all you can eat latency and an oversaturated network on devices with a limited lifespan.. what else could you ask for!
post
Hmmm ditch lightning fast and stable fiber for the mediocre speed and unstable micro satellite internet connection controlled by a petty asshole...
What to do, what to do?
Just think how much control he can have if he owns the medium which people access the internet.
And he'd only do good things with that power /s
"Give me all your money" says world's richest person, in a fit of originality.
Basic physics says satellites using Ku-band or whatever they use can't compete with fibre.
Satellite internet has its uses like for ships at sea.
Or rural and remote areas. I think it would be a ton of fun to go backpacking with a foldable solar panel and an antenna.
4G towers can have 100km radius
Fibre is an investment that can be used and upgraded for decades. Starlink is a subscription service forever to a private company.
And upgrading is piss cheap. Just change transceivers.
Same fiber cable that does 1gbps can do 100tbps.
That's not true... There are different types of fiber with different throughputs depending on the class of the cable and the length of the installation.
It is true.
Multimode (what I think you're trying to reference) isn't used in distance applications at all, it’s only for short in-building links. Anything that your ISP would provide you would be single-mode. Carrier/Backbone is virtually 100% SMF as well. SMF (OS1 and OS2) don't really have a bandwidth cap. It's all transceivers not the fiber.
But the point is that fiber that ALREADY in the ground, you can upgrade simply by changing the transceivers. It doesn't matter the length, SMF/MMF, or anything else... you just get a transceiver rated for the length of run (power of the led/laser, and the optics). The length is irrelevant otherwise as the presumption is that the install in the ground has been shown to work in the past already.
Old standard ITU-G.652 single-mode has been made to push multi-petabit transfers in lab environments. The only change was the transceivers. And to be clear, ITU-G.652 was standardized in 1984. Nobody rips out the fiber from the ground (caveat is that the cable itself hasn't degraded). You just upgrade the optics/transceivers.
“It's not the fiber that’s limiting—ITU-T G.652 defines physical specs (dispersion, attenuation), not throughput. Field trials over 96.5 km of real-world G.652 fiber showed 56.5 Tb/s using advanced DWDM and modulation
source: https://arxiv.org/abs/2108.01873
Your statement didn't specify an application!
The context of the discussion does...
SpaceX doesn't provide in rack or in-building connectivity.
SpaceX is an ISP. You wouldn't have an ISP running multimode.
ISP's absolutely run multimode. That's how you get fiber into a building or between buildings. Different types of fiber all play a role in a network deployment.
Broad statements are misleading. OM4 multimode won't push 10gb at 500meters no matter how good your hardware is.
Broad statements are misleading.
Ignoring the context of the discussion is even more misleading. In the context of this conversation, ISPs providing consumer connections and obtaining grant money, my statement is 100% accurate.
That’s how you get fiber into a building or between buildings.
You just said multimode can’t do significant speeds at distance, yet claim that buildings separated by distance would be connected with it? That logic doesn’t hold.
Intrabuilding or intrarack Yes, you’ll find multimode fiber occasionally. But even these rare cases are increasingly replaced by single-mode as costs drop and bandwidth needs rise.
Everything else (ISP deployments, backbones, FTTH) Single-mode fiber dominates. I haven’t seen a single ISP deploy multimode for consumer-facing services over a typical network radius (~hundreds of meters to kilometers). The only minor exception is MMF from the building network room to an apartment unit, which is irrelevant for this discussion and would be EXCEEDINGLY rare as most buildings would just copper line to the unit. But even in that case... the 20+km from the head end to the building counts for much more than the 20meters to the unit itself.
For all practical ISP purposes, single-mode fiber is what’s in the ground/on the pole, and upgrades are handled via transceivers, not ripping out the cable.
OM4 multimode won’t push 10gb at 500meters no matter how good your hardware is.
But just because you said it...
https://www.corning.com/catalog/coc/documents/application-engineering-notes/AEN075.pdf
and OM4 is suitable for distances up to 550 m
https://www.fs.com/uk/blog/om4-multimode-fiber-faq-highspeed-connectivity-guide-9499.html
OM4: Supports 10 Gbps up to 550 meters.
https://www.timbercon.com/resources/calculators/om1-om2-om3-and-om4-fiber/
OM4 Not specified 500 m* 150 m 150 m
*The IEEE has yet to officially give a distance for 10GBASE-S on OM4 fiber. The distances are decided by the IEEE in 802.3, not The TIA or ISO/IEC cabling standards. Some glass vendors say 500 m, but most are now quoting “up to 550m.”
You absolutely can run OM4 at 10gbps at or over 500m depending on your optics/laser.
But Multimode was never the point of discussion as the whole thread is based around broadband services (virtually none of it serviced by multimode, if any at all) and grant money for rural area coverage. Any fiber upgrade in this scenario will 100% be SMF with no qualifiers. In my past 30 years of IT career all buried and pole mounted fiber is SMF that I've ever seen for an ISP. I can tell you for certainty that ever fiber I've buried in the past 10 years for several companies has been SMF. I'm not even sure that I've touched MMF in the past 5 years even in intra-rack setups, I think I might have gotten some with a government auction win about 8 years ago I wanna say? With costs of SMF at near parity for the cable itself and getting closer every year in the modules... it's a dying form factor and was never really in use for ISP services to begin with.
My fiber ISP runs single mode to my home. And it means it's all single mode on the pole for the neighborhood.
Multimode just isn't used all that much anymore. We replaced all fiber links in our DCs with single mode recently. It's just cheaper, and as the other person said: Its far more future proof.
Technically that's true.
You just need to lay the 100Tb/s capable cable in the first place.
Going from the most secure, hard wired formats to a con man's satellites would be a fatal error. Any sort of military conflict and the network is all down, atleast broadband keeps secure networks intact.
A society grows great when old men plant fiber whose speed they know they shall never download from.
American taxpayers paid for both Starlink and Space X. Overpaid, actually, that's why he's the richest man in the world. None of his businesses are profitable, he just skims hundreds of billions off the enormous government grants he gets.
Since we overpaid for that tech, we should just confiscate it from him. He can be thankful that he doesn't go to prison for misappropriating government funds.
He can keep Tesla. It'll be bankrupt in 2 years anyway.
I say this as someone who actively pays for starlink out of necessity.
Fuck you, no. Fiber is much better for everyone. Eat shit muskrat.
I had Starlink for over 2 years while waiting for my fiber to be installed. Worlds better than the marginal DSL I had available before (5 Mbps down, 1 Mbps up), but I'm far happier with the fiber I have now.
I feel that man. Right now I load balance between tmobile and starlink cause the towers near me suck. I work from home so having consistent internet is really important and in my area, the fiber build out is really slow and expensive. Luckily I'm moving here soon but its been a pain in the ass to say the least.
Starlink is great for what it is. Very important tech but yea, I'm sure most everyone would be happier with fiber.
Company says that everyone should give them money and stop using competing products.
Obvious thing to say in the land of self-interest.
"Oligarch mouthpiece demands diverting of major public funds to oligarchs instead"
Story of America, really.
Nah we don't support nazis
glances at the state of the nation
What would things look like if we did?
Isn’t there an army walking the national capital, and masked goons grabbing people off the street into unmarked vans paid for with a budget /almost/ as big as Canada’s entire military?
Honestly, I think starlink is a fantastic idea in general, but this is clearly bullshit. Starlink works well in tandem with fiber, not as a replacement.
It's just never going to be as cost effective as installed fiber. Fiber is obviously the right technology to use in heavily populated areas i.e. for the vast majority of Internet users. And where the population is sparse and laying fiber for individual customers is cost prohibitive, that is where satellite connectivity shines. If SpaceX or anyone else is pretending otherwise, they're being blatantly deceitful and malicious. That's not in Internet users' best interest.
lets get down to the real reason he wants to do this. he would be able to turn off connection for millions if they piss him off, or hand over the data to said political actors like putin or trump, also to manipulate future elections like he did last time.
In France they authorised air hanging fiber, so they just use electric poles and hang the fiber under the 220 volt lines, as a last resort.
Cheap as hell. Or, where there's a will there is a way.
Starlink works well in tandem with fiber, not as a replacement.
It doesn't even work well in tandem.
Starlink has a single benefit going for it right now: Lack of uptake.
They only have a swath of spectrum, and that has a physical upper limit to how much information it can carry, in total. So does fiber. But, Starlink gets to share that with all users (Much like how cable internet works, its shared bandwidth for everyone on the loop). Fiber, you get your dedicated pipe.
This isn't even getting to view obstruction (A plane will cause a drop out), latency, jitter, etc. These are all physics problem that just cannot be solved without violating the laws of physics. Latency, at a minimum, is 2.6 ms, and that's just for the first leg.
It's crazy to say it doesn't work well in tandem... I mean, it's demonstrable, If it didn't work, people wouldn't use it, but they do. And there is no other way to reach users in some places. Starlink can reach users that only a long range wireless solution can work for. There are some other long range wireless solutions, but this one does work.
Look, I don't like Elon, I don't like monopolies, I'm not a secret shill for SpaceX, but I can admit the truth right in front of me. You don't have to stretch the truth to say Starlink isn't a good system for the vast majority of people, so why do it? Why create a false narrative? Why get all defensive about a technology?
And finally, I do not see any reason to care about an extra 5 ms latency.
And there is no other way to reach users in some places.
There is, if we decided to instead of giving Elon billions every few months, we used that money to expand the fiber networks.
Starlink can reach users that only a long range wireless solution can work for. There are some other long range wireless solutions, but this one does work.
There are myriad technology solutions that are both viable, and already being used. Capitalism means we don't deploy them. Oligarchy means we instead choose to do things that are more expensive, but happen to benefit a friendly oligarch.
You don’t have to stretch the truth to say Starlink isn’t a good system for the vast majority of people, so why do it?
Except, it isn't. Its just the one with the hype.
Some people live in places that aren't connected to large electrical grids, they have local generation and micro grids for a small community. Isolated mountains or small islands, or deserts are good examples of these situations. So if connecting to the electrical grid wasn't realistic I'm willing to bet that a fiber connection also isn't realistic.
It's hard to believe you think fiber can work for literally everything. I really don't know why you're bothering to dig in on this issue, it's so easy to prove otherwise. I hadn't even mentioned the use case of vehicles yet, boats, planes, trains, trucks, campers, obviously you can't run fiber to a vehicle. Or truly remote locations where people don't live, but researches work there, Antarctic bases, etc.
Also, I think you misunderstood my last line. I'm saying Starlink isn't right for most people. I'm just not making things up to say that.
So if connecting to the electrical grid wasn’t realistic I’m willing to bet that a fiber connection also isn’t realistic.
If fiber isn't possible due to electrical grids being non-existent... A power hungry sat transceiver will likely be a non-starter, too.
It’s hard to believe you think fiber can work for literally everything.
I don't think it can. I also don't think Starlink can work for literally everything, either. There are better, and faster, and cheaper solutions like Microwave backhauls and cellular data service for the last mile.
I hadn’t even mentioned the use case of vehicles yet, boats, planes, trains, trucks, campers, obviously you can’t run fiber to a vehicle.
Boats are the one outlier here, that cannot be reach via cell service, with a fraction of the cost of Starlink. And sure, boats can use it, and boats should pay the full cost of the package. No need for government money to fund them, they didn't need it before, and don't need it now. Boaters were quite satisfied paying their Iridium bills in the past.
Starlink still requires ground stations, and those ground stations can and are a limiting factor. I was up at a cabin that had Starlink, and service is still in the "better than nothing" phase.
There is concern for fucking up things like radio telescopes. Also, creating a Kessler syndrome event. "But LEO wouldn't have an issue with that because it would burn up". Two things:
- Everything in LEO being destroyed is still really bad. Astronauts would likely die.
- Objects in lower orbits can get ejected into higher orbits and hit things there. Kessler sydrome in LEO could potentially start a chain reaction in higher orbits.
Plus, the EU and China are understandably worried about Musk being the only game up there and want to deploy their own equivalent systems. So now there's not just one system of satellites threatening Kessler syndrome, but possibly three.
Just roll out fiber everywhere like we have with electricity.
While it is possible for objects in orbit to be knocked into a higher orbit, it's certainly not common. It basically requires a collision with another object in a highly elliptical orbit, this is not a kind of orbit we use very often.
Also, these low orbit constellations are simply nowhere near the majority of satellites, up in geostationary orbit. It's not realistic to imagine any debris from LEO ever reaching GSO, the distance between is just too vast. In general, Kessler syndrome would only extend downward from higher orbit, extending up to a higher orbit would be extremely unlikely.
Also, while astronauts could die, we keep enough emergency escape vehicles docked for the entire iss crew. NASA is full of smart people and they're generally risk adverse these days, I don't think anyone would die, but it would certainly be a shame to evacuate the iss.
Plus, the EU and China are understandably worried about Musk being the only game up there and want to deploy their own equivalent systems. So now there's not just one system of satellites threatening Kessler syndrome, but possibly three.
This is in fact a worrying situation. Not because I think Kesler syndrome is a realistic concern, but because there's only so much space in low earth orbit. I really don't like one company having a monopoly on low orbit communications, but having layers and layers of satellite constellations also seems like a dangerous situation.
Just roll out fiber everywhere like we have with electricity.
I'm all for that in theory, but whenever we dedicate funds to that cause... telecoms just walk away with it. If the US isn't interested in holding them accountable, I don't really see any reason to throw more money their way. That said, Starlink is doing fine, I see no reason to throw money at them either.
top 93 comments