Hacker News Reader

Comment by geerlingguy | Hacker News Reader

I was able to place an actual order just last week in St. Louis, MO.

I will be using it and doing some tests in an area that actually has cable Internet (though spotty, and until very recently was limited to less than 20 Mbps up, no matter how much you paid).

But you have to realize, when you look at the $600 initial cost and $100 monthly fee—this will bring 100+ Mbps down and 15+ Mbps up to so many places (even in suburbs) where people couldn't get more than 10/2 or so for any price.

I already know a few people who don't even live in rural parts who would happily pay triple just to get off their junk DSL or flaky Cable connection, even with Starlink in beta. They are already used to spotty Internet, so Starlink is a massive improvement even if it's spotty (which it hasn't been, according to these users).

View on HN

Let's see how well it copes with more users. My guess is it won't cope well. There is only 20gig/sec of capacity per satelite, and I believe that is raw speed, which will get hammered in poor weather conditions and more FEC or less efficient QAM is required.

20gig/sec is not a lot of capacity for many thousands of square kms. It's 1-2k users at peak times watching Netflix for example on 1080p or 4K streams, nevermind everything else. It is probably not far off many DOCSIS CMTS allocations, which get horrendously overloaded being only a few city blocks.

My guess is after beta they will get rid of the unlimited aspect pretty sharpish, and probably introduce aggressive traffic shaping.

It is likely to work very well in very rural areas with no major cities or towns in line of sight of the satelite. However, for suburbs/exurbs close to cities and towns which have poor internet I think the experience will be pretty poor once it has more customers (especially in Europe. There are basically no truly remote areas).

Starlink will be a huge improvement in rural areas.

That's mostly OK though. It makes no sense financially in cities anyway.

I would love to get Starlink just because it's awesome, but I'm in a city and can get 350gbps broadband for £50 a month so it would be kinda pointless.

Could they add more satellites to increase bandwidth? I can imagine a continuous stream that passes over multiple cities. Urban ground stations would have a choice of 5-10 satellites visible at any time.

That's not true. Viasat and hughes cover everywy in the USA with speeds faster than that.

Source? I'm not from US but I googled around and Viasat offer 30/3 mbps with a soft speed cap after 100GB for 150usd/month [1] Hughes offers 25/3 mbps priced per data cap 50GB is 140 usd + equipment lease of 15usd /month.[2]

Neither of which is even remotely close to Starlink's offering.

1 - https://promo.viasat.com/plans/standard-30-715T5-11796UC.htm... 2 - https://www.hughesnet.com/get-started

OP said "where people couldn't get more than 10/2 or so for any price."

Your response directly refutes that, so I don't know why you are asking for a source. I didn't say either of those are competitive with a non-existent service offering from starlink. Wait until they have a real service that is oversubscribed, then we can discuss what plans they'll offer.

Hughesnet and other geostationary providers have 400+ Latency that are rather terrible.

Again, OP was talking about speeds, not latency. Nobody is going to refute geo has higher latency than Leo.

400+ms pings make the connection useless for anything resembling interactive use, which is pertinent.

If my internet provider had 100mbs symmetric but I got that kind of ping I wouldn't even consider it a functioning internet connection, let alone 100 mbs, even if in some theoretical tuned workload I could actually move data that fast (UUCP with tuned TCP window sizes? UDP with some sort of custom packet loss recovery algo?)

I'm not sure why you keep going down this rabbit hole, but OP was talking about bandwidth, not latency. Bandwidth on current generation satellites is just as high, or higher than Starlink. Satellite does not use TCP over the air. They have technologies for higher-latency links that make latency appear lower by controlling the window at the gateway side.

With latency >600ms to the first hop

To the first hop? 600ms is the round-trip latency of geo.

Indeed, so 600ms to the Ground station (and then normal latency on top of that)

No, 600 is the total. The propagation delay just to and from the satellite is roughly 500ms. The rest is internet latency and queuing delay.

The real limitation of rural satellite services are the data caps. When I was looking at options that was a serious limiting factor. All the bandwidth in the world doesn't matter when you hit your data cap after 2 netflix kids' shows.

Starlink will have data caps as well. All models show that when loaded with the amount of subscribers they're targeting you will either get very low speeds (no throttling), or they will need to cap/throttle the heavy users. I'm not sure why people keep misunderstanding that the beta performance/price/rules are exactly what will be there for the real service launch. It's simply not feasible.

UH???

Serious Question: So are the early adopters subsidizing the masses down the line?

i.e. so at $100/mo + $600 start-up/set-up fee -- thats for early adopters ... so the "african farmer who can pay <$1/mo for the service can afford it... and corporate subscriptions are obv going to be big bulk data contacts?

And the encryption policy? The Privacy policy? The data-mining-of-location-history policy?

Yeah, it's a Nope from me dawg...

And I folow Elon closely, and I like a lot of what he says and stands for - but he is at the behest of the USG - Starlink must not be trusted.

> And the encryption policy? The Privacy policy? The data-mining-of-location-history policy?

What's the difference between this and any other ISP?

Many folks don’t know or forgot that there are ISPs in the US that do deep packet inspection and inject banner ads into http traffic.

Seems like this has to be giving dwindling returns with so much of the web being on https these days.

One thing my current ISP still does is returns an ad-laden search page any time you visit a domain that doesn't resolve.

Super annoying. FYI every ISP I’ve had that does this has some setting buried somewhere that lets you turn this off.

If not that, you can usually bypass this by setting your router (or computer directly) to use 8.8.8.8 (Google dns) or 1.1.1.1 (cloud flare dns) as a resolver.

I’ve only heard of a couple of cases where ISPs sniff all dns and rewrite NXDOMAIN responses, and the answer for those is some kind of unpleasant tunneling.

Does DoH help with this?

Yes, it would mitigate it, until ISPs outright block all known DoH endpoints (probably not feasible though, too easy to fire up another one.)

I wonder if you could block unknown DoH endpoints via traffic analysis. They probably have pretty unique signatures, with every request/response being approximately the same size.

Not that I want to give anyone ideas.

I haven't looked closely at DoH, but based on my knowledge of plugable transports and website fingerprinting attacks, the answer is: probably, but with just a little effort on the client side, detection could be made pretty expensive -- probably expensive enough that ads on unresolved domains wouldn't justify it. Basically, if live classification of network traffic on (obfuscated) metadata were easy, China would have deployed it by now.

I've always wondered how muc CF paid to get 1.1.1.1

For the specific behaviors that GP mentioned, that's true perhaps.

But who is your DNS provider? If it's your ISP then they know who and where you are, when you're home, and what you browse...TLS or not.

It’s in the US but can have non-US customers, a potential one of which I presume is the GP.

Citations, please.

Can you share more info on 'the encryption policy' you're worried about? I'd not heard any cause for concern regarding encryption on starlink and was debating potentially signing up for their service.

It's an ordinary DIA link, except right now it's in cgnat IP space. There are no restrictions on crypto. You can run whatever VPN or crypto through it you want. I'm doing so right now.

Ipv6 is working you get a routable /56

Nothing about their privacy policy looks intrusive, at all. [1]

The data they collect is either necessary for them to keep (your name, address, payment information, etc.), diagnostic (IP address, HTTP headers, etc. ONLY on their own websites, plus quality of starlink connection e.g. latency and throughput), or optional (survey responses). They don't sell the data or share it with anyone unless necessary for legal or business reasons (payment info shared with payment processor and such).

[1]: https://www.starlink.com/legal/privacy-policy

I don't understand. Starlink is a narrowly tailored cool tech solution for internet access in remote or rural areas, not world hunger, inequality or evil.

> So are the early adopters subsidizing the masses down the line?

"Subsidizing"? No, they're paying for a service. It may or may not get cheaper over time... but that's rather irrelevant. SpaceX isn't a charity, they are in this to make money. It's also not like SpaceX is making a ton of money off of the early adopters. The satellite dish almost certainly costs the vast majority of those 600 dollars to produce, and the $100/mo service fees from the small set of early adopters aren't going to pay for the satellites and launches any time soon.

> the encryption policy

They don't prevent you from encrypting your traffic... just like any other ISP. What more do you want?

I don't know if they further encrypt traffic inside their network, but I rather assume not because that would be rather pointless. The traffic you send out should be considered more or less public because the internet will send it to malicious third parties regardless of what SpaceX does to it, see all the BGP highjacking attacks that have taken place for instance. SpaceX is obviously unable to encrypt things once they're out of their network into the internet at large... that's just not how the internet works.

> The Privacy policy

Is quite good for an ISP really https://www.starlink.com/legal/privacy-policy

Estimates for the satellite dish from people in the industry are well over $1000 (see e.g. [1] https://spacenews.com/news-analysis-spacex-has-a-lot-riding-...)

Unless SpaceX has made a huge unannounced breakthrough in phased arrays they're actually taking large losses on the up-front cost at this point.

Dead industry projects incompetence onto disruptive newcomer, news at 10.

More seriously, the phased array antennas seem to be a big part of Starlink’s innovation. Launching tons of satellites cheaply by itself isn’t a super protected moat because if they block competitors from using the launch platform at a reasonable price, they risk antitrust action.

I don't see antitrust legislation being a real factor to anyone. Several companies have gone trillion dollar market valuation and several decades without such action. Even if Microsoft had a crystal ball of the future legislation when they were starting up they probably wouldn't have changed a thing. Antitrust in modern history is a thing without teeth when used and almost never used.

(I say this as someone who 100% wants market regulation that actually reinforces the natural market efficiencies - including sane, forceful antitrust legislation.)

> don't see antitrust legislation being a real factor to anyone.

You’ll suddenly see it change when it’s a powerful company (Amazon) upset that they get boxed out of the only US launch platform.

They should buy Ubiquity and build extant mesh networks...

>*SpaceX isn't a charity, they are in this to make money.*

Exactly - but what are the EXACT revenue streams they expect?

How are they going to make money off the largest continent on the planet; Africa which also has the poorest population on the planet?

Mate, you are something.

And you're suggesting the encryption, privacy and data mining of location history policies exist and are diligently followed by current ISP's?

Your average African farmer may not get their own dish but I bet there are charities and NGO's who would be willing to pay for the dish and subsidize a local school's Starlink service. Throw a WiFi hotspot on the side of the building and now the village has access. This is a game changer that will open up the remote corners of the globe.

I used to work with an NGO that funded schools and orphanages in some particularly under-serviced parts of SEA. The retail prices of this would be affordable for a lot of institutions as it stands (places where the school fees were typically between $100-$200 USD/year). It doesn’t solve the problem of having reliable electricity, or any actual computing equipment. But I can certainly see some rather cool use cases for it.

Quite the opposite. VCs are subsidizing those $500 terminals that really cost $1000+.

Reply on news.ycombinator.com