SpaceX just launched two of its space internet satellites | Hacker News Reader

SpaceX just launched two of its space internet satellites

theverge.com

121 pointsawiesenhofer70 comments

The article doesn't mention what I find to be the most interesting part of this mission:

From Ars Technica:

There is heightened interest in this launch because, for the first time, SpaceX will attempt to "catch" one of the two payload fairings that enclose the satellite at the top of the rocket. The value of these fairings is about $6 million, and recovering and reusing them would both save SpaceX money and remove another roadblock on their production line for Falcon 9 rockets. These fairings will separate from the rocket at about three minutes after launch and are "steerable" in the sense that SpaceX hopes to guide them back to a target located in the ocean.

"Going to try to catch the giant fairing (nosecone) of Falcon 9 as it falls back from space at about eight times the speed of sound," SpaceX founder Elon Musk said on Instagram Thursday morning. "It has onboard thrusters and a guidance system to bring it through the atmosphere intact, then releases a parafoil and our ship with basically a giant catcher’s mitt welded on tries to catch it."

https://arstechnica.com/science/2018/02/spacex-to-try-again-...

They didn't 100% succeeed in their first attempt at this, but got really close!

https://twitter.com/elonmusk/status/966692641533390848

I love this so much. The way Musk described it makes perfect sense. It's as if you had big bag of cash with a million dollars just falling from the sky, you'd probably go out there and try to catch it. It's the same thing.

Privatization of the space industry is going to lead to all sorts of efficiencies. I believe its also going to lead to some new areas of innovation that would never have been considered. For instance, this flight is putting 2 satellites in very LEO for internet service testing. Currently satellite internet service is at a much higher orbit which results in greater latencies.

ULA is a merger of boeing and lockheed martin. All three are private companies. This didn't lead to any efficiency in price. Multiple launch providers are what is leading to better price.

Exactly. Launch prices went down when countries other than the US started to offer payload services. They're going down even more sharply now because of increased competition, not necessarily privatization, though that has been a factor in increasing competition.

I really hope that they manage to get this working on one of the next few launches; they got so close this time!

From what I've read, they're approaching a situation where they are fairing-limited rather than core-limited, as a result of the fairings taking forever to fabricate for each launch.

> 40 million subscribers to the service by 2025, amounting to $30 billion in revenue that year.

$30,000/40/12 ≈ $63/mo

For internet in a remote location of the world? Where do I sign up?

More interestingly, how will Internet work on Mars?

It will probably have its own satellite network that serves as the Mars internet, with a high-latency intermittent bridge to the Earth internet.

https://en.wikipedia.org/wiki/Interplanetary_Internet

Where high-latency is a ping of somewhere between 364,000ms and 2,684,000ms depending on the date.

Tcp seems like a really bad protocol for this.

There already is a bunch of RFCs in place for high latency protocols.

Famously RFC1149: https://tools.ietf.org/html/rfc1149 . I suspect that wouldn’t fly in an interplanetary scenario though.

I think https://tools.ietf.org/html/rfc1217 might be more appropriate, and even touches on the relevant topic in section 5(e).

on Mars a NAS is must have.

We should call it the InterPlaNet if that ever happens.

https://en.wikipedia.org/wiki/InterPlaNet

Heh, cool :) Thank you for that link.

$63/month is the average though, so it seems like it will be really expensive for customers in developed countries.

As far as I understand you also need a relatively large antenna on your roof for it to work.

Depends on how they expect growth to go. They may design things for a break even of 100 million users with minor 'subsidies' to assist growth.

It also doesn't include revenue from backbone services that the same projections planned to form the bulk of Starlink's business. Consumer revenues (and thus average prices) are lower.

Currently you can get Iridium satellite phone with unlimited data minutes for 125/month. The connection speed is less than 10kbit/s.

$63/month 1Gb/s for a remote village, island or a ship with 10-500 users who share the bandwidth is not that bad. 40 million subscribers might serve 100-500 million people.

Please check this comparison:

http://www.globalmarinenet.com/compare-iridium-inmarsat-and-...

I don't understand how this network will work in practise. I can broadly understand the physical layer. There will be a big increase in bandwidth over other systems. More satellites and higher frequencies means more data in the system overall.

The link layer and the routing layer however are going to be key to how this system will deliver at such a large scale and with low latency. Negotiating links and routing is currently done with huge, power hungry switches in giant data centres. And those only have to deal with relatively static routes, with a satellite constellation, routes and constantly changing.

My best guess is that for the most part, traffic will go via intermediaries rather than from handheld devices. Having fixed base stations that communicate with the satellites means fewer links to maintain and easier negotiation if they maintain a fixed location.

Does anyone have any more technical information beyond what is provided by a paper by Samsung [1] on a similar space network?

[1] https://arxiv.org/ftp/arxiv/papers/1508/1508.02383.pdf

This isn't meant for handheld devices. Musk has mostly been talking about pizza box sized dish.

How practical would it be for radar stealthed balloons to act as floating cell towers? I wonder if such devices could be used to provide internet to regions under totalitarian control?

I think a receiver the size of a pizza box can still use the traditional pragmatic approach of hiding it on a rooftop.

You can still track signals originating from rooftops using drones and then prosecuting the people doing that for spying.

Iran is one of those countries that really don't like two way satellite internet.

Predator drones, or quadrocopter drones?

I thought that this could be a possibility, but has any government actually gone that far yet? This sounds a level too dystopian for the current time.

> but the vehicle won’t be attempting a landing after takeoff. (SpaceX has not given a reason for that.)

From watching the stream of the launch it seems it was an old model rocket that they don't need any more. I guess there is a cost to recovery

Block 3 booster that was already launched once.

I think it was mainly because they were going for a polar orbit.

They've landed cores out of Vandenberg before (ie. polar orbits). These landings were all on ships, as I guess their landing pad is not complete or not approved.

I saw the launch by accident this morning while walking the dogs. I watched the first stage separating and was waiting for it to fire again but nothing happened.

I wonder what the latency is going to be, in particular for intercontinental traffic. Will it be better or worse than with the submarine cables we use today?

Probably around ~50ms (see https://arstechnica.com/information-technology/2016/11/space...)

50ms sounds fantastic even for fast-paced multiplayer games.

That depends if it's average or median or 95th percentile, plus how much variability there is.

That's also true.

Don't have a link but in one of the ars article, the latency of full operating network would be around 25-35ms.

Yeah important note, per Ars, is these are very much low orbit, at ~800mi altitude, not geostationary which is ~22,500mi where Hughes is parked for example. So, good latency if you do 2 x 800/c. This is great if you happen to own a rocket company and have tons of low, cheap launches instead of one big expensive launch.

Another thing to consider, HughesNet is up at 22,500mi but also at the equator whereas StarLink is basically going to be much closer to straight up.

It should be better than submarine fibre for longer distances. The satellites will be in low earth orbit, and the constellation will be dense enough that routing ought to be pretty direct. You only get about 2/3 c in glass, compared to c between satellites in a vacuum. If the satellites are 500km up, so long as the distance is greater than roughly 2000km, you should get lower latency via the satellite path. Of course this assumes you don't lose too much to the MAC, and the satellites manage a routed path not too different in length to the route via the fibre (which often involves a detour to get to where the fibre comes ashore).

Of course, they did no such thing. Side note: will we ever have real journalists again?

They launched a pair of temporary test satellites to prove some of the concepts they plan to use. These will remain on orbit only for a short time.

I was at Vandenberg on Wednesday morning for the scrubbed launch - drat!!

> These will remain on orbit only for a short time.

This statement is directly contradicted by the FCC filing:

https://apps.fcc.gov/els/GetAtt.html?id=185534&x

> The designed lifetime of each satellite is six months. If this lifetime is exceeded, SpaceX plans to continue operation until such time as the primary mission goals can no longer be met, at which point the spacecraft will be deorbited.

The text in this particular filing, if read plainly, would indicate that SpaceX could easily let Microsat-2a and 2b stay in the air potentially through production deployment of Starlink if feasible.

Nowhere near the target design lifetime of the actual system satellites. No production functionality beyond test-bed.

Of course they will use them as long as they can; that’s consistent with what the FCC was told.

(Comment on the side note) It's all the editors. The journalism and content is there, its the silly click-bait title (normally editors job?) that muddles the journalism.

Come on, they said this in the actual article. "But the vehicle will also have two additional satellites hitchhiking along for the ride: prototype probes built by SpaceX to test out the technology needed to beam down internet from space.

Sending up these two test spacecraft — named Microsat-2a and Microsat-2b — is a big first step in SpaceX’s long-term plans to create satellite internet." Sure, headlines nowadays tend to the buzzworthy but this article seems well written otherwise.

Absolutely! The article said exactly the correct thing. The headline - something entirely different.

What are you saying is not "real journalism"?

I don’t know about most things - but what I do know, almost never matches what is reported. In this case, it’s pretty minor; as others said, probably not even the journalist fault (article was good), but an over-eager editor making a click-baity headline.

However, do we allow journalists to disclaim all responsibility for ensuring their good work isn’t sullied by salacious editors? I’d be pissed off,and perhaps they were.

Let’s just leave it at that - we are the dumb slobs left to clean up the mess when our family and friends are misled about the state of reality by editors/journalists muddying reality...

How can you say they "did no such thing" and then say they did exactly such a thing?

I’m super excited about what they’re doing; but the average reader would be reasonably expected to interpret this headline as SpaceX starting their deployment. It is no such thing.

If I went to a friends house in Lompoc and we went down to the shore would there be a decent view of the launch from Vandenberg AFB?

Good viewing area up the road past the main gate. There are instructions online to find it:

“There are two viewing areas for launches at Vandenberg Air Force Base (VAFB). * Weather Station: The public viewing site for Vandenberg launches is off of Corral Road near Vandenberg's main gate. To access the area, take Highway 1 to the Santa Maria Gate and proceed on Lompoc Casmalia Road. At the barriers, turn right onto Corral Road and bear left to the top. The launch countdown net and port-o-lets are set up near the bleachers.”

Oh, awesome. Thank you for the reply!

I've seen numbers on proposed download speed (~1gbps) and latency (~35ms) [0], but I can't find anything on proposed upload speed. Any ideas? I know offerings like Hughes net have claimed 3mbps.

0 - https://arstechnica.com/information-technology/2016/11/space...

Isn't it similar to existing tech being rolled out on airplanes?

Edit: Just read that it is using the same bands. Maybe it'll mean gogo will finally have some solid competition.

Launched with so much success, Elon is giving away thousands of dollars in ethereum /s

https://twitter.com/eiiilonmusk/status/966844047032832000

backup: https://imgur.com/a/8OsbX

(this is a third or fourth one I've noticed so far)

11,941 to go!

It's actually surprising Ajit Pai would publicly support competition when his entire schtick seems to be locking down on it.

Maybe... you've misunderstood him?

It probably indicates he thinks the project is a pipe dream. Since otherwise, it's a major threat to his monopolistic bosses.

> it's a major threat to his monopolistic bosses

I don't think he's being this cynical, but in what world is SpaceX competition to Verizon et al? SpaceX's satellites won't beam Internet to your phone. They'll beam it to the towers.

> but in what world is SpaceX competition to Verizon et al?

In the world of their fiber optic networks, which they refuse to build out.

Verizon might be one of the good ones. They've at least started to build them out and actually, in my opinion, is very good.

Sure, what they already built is decent, except for the lack of IPv6. But they simply don't want to expand anything.

It isn't, really. It's actually going to save them money. The system will not have enough aggregate bandwidth to really compete for any dense areas. Right now, for the major ISPs the dense areas are profit centers that are used to subsidize the rural areas in the states and counties that require an ISP to serve both if it wants to serve the cities. Starlink will not be able to meaningfully compete in the cities, while it will ease the burden of the rural net-negative customers.

Right, and it takes off the monopoly burden from those telcos so they are then free under the same Pai to ratchet up their underhanded techniques. If I were ATT and Verizon, I would fund Starlink.

Notice he says he's excited about providing high-speed internet to "rural Americans."

So it's not for anyone who lives in an urban area. He'll be sure to outlaw it because radio signals might interfere with "emergency services" or some BS like that.

Reply on news.ycombinator.com