The Dunant subsea cable(cloud.google.com) |
The Dunant subsea cable(cloud.google.com) |
(And I assume that Google has enough demand: If it didn't, why would they build such a large cable?)
No. Good market socialist solution in situations where network investment (electric grid, railway, telecom) creates natural monopolies, is forcing separation of the network and content.
For example electric grid owner must allow other sell and buy electricity trough the network. They can only get maintenance fee determined so that it cant be used to distort energy markets in favor of the company owning the grid.
In telecom it usually applies only for the last mile.
Not a big deal, but...sheesh. It's not like it was a troll comment; just a relatively lighthearted poke.
They are losing billions because they are paying for growth. It is the proper strategy.
This is what happens when a marketing company starts a cloud right? turns it into a loss leader and everyone who buys it becomes and apologist at all cost.
i don't get it.
GCP is not a loss leader. the unit economics are fine.
long click the back button, a popup will show your navigation history and you can click the last link before entering the broken site.
That said, the behavior is absolutely unacceptable.
[1]: And in Firefox desktop you can also do this but I can't remember if it is long-click or right click.
How the great have fallen...
So, we rang up, a sales guy picked up the phone and got a million pound pay day, and resigned that evening.
But our customers were happy so thats what counts :-)
Good times - they were a wonderful company, thank you
Reference: https://en.wikipedia.org/wiki/T-carrier
Light is only traveling at around 2/3 of speed within Fibre.
The previous decades have been around Bandwidth. Is time we shift out focus to latency. 5G is already part of that , and 6G is further pushing it as standard feature. I wish other part of network start thinking about Latency too.
May be not network, but everything. From out input devices to Display. We need to enter the new age of Minimal Latency Computing.
I also wonder what kind of permissions and licenses you need to seek to run a cable across the ocean floor?
https://cloud.google.com/blog/products/infrastructure/announ...
I presume that the other trillion-dollar companies are getting in on the action too.
[0]: https://zeenea.com/metacat-netflix-makes-their-big-data-acce...
1 - https://blogs.loc.gov/thesignal/2012/04/a-library-of-congres...
What is the limit on how many fibres can go in a cable? Should we expect future cables to have 50 fibres, or 100, or 1000, or more?
I suspect that the repeaters and associated power equipment along the line is pretty big stuff. So the fact that this cable is able to "share" that equipment across the 12 fibers is a breakthrough in technology.
The "state-of-the-art" AFAIK is to use many wavelengths per fiber, each one carrying ~192 wavelengths each wavelength transporting at up to 100Gbps (this is known as DWDM).
So so with SDM, you just have more fibers? So what? It seems like I am missing something here? Why is "SDM" the key concept rather than "DWDM"? Why not just say DWDM with 12 fiber pairs?
Multi-mode fibers are not feasible for long distance transmission. For long distance communications, using suggested approach, may be better to use multi-core fibers.
[1] https://siliconangle.com/2013/07/19/how-the-nsa-taps-underse...
A lot of surveillance is done both *illegally* and secretly.
Forcing carriers to install black boxes next to their routers is not always the preferred choice.
I'm sure a Shakespeare play or The Great Gatsby are barely a few megabytes.
But if you asked Joe Shmoe on the street "In Great Gatsbys, how big was the last picture your iPhone took", they would rightly have zero idea.
It's so useless.
I used to use that metric when folks ask why it took so long to debug. Like, our project is 600,000 LOC and more complicated than any of his works. He didn't have it all memorized and neither do I. It's a metric PMs can understand.
So only the raw texts, probably. 10TB sounds about right for that.
[1] https://nplusonemag.com/online-only/online-only/the-library-...
Seems crazy since oversea transit (tcp & single-channel) is usually latency (or loss) bound.
I would expect it's better than going over public transit and legacy subsea fiber, but it would have been useful to see some comparison tests between POPs.
Damn. Anyone else just agog at this figure?
Well, yes and no. I recall they wanted to pursue hollow cables in the early days of optical cabling, but it turned out solid fiber was the answer.
(sorry, can't find a good reference)
So FTTC (Fiber Through The Core) is what you want.
I don't know of any being deployed long distance, though in principle they'd be really valuable for intercontinental backbones. Starlink fills a huge gap in existing infra, and there are places that won't see any sort of fiber, let alone fancy microstructured fiber, for the foreseeable future (or ever, obviously in the case of ships/aircraft). But the bandwidth isn't great. Each current sat does I think 20 Gbps, and though no doubt that'll increase over time that's literally orders of magnitude from this single cable alone. Having the sats support direct ground optical links for backbone usage might be interesting someday, but weather attenuation will never stop being a problem with that. Starlink is filling in the gaps for fiber infrastructure, not replacing it. They're complementary.
So I agree it would be great to see more advanced fiber deployed long distances and start to shrink latency for everyone, and interesting to know what technical obstacles remain if any (maybe a lot remain?). A 40% speed boost while still having massive bandwidth isn't nothing.
Each hop will add latency since signal needs regeneration. So it’s not clear to me a swarm of satellites is a real winner from a latency POV. Furthermore, given costs to put the constellation up there, it’s extremely expensive on a $/bit basis and not sure how it could compete against fiber.
The value of Starlink is providing service in areas lacking existing broadband infrastructure where the cost to provide service exceeds the cost of Starlink.
I mean input lag [1] is easily 50ms. But some of them requires software changes. And any thing software is expensive. The cost of this new Cable is only $300M. Hardware innovation is getting faster and cheaper than Software.
Webdevs: is there a reason why a page would be designed so that JS being on is mandatory? Especially for something as prosaic as a couple of paragraphs of text.
If you meant mandatory in terms of the actual medium requiring it, I can only hint to interactive applications, aside from that I don't think it would actually be mandatory.
Here's an idea: add some HN logic to automatically move a comment that begins with "TL;DR" to the top of the thread.
I think in the case of Google, it's because they've been told they are the best developers, the top 1% of SWE's, they went through rigorous interviews, are paid a small fortune twice as much as they would get at a regular coding job, etc.
So it's dick shaking. They need to show to the world that they're better than plain HTML websites, that they have a massive schlong, that they out-chadded the vast majority of software devs. Plain html? Psht, we can invent our own language, gonna put those six years of uni to work! Wordpress? This is beneath us! It has to be a client-side rendered JS-pulled-through-GWT behemoth because on my system it's... wait it's slower, but nevermind that it's technologically ALPHA.
edit: actually looked at the source, looks like a Polymer / Web Components website. I've had to work in that once, it was dreadful compared to libraries used by real people.
Is this due to more and more content simply generated by javascript frameworks?
Yes it's long, but it's so worth it!
And there are likely laws about not obstructing the shoreline.
https://www.cnn.com/2019/07/25/asia/internet-undersea-cables...
As for physical security, there isn’t much on the sea floor. There are various instances of nation states tapping cables due to the ease of access when it comes to actually “listening” to the data. Obviously the issue there is getting to the undersea cable.
https://www.theatlantic.com/international/archive/2013/07/th...
Multiple nations have specialised subs to tap into them. I doubt you'd find anyone willing to make such a guarantee. They are impossible to secure in any way and you need to rely on security assurances at different layers instead.
Now we just compromise the servers/routers. https://gizmodo.com/the-nsa-actually-intercepted-packages-to...
[0] https://www.geoportail.gouv.fr/carte?c=5.372494831681247,43....
[1] https://www.sigcables.com/index.php/cableliste/fiche_cable/5...
[2] https://twitter.com/jlvuillemin/status/1238414261774401537
[3] https://twitter.com/jlvuillemin/status/1238433769935319042
[4] https://twitter.com/jlvuillemin/status/1238479381145751553
It is generally buried either under the sand or inside concrete. But yes, there are places where you can get very close to these things if you know what you are looking at.
https://en.wikipedia.org/wiki/Cable_landing_point
Here is a pic of the landing for the US base in Cuba.
https://www.dvidshub.net/news/186633/uct-1-unit-choice-gtmo-...
https://media.wired.com/photos/59546c71be605811a2fdcfd0/191:...
Their Oregon to Japan cable, 9000km and laid in 2016, cost $300M.
https://www.computerworld.com/article/2939316/googles-60tbps...
The T1/T2/T3 and E1/E2/E3 hierarchies join at the STM-1 level: An STM-1 can be subdivided as 4 x E3s or 3 x T3s.
This means that on a EU<->US SDH link, an STM-1 can be demuxed into either E3s or T3s, so you can have both standards on the same fiber.
What's terrifying is that Google described each of their B4 sites as having 60tbps uplinks in 2017, growing at 100x per 5 years. So a 250tbps undersea cable is nice but when you think about it probably not enough to make intercontinental transfer too cheap to meter.
How do you splice a hollow optic fibre?
Current practice is to use erbium doped fiber amplifier or raman amp for boosting optical signal at long intervals for transoceanic runs. Given the complexity of spatial signal, I don't think a regular optical amplifier will work? I could be wrong, this tech is changing but submarine fiber-optics tech is necessarily conservative and slow moving.
While a cable is being tapped, there will be a suspicious change in signal strength, and various signal reflections will tell the cable operators where the tap is. Thats bad for a spy agency who want to remain undetected.
Instead, they break the cable in three points deliberately. The middle point is where they put the tap, and the spy agency will repair it. The points either side are simply so that the cable operators don't know where the tap has been inserted, and have to be repaired by the cable operator. That gets expensive, since it will typically happen 3 or 4 times for a new cable install (3 or 4 countries want access to the data).
Cable repair operations are typically public knowledge (they require specialized ships), so anyone who fancies can crunch the data and see how often a cable breaks in multiple places before being repaired to know how often it's tapped... Mediterranean cables seem to see the most taps.
Please don't make guesses. I'm aware of the tapping process.
> Thats bad for a spy agency who want to remain undetected.
Yes, this is inevitable and it's still extremely more stealth that plugging network taps in somebody's else NOC. Especially if the tapping is done illegally.
Gatsby + Netlify with a CMS-as-a-service like Contentful or Prismic will lead you to a good result. We made e.g. https://fox-it.com/ using that, its back-end is Wordpress but it's drained empty to rebuild the website. Note how it works without JS, the dropdowns don't work but they fall back to full page navigation page. Note how with JS enabled, all the content shows up instantly. This is how it's supposed to be done.
It's very much still happening. Metadata is enough for intel purposes, storage is ridiculously cheap and post-quantum breaks of key exchange is forever 20 years away like fusion.
https://www.nytimes.com/2015/10/26/world/europe/russian-pres...
https://www.theatlantic.com/international/archive/2013/07/th...
https://www.zdnet.com/article/spy-agency-taps-into-undersea-...
The second is from 2013; Google and others encrypted those comms shortly afterwards after Snowden revealed those taps. https://arstechnica.com/information-technology/2013/11/googl...
> "The traffic shown in the slides is now all encrypted and the work the NSA/GCHQ staff did on understanding it, ruined."
The third link is twenty years old, and no longer very doable for the same reasons as above. Anyone still sending unencrypted stuff along these cables deserves to get stung.
But Starlink will never match the bandwidth and reliability that fiber can do, nor is it meant to. So it's not a replacement, just another awesome option.
If they don't interact with the thousands of miles of earth between the source and the destination, they probably also won't interact with the receiver! :p Imagine the retransmission rates!
https://arxiv.org/abs/1203.2847
Anyone from an HFT firm who wants to look into a partnership researching a neutrino link to the CME data center feel free to reach out :)
I think it could become possible before too long to use this to transmit data. It would probably be a ~billion dollar project, but the HFT arbitrage market is essentially winner-take-all, and may be large enough to support this size investment.
The low detection rate isn't so terrible either -- one only needs the bits that are detected to be tradably-correct almost-all the time.
The hard part is arranging to make enough money to fund the accelerator and detector.
It just needs to feel gratifying.
https://golem.ph.utexas.edu/category/2014/10/new_evidence_of...
If the encryption used is flawed, they could see whatever they want.
They can see who is talking to who, and when:
But Snowden showed us that a lot of it is scooped up and warehoused. Maybe they can see your traffic in a decade or two?
They can't. That's why they call it "bulk collection."
https://en.wikipedia.org/wiki/MUSCULAR_(surveillance_program...
The most-likely outcome is a few happy geologists/geophysicists and a number of very-sad HFT underwriters.
https://en.wikipedia.org/wiki/Mohorovi%C4%8Di%C4%87_disconti...
US, sea, 183 m below the sea floor in 3,600 m of water: https://en.wikipedia.org/wiki/Project_Mohole
I long for the days when the US loved building infrastructure.
Neutrino detectors work by maximizing dumb luck through being both very large and very, very clean (low radioactivity). The transmitter-detector systems work by sending oodles of very energetic neutrinos at a well-defined time and looking for a rare coincident flash in the detector.
Also multiplayer gaming is rather popular, and that's just not possible with that much latency.
VOIP is a pretty terrible experience with that much latency too.
Well, as someone who grew up in the modem era(90's) and was trying to play online fps games. I cared quite a bit about latency. Normal people also like things to be quick you know :)
Based on that experience in the 90s to this day i want my internet connectivity to be as fast as possible and i'm willing to pay.
Low latency enables video/audio chat amongst other things and just a better experience.
A quick google search gives 600 ms latency for satellites(not Starlink) thats quite a alot. Also bandwidth is a issue with existing providers i think.
PS: I am on the waiting list for starlink.
Might want to check with Pythagoras on that one..
The further you are from the other end, the less additional distance the satellite adds on.
Consider a right-angled triangle with base length d and height 550, corresponding to transmission from a base-station to a satellite. The hypotenuse has length sqrt(d^2 + 550^2), so the difference in length between the hypotenuse and base is sqrt(d^2 + 550^2) - d.
This has a maximum of 550 when d=0 (i.e., shooting straight up), and decreases as d increases: https://www.wolframalpha.com/input/?i=plot+sqrt%28d%5E2+%2B+...
Alternatively, consider the triangle inequality: the sum of the lengths of any two sides must be greater than or equal to the length of the remaining side. This directly implies that the difference in length between the hypotenuse and base is less than or equal the height [base + height >= hypotenuse implies height >= hypotenuse - base].
1.) Think of the precision mirrors in the so often mentioned EUV-lithography equipment from ASML for latest generation chips from TSMC.
2.) Now imagine something like that on board of a satellite, maybe smaller.
3.) Have 2.) moveable with sufficient precision to bounce the rays from satellite to satellite in realtime, without having to regenerate them in any way for about 4 to 5 hops.
4.) problem solved by purely 'optical' mesh while signal is 'in orbit'.
kthxbaiiii!
More seriously, those mirrors are special because they use bragg reflectors to handle 13.5nm light. They're not special for their precision, nor their reflectance.
Setting that aside, the major problem with your proposal is that laser still have significant bream spreading. So the mirrors would need to be large enough to encompass a spread beam at every step, which adds weight and volume for both the mirror and the tracking mechanism. The tracking mechanism is particularly problematic because moving mass on a satellite affect the attitude, so you either need precision counterweights to null it out, or large reaction wheels.
Using MEMS mirrors instead would solve some of the mass issues, but MEMS mirrors have very limited tracking (typically limited to a single axis) which would probably render them impractical.
Far, far easier to just send and receive the signal at every step.
> Flatness is crucial. The mirrors are polished to a smoothness of less than one atom’s thickness. To put that in perspective, if the mirrors were the size of Germany, the tallest ‘mountain’ would be just 1 mm high.
edit: What I meant to say was rather something with that precision reflecting whichever wavelengths are used for laser communications. Which would be infrared, I guess? Or are we talking Maser?
edit: arrgh, forget it... one beam, reflected multiple times until 'end of the line', got it...(sigh)
1. For the deep ocean parts of the route, cables and associated equipment (such as repeaters) are simply spooled out from the back of the cable laying ship, to settle on the ocean floor.
2. For shallow waters, the cable is buried. This is done by dragging a plow along the bottom which cuts a furrow and puts the cable into it. The plow has an altitude control and a camera so that an operator on the ship can control it, and a magnetometer to check if the cable is properly buried behind it.
3. For areas where burying isn't practical but they anticipate ships will anchor, they use armored cable.
For #1, the costs are going to be the cost to operate the ship while it slowly spools out the cable and the cost of the cable. For #3, same thing, but with more expensive cable. For #2 I'd expect it is similar, except the ship goes a lot slower (about 0.5 knots when using the plow, compared to about 5 knots when laying surface cable).
Finally, there is this.
#4. At the shores, they need to avoid damaging reefs and other habitats, not wreck the beach, and things like that. The cable needs to be in conduits that are buried or anchored. And building those conduits needs to be done in a way that does not mess up the environment.
So what you've got then for a long cable project is two ends that present underwater construction projects, the shallow waters near the two ends where you have to bury the cable, and then the long deep ocean stretch where you are just spooling the cable out.
This suggests the costs are going to have a component that doesn't really depend on how long the thing is (the two ends and the shallow waters near the ends where burial is needed) and a component that is proportional to length (the long run between the two shallow waters near the ends).
At 5 knots, it would take about 1000 hours to lay the deep sea part of the cable. If the ship costs $50k/hour to operate, that would be about $40 million. (I have no idea what it costs to operate these ships, but Google tells me that big cruise ships cost about that much to operate, and I'd guess that a cable laying ship is cheaper).
Assuming the underwater cable itself is 10 times as expensive as regular cable, its about $150 million for 9000 km.
That's brings us to about $200 million for the deep ocean part.
Still sounds really inexpensive when I consider it contains a large number of repeaters and is meant to stay at the bottom of the ocean.
Edit: Forgot to write, I haven't run the numbers myself but I enjoyed your reasoning here, you put a smile on my face :
> At 5 knots, it would take about 1000 hours to lay the deep sea part of the cable. If the ship costs $50k/hour to operate, that would be about $40 million. (I have no idea what it costs to operate these ships, but Google tells me that big cruise ships cost about that much to operate, and I'd guess that a cable laying ship is cheaper).
Repeaters aren't terrible expensive, so they only add a few million to the total cost.
Looking at what I can find, it looks like way more than 10 times the cost.
The Google project was $33k/kilometer, so I don't think I could have been too far off on the cable itself. Looking at other undersea fiber projects, that seem about typical. For example, this one [2] estimated $27k/kilometer [1].
Here's an Alibaba seller with submarine fiber for $2000-9000/kilometer [2].
The submarine cables have an aluminum or copper tube around the fiber optics, an aluminum water barrier, and a sheath of stranded steel wires, and an outer polyetylene layer, with various other layers of mylar, polycarbonate, and petroleum jelly in between.
I'd expect the metal layers to be the most expensive parts. Looking at the cost of tubes or cables of those materials, it looks like each of those would be in the $1000-2000/kilometer range.
[1] http://infrastructureafrica.opendataforafrica.org/ettzplb/co...
[2] https://www.alibaba.com/product-detail/Submarine-Fiber-Optic...
Source: I've laid subsea cable.
The fun thing was the company hand book had a whole other section of T&C allowances etc if you worked on a ship.
Repeater design is inherently very, very conservative because if the repeater fails, the cable fails. This results in an outage lasting days, if not weeks, as a cable ship is dispatched to the failure location.
The cable ship has to trawl for the cable and pull it up to the surface. Then the cable is cut and replaced with a new section that includes a new repeater to replace the failed one. Expensive.
You can buy subsea cable for $10-$20 per meter.
EDIT: the cost depends on how many layers of armoring you require. Deep sea cable requires less, shallow sea cable more.
The hypotenuse is cos(angle)*base.
If you think about it at a minute if a sat is 500 miles up directly overhead that's the closest it ever will be, as it flies off the hypotenuse gets longer, not shorter.
So ideally you bounce off a sat overhead, (distance of 1100), any single hop will be longer, and to get across an ocean you'll likely need more than one hop.
Basically the sin(beam path) will will never be less than 550 and the length of the beam will never be less than 550.
Edit: Actually, this is always true: we are considering a right-angled triangle where the base is the horizontal distance from the ground station to point under the satellite, the vertical part is the 550 miles between the point under the satellite and the satellite, and the hypotenuse is the line joining the satellite and ground station.
> if a sat is 500 miles up directly overhead that's the closest it ever will be, as it flies off the hypotenuse gets longer
Yes: as the horizontal distance d increases, then the length of the hypotenuse (sqrt(d^2 + 550^2)) increases.
However, the difference between this and the horizontal distance (sqrt(d^2 + 550^2) - d) decreases.
-----------------------------------------------
If the angle from the horizontal to the line between the satellite and base-station is theta, then:
sin(theta) = 550/hypotenuse => hypotenuse = 550/sin(theta)
tan(theta) = 550/base-length => base-length = 550/tan(theta)
difference in length = 550/sin(theta) - 550/tan(theta)
[which simplifies to 550 tan(theta/2)]
We are interested in angles between 0 degrees (horizontal - corresponding to the limiting case of infinite horizontal distance between the satellite and base station) and 90 degrees or pi/2 radians (straight up): https://www.wolframalpha.com/input/?i=plot+550%2Fsin%28x%29+...
This is always between 0 and 550. The triangle inequality holds: for a single hop from base-station to satellite, the increase in length is never more than 550.
But as you point out, there may also be multiple hops.
> So ideally you bounce off a sat overhead, (distance of 1100),
This is the shortest total ground-satellite-ground distance, but as you cover 0 horizontal distance it is the worst case: the difference between the ground-satellite-ground distance and the length of the direct ground-ground line is maximised.
I think this is an over-simplification if we are chasing pedantics; There are cases where it will be more and others less so the slightly more precise wording might actually be "about 1100km."
To the larger picture: it seems we often lose that order of length on the ground due to existing network topologies and geographical limitations.
