a lot of cables lands. So, those are just places of extreme infrastructure vulnerability. Like if something were to happen there, it could be really disruptive. Just because you'll have a dozen cables landing. Myrtle Beach is one that a lot of stuff is landing at now. So, these points, they're not necessarily near big population centers. They might be near big data center clusters, but they're usually, these landing stations are kind of anonymous looking shacks on the side of the ocean, a lot of like heavy power lines going to them. It's a little bit, it's always kind of been security through obscurity. They're not really marked.

I think there's been talk about kind of hardening them, and creating more security infrastructure around them, just because they're so vital to the global infrastructure. And then, you have kind of the repair geography, which is slightly different, where you have cable depots around the world. That basically just a port somewhere that is both central, but relatively out of the way in terms of, you know, out of the path of hurricanes in a geopolitically stable place where you can store a bunch of spare cable, and more a couple ships, sort of maybe just one ship.

They can be a bit random-seeming, like Curacao is one, just an island in the Caribbean that's out of the frequent hurricane path. There's one in New Jersey. Marseille is a big one. Then, there's various ones. The Yokohama is one for that region. I'm forgetting where the sort of major South Pacific ones are, but you have sort of these somewhat random-seeming installations where they have just a cable depot and a ship, basically.

[0:34:54] GV: I think it's really fascinating, in the sense of, I think most people can sort of understand how, for example, oil infrastructure works because, well, you've got a well somewhere and the well is there. So, we figure out how to get near to it. But the cable infrastructure is more, all these little variables that come in, as you say, like not near hurricanes. Then, obviously, we've got just a bunch of sort of political things of who's willing to fund a landing station where and all that kind of stuff. I think that's very interesting.

I mean, you sort of touched on it there, sort of two parts. One is the secrecy, so to speak, of where these places are. In the article, you mentioned almost the secrecy of the maintenance crews themselves. Then, just looking at, well, what's the path forward here? It feels like redundancy and resiliency is probably where things should be looking, where if the maintenance requires all this effort, to say, almost secrecy around it, because it's getting a bit tangled politically as well. What are you seeing in terms of how this is being looked at as how do we make these cables stronger, and more resilient, and redundancy, i.e. laying more and more cables to do exactly the same route, for example? What are you seeing in that sense?

[0:36:11] JD: It's interesting. It's kind of happening on a couple of different tracks. A lot of what's actually getting done is currently, is just private companies, basically, funding new cables, funding their own cables. That's not affecting the maintenance industry significantly yet, but it's increasing redundancy. There's more and more cables being laid all the time. Although, just sort of a side note, there are questions about whether it will increase redundancy if you have, that you could end up in a scenario and some models with fewer but higher capacity cables, which would decrease redundancy. Because then, if one of these mega cables goes down, it's a bigger problem than a bunch of smaller cables.

So, that's where a lot of the actual work is getting done. Then, kind of at the same time, you have governments that are increasingly talking about cable infrastructure from just sort of a security standpoint, and the importance of securing cables. It's unclear exactly how that's going to play out, because a lot of it is kind of like talking about how we need to patrol cable routes, and there's talk of deep-water drones and things to protect cables. But in terms of the actual threats to the internet, actual attacks like that have been really, really rare compared to a big earthquake or just sort of fishing. Fishing is the biggest threat. It's like someone dragging a trawl net across the cable. or an anchor, like we saw with this Chinese freighter, whether it was deliberate or not.

The biggest threat to the industry, the repair industry is just sort of lack of investment. The ships are old, crew, the crews are old, there aren't young people joining the industry in part because it's so secretive that no one really knows it exists to join. So, you have a lot of skill that's been acquired over decades aging out. So, those are the sort of more mundane threats than foreign attacker or something that I think people are mostly talking about at the government levels. 

[0:38:14] GV: Obviously, this is not a sort of political podcast, but it is interesting to think about sort of a government that's willing to cut cables, and then think, well, did they really know all the things that that cable that they're going to cut is going to do? And then, they realize, their country has lost access to some services that they desperately need right now. So, it's a bit different to sort of classic weapons, if you think, if you sort of look at it like that. There's maybe sort of an aspect of not really knowing what's going to happen.

So, actually, the likelihood that government goes there as their sort of way of threatening a nation. I mean, I believe, again, to bring up just China once more, something happened with a cable near Taiwan, and that was sort of deemed like the start of a sort of war games, if you like. But again, that maybe feels like they just - that's sort of the point. It's war games where governments would be less likely to really go big on this for many reasons.

[0:39:05] JD: I think that's right. One of the interesting things about cables from a national security standpoint is like, we haven't talked about it so much, but the financial system depends on these cables. All the central banks, all the hedge funds, and stock exchanges, they need these cables to be running globally to function. So, while cable attacks are talked about as kind of a gray zone tactic, that might be short of total war. When you look at the actual effects, and how unpredictable and total they would be of a big cable cut, it looks a lot more like total war. You don't want to just sort of wreck the entire global financial system, and possibly cut your own country off, and have, who knows what kinds of knock-on effects.

I think the different scenarios are sort of island or isolated cases in Taiwan is one that comes up a lot, where actually, if a siege of China were to begin, a siege of Taiwan or the war games, a lot of that start with cutting the cables because you could sort of isolate that impact and just cut one country off. But in a lot of other cases, it's hard to do.

[0:40:08] GV: As we're sort of starting to wrap up, just a couple of things. I guess, just maybe future on the technology side, like again, have you sort of been aware of how the pure technology of the cables might evolve, like anything to do like quantum communication effectively or anything like that? Like, how is that being assessed?

[0:40:26] JD: I mean, this is an interesting area where the hyperscalers, I think, are driving a lot of innovation in cable development. Like, you just look at the cables that are being announced, they're so high capacity. And that, you know, they're just spending a lot of money driving the research forward in how they do that, how you get more data out of whatever fiber optics you have in there, or how you lay new cables that have higher capacity. I think that's where a lot of the innovation is happening.

I think the maintenance and laying stuff, I mean, that's pretty much just how they did in the Victorian Era, and it is going to continue to be that way. I think for the foreseeable future, because you're dealing with sort of not - I mean, the forces you're dealing with, they're just the ocean and the geography. But in terms of what's inside the cable, that is moving forward quite remarkably. It does kind of impact the maintenance side when people will talk about, "Well, now, now we have 24 fiber repairs, 48 fiber, more and more fibers crammed into a cable, more sensitive." It makes the repairs more delicate and be like, "Now, we need to be stuck in place in the ocean for two days or something to splice the cable." So, that is something that they talk about, that is making maintenance a little trickier.

[0:41:42] GV: I think sort of, I don't see it all to play for, but it's sort of all very interesting to keep our eyes on, I think. The cables have not really come into the sort of limelight until maybe the last year or two years. I say, I think it's just dawning on, on everyone, just how much they run our lives, so super interesting. Just a kind of final closing question, just to you. When you did this research, obviously for the article, and you spent time with people, what sticks out as maybe one thing that was just the most surprising thing that sort of came out for you personally about any of this.

[0:42:12] JD: There were a lot of surprising things. I think the scale involved was always striking. Just how much time it takes to do everything and sort of how precise you have to be. Just working on a project where it's kind of every step could take a day or something, and how you have to be so focused on, what does the tension meter say, and what exactly do we think is happening here. I think it's an interesting example of, like, you're working with really sort of cutting-edge information technology on one level, but on the other level, it's just so, so analog. You don't really know what's happening on the bottom of the ocean, you're really just kind of fishing, quite literally just seeing what is happening with your line. I think that was a striking juxtaposition.

[0:43:06] GV: Yes, I remember this is some part in the article where it's talked about how exactly it takes a lot of time. It's a lot of just waiting around to see what surfaces literally. When it comes to the engineering side, it is a lot of - the longer you're in it, the more these little feeler points drive an engineer being able to say, "Well, they can make a judgment on something," which is, I think, again, it feels very in a nice way like old-school engineering versus today, when we get a response back in milliseconds and can fix something that way.

Josh, thank you so much for making the time today. I think it's been amazing to bring out a lot of the article you wrote into today's episode and kind of go beyond that as well. Where can people find you and sort of find more of your articles and all that kind of stuff? 

[0:43:53] JD: You can find me at theverge.com.

[0:43:57] GV: I think everybody knows where that is. Again, thanks so much for coming on, and yes, hope we get to catch up again in the future.

[0:44:03] JD: Thanks. Sounds good.

[END]