Hacker News Reader

Comment by _Microft | Hacker News Reader

The maritime coverage map shows coastal coverage around countries with inland coverage right now, i.e. they are bouncing the connection via one of their satellites to a nearby groundstation.

Starting in Q4/2022 they want to cover mid-latitudes around the globe. That might mean that they plan to enable inter-satellite links then? (This is a bit surprising - on one hand, this step extends coverage towards higher and lower latitudes but on the other hand not as much as they already have inland-coverage (cf. Brazil). It also extends longitudinally around all of the globe?).

Coverage above mid-latitudes requires satellites in polar orbits to join the network. (Their non-polar orbits have an inclination of 53° which means that satellites go no further north or south than that (plus a bit whatever their range is)).

https://api.starlink.com/public-files/maritime-coverage-map....

View on HN

If you zoom in on the coverage map, you notice that it's splines manually drawn by whoever made the map.

This isn't the output of some fancy RF model figuring out exactly where you'll get coverage. Nor is it circles of a specific diameter around ground stations. Or even hexagonal cells where coverage will be allowed/denied in their backend config.

In places it seems to follow countries territorial waters (which would be expected due to regulatory issues), but in others it spills out into international waters.

Overall, the map has sufficient 'oddities' that I think there is a good chance it's a rough hand made 'guestimate' coverage map, and won't perfectly reflect where coverage will really be delivered.

They need to have a minimum number of satellites before they can enable inter-satellite communication. Apparently its getting there.

Have a look at starlink.sx, it’s possible a few planes have enough sats to establish laser links.

Is it just me or is this map super heard to read? Three shades of similar blue?

Is it saying that in Q1/2023 the top of the earth (rest of Canada, etc) should be covered?

> Is it saying that in Q1/2023 the top of the earth (rest of Canada, etc) should be covered?

That's how I read it.

> That might mean that they plan to enable inter-satellite links then?

Or maybe they'll have stationary ships with "ground"-stations until the inter-satellite thing is working.

That wouldn't solve the problem. Those ships would still need a backhaul.

If they stationed them near the edge of current coverage, the ship could be the inter-satellite link until the actual inter-satellite links are working.

That would only buy you one increment of offshore range gain; it wouldn't cover the whole ocean unless you had a huge network of ships and used multiple up/down hops. And at that point the overall latency would suck so bad your customers might be better off with GEO.

It would depend on how fast the internal relay latency is on the ground terminals.

Geostationary satellites are up at around 35,000 km. Starlink is about 500-600km - you can get a whole lot of hops between Starlink satellites and terminals before you're looking at the same distance.

There's talk that they might be able to use either existing terminals, or a special set of terminals to do this, without the need to deploy them in fixed locations.

They're already in talks (trials, too I think) with Airlines to equip their aircraft with terminals. I'm sure commercial shipping companies would be interested too.

There's enough of those around to provide a massive network of potential relays.

But they are deploying inter-satellite laser links on newer launches, so those will come on and provide more coverage too.

>They're already in talks (trials, too I think) with Airlines to equip their aircraft with terminals. I'm sure commercial shipping companies would be interested too.

Would an airplane using Starlink be more easily tracked in an emergency to avoid MH370 type of mystery?

No idea.

They may not track/keep that information, but there's no reason they couldn't.

Inmarsat (satelite phone provider) actually helped reduce the search radius based on their contact with MH370.

> Would an airplane using Starlink be more easily tracked in an emergency to avoid MH370 type of mystery?

The thing that makes Starlink so appealing to airliners is that they can now do real-time telemetry of virtually every sensor on the plane without going bankrupt from the satellite data cost - not just for disaster recovery, but also for regular maintenance. Think of some random but unimportant component failing and the airline can dispatch a spare part and a repair crew to wait for the plane and do the maintenance right when it lands.

The idea is that the ships go up to another starlink and then down to a base station.

You could have all the maritime subscribers participate in a sort of mesh network, pingponging packets toward the nearest on-shore hub.

highly unlikely, the per day cost of operating a medium sized ship in an offshore environment with actual human staffing, fuel, maintenance, purchase or lease cost of the ship itself is extreme.

the one thing any satellite operator absolutely does not want is absurd ongoing monthly recurring costs to run their earth stations.

They are charging $5k USD per month per customer for this per TFA. Depending on demand (and 350Mbps downlink can generate a lot of demand) that may still be feasible.

Cargoships have operation costs in the 10s to 100s of thousands per h. Sure you could bring that down for what's required here, but you'd still sink a lot of money into these ships. How many maritime subscribers would there be?

$5k USD a month is significantly less than the monthly cost for any usable bandwidth allocation to a current tech ku band geostationary maritime vsat terminal, such as for the smallest "expedition" cruise ships, nevermind a big one.

Starlink should rent space on oil rigs at least those areas could have a "ground station" in a way. Oil rigs don't have physical undersea comm lines to land. But it can be a repeater since more power available on a rig.

Call me confused, but if these satellites are not geostationary, why are the both the inland coverage and coastal coverage mostly limited to political boundaries? Shouldn't coverage be available anywhere there are satellites overhead?

Just because something is technically possible doesn't make it legally possible, yet. They still need regulatory approvals where they operate because EM spectrum is a public resource.

It’s purely a regulatory issue.

Technically: yes.

Legally: no.

Like other mention, its regulatory.

But its also related to groundstations, the satellites bounce the signal down to land. They're transitioning to satellites with the capability to network between themselves which will reduce the need for groundstations.

I don't know how much they really limit coverage due to borders. Like if you get one in colombia and just move it to venezuela, does it still work? They dont have permission in venezuela but they might just not region lock it until venezuela actually complains or something.

I know for a fact that other sat internet providers do work cross border in this exact situation.

My family is using a directv antena bought in colombia and paying service in colombia, in Venezuela, so I'm pretty sure the market will try to do that as soon as they lunch in colombia

Normal satellite TV services work by restricting the area the signal is broadcast to. It has limitations on how precise it can be, and the receiver is pretty much passive.

Starlink is an active two-way system. Not only are they using high precision beam guidance on both ends, the receiver also has a GPS antenna and will report it's location back to the network.

Starlink is capable of turning off groundstations that are not where they're supposed to be. Not only capable, but for 'fixed location' groundstations (i.e normal service), they actually do block service if you move too far from where your assigned service area is.

How did you figure out that their connection isn't being censored by Venezuela?

The user terminals have a GPS receiver in them and would be able to determine where they are in the world with attendant precision. This would allow SpaceX to enforce fine-grained availability restrictions.

it's a political problem because right now you need a starlink earth station in a cooperative location (politically and economically) with access to decent fiber based terrestrial ISPs

for instance right now without fully operational satellite-to-satellite laser links, if you wanted to have live starlink services in afghanistan, you theoretically could, but you'd need to have a starlink earth station in somewhere like Dushanbe, Tajikistan. Or Peshawar, Pakistan. Or southeastern Uzbekistan. All politically problematic and very protective of their own domestic telecom companies.

Not the sort of place you can just drop a starlink earth station and buy a protected 100Gbps protected DWDM circuit to the nearest major city for an IX point, as starlink has done with their earth stations colocated with DWDM ILAs in the US pacific northwest.

Will constellation completion and intersatellite links expand the reach to polar areas as well? It would be a game changer for Arctic and Antarctic.

OneWeb already offers service above 50 degrees north, so it's not really a game-changer. It's the other game-changer.

My South Pole colleagues would love to have a better network.

Looking at the orbits for Starlink, they are currently not flying over the north / south pole directly. I don't know if there's plans to do so later, probably if the investments start to pay off and they can afford more launches and sattelites. But for now it's a big money sink and I'm sure the service isn't yet paying for itself - if it ever will.

> But for now it's a big money sink and I'm sure the service isn't yet paying for itself - if it ever will.

Given that the estimated cost of building and launching the satellites is estimated at only 600 million dollars [1], it is estimated to take something around two or three years until it's been paid off, followed by two years of generating profit and the next five-year cycle starts (assuming that their estimate of five years life time per satellite holds [2]).

Personally, I think Starlink was the best bet SpaceX ever made. That thing will be a permanent cash cow.

[1] https://www.forbes.com/sites/johnkoetsier/2022/02/14/starlin...

[2] https://subspace.com/resources/spacex-is-giving-the-internet...

If only - it doesn’t appear so. Connectivity in the Antarctic is laughable - McMurdo has 17Mbps iirc, and it flakes out not infrequently, Amundsen-Scott is “when a satellite is visible”, and being on a ship in the polar regions is typically even worse, as more often than not you’re dealing with crappy weather and a rolling vessel, and a poor and intermittent iridium signal.

It also all costs an absolute fortune - iridium is about $1000/mo for 100Mb, and $10 for each additional Mb - and there are surcharges for use in polar regions.

Musk could probably bankroll all of starlink just by serving the scientific communities in Antarctica.

There's a polar shell and there's a 70 degree shell, both with lasers, in addition to the 2 ~53 degree shells.

Not all the satellites are the same, some are enabled for inter-sat communications

Yeah I don't see this getting pushed back at all lol

Reply on news.ycombinator.com