Amazon Project Kuiper Broadband Constellation

[gongora]

[This topic split off from Amazon Web Services hiring engineers for “big, audacious space project” so each of the Amazon projects can have its own thread.]

[Geek Wire] Amazon to offer broadband access from orbit with 3,236-satellite ‘Project Kuiper’ constellation

Amazon is joining the race to provide broadband internet access around the globe via thousands of satellites in low Earth orbit, newly uncovered filings show.

The effort, code-named Project Kuiper, follows up on last September’s mysterious reports that Amazon was planning a “big, audacious space project” involving satellites and space-based systems. ...

Project Kuiper’s first public step took the form of three sets of filings made with the International Telecommunications Union last month by the Federal Communications Commission on behalf of Washington, D.C.-based Kuiper Systems LLC. ...

The filings lay out a plan to put 3,236 satellites in low Earth orbit — including 784 satellites at an altitude of 367 miles (590 kilometers); 1,296 satellites at a height of 379 miles (610 kilometers); and 1,156 satellites in 391-mile (630-kilometer) orbits.

[matthewkantar]

Any ideas about where the necessary spectrum would come from?

[gongora]

It will be interesting to see how a big constellation just getting filed now can play nice with the already filed constellations.  (Not that Amazon has a reputation for playing nice, and some of the earlier constellations may have U.S. approval but still be pending ITU approval, and the priority rules can be overriden to some extent like the FCC did in the U.S., but there are rules Amazon will have to follow to some extent.)

[ncb1397]

It will probably work similar to Amazon Whispernet now. Whispernet allows free access to Amazon.com and wikipedia. I guess this is the e-commerce equivalent of putting a brick and mortar in every city.

[DigitalMan]

I wonder how Mr. Wyler feels about this considering he contracted a bunch of Blue launches and actively avoids SpaceX.

Also, it seems likely to me that Kuiper is going to use the same frequencies as OneWeb.

[gongora]

I wonder how Mr. Wyler feels about this considering he contracted a bunch of Blue launches and actively avoids SpaceX.

Also, it seems likely to me that Kuiper is going to use the same frequencies as OneWeb.

Most of the constellations are using the same frequencies.  OneWeb has contracted a bunch of different launch vehicles.

[DigitalMan]

I see one of the links announcing the constellation has references to three ITC filings.  There is a lot of stuff in there, I don't have time to dig in today, but I think I saw a reference to 54 planes.

[gongora]

I see one of the links announcing the constellation has references to three ITC filings.  There is a lot of stuff in there, I don't have time to dig in today, but I think I saw a reference to 54 planes.

I didn't even notice those links in the article until they were pointed out.
The number of planes and number of satellites looks odd, I wonder if the forms are filled out correctly.

USASAT-NGSO-8A
https://www.itu.int/ITU-R/space/asreceived/Publication/DisplayPublication/8716
34 planes, 34 sats per plane
630km altitude, 51.9 degree inclination

USASAT-NGSO-8B
https://www.itu.int/ITU-R/space/asreceived/Publication/DisplayPublication/8774
36 planes, 36 sats per plane
610km altitude, 42 degree inclination

USASAT-NGSO-8C
https://www.itu.int/ITU-R/space/asreceived/Publication/DisplayPublication/8718
28 planes, 28 sats per plane
590 km altitude, 33 degree inclination

[jongoff]

This is interesting news. They're far enough behind OneWeb, SpaceX, and Telesat that I wonder if they'll be able to catch up. Definitely gives me another company to talk with about DogTags, MagTags, and our Bulldog servicing vehicle...

That said, I also wonder how this is going to impact the desire of commsat companies to launch on Blue Origin in the future. Now one of the big things going for them relative to SpaceX (that launching on them wasn't funding a competitor) is no longer true.

~Jon

[ncb1397]

This is interesting news. They're far enough behind OneWeb, SpaceX, and Telesat that I wonder if they'll be able to catch up. Definitely gives me another company to talk with about DogTags, MagTags, and our Bulldog servicing vehicle...

That said, I also wonder how this is going to impact the desire of commsat companies to launch on Blue Origin in the future. Now one of the big things going for them relative to SpaceX (that launching on them wasn't funding a competitor) is no longer true.

~Jon

A small difference in the eyes of competitors might be that Amazon has more money than God while SpaceX can't close a $750 million loan. OneWeb signing a large order with SpaceX might mean the difference between having another competitor and not while with Amazon it is nearly a foregone conclusion (at least as far as funding goes). There also is technically a firewall between Amazon and Blue Origin while there is no legal seperation between SpaceX launch services and SpaceX Starlink.

[Wudizzle]

A small difference in the eyes of competitors might be that Amazon has more money than God while SpaceX can't close a $750 million loan. OneWeb signing a large order with SpaceX might mean the difference between having another competitor and not while with Amazon it is nearly a foregone conclusion (at least as far as funding goes). There also is technically a firewall between Amazon and Blue Origin while there is no legal seperation between SpaceX launch services and SpaceX Starlink.

Alternatively, OneWeb signing a large order with a company that has never put anything into orbit may mean the difference between SpaceX having a competitor or not.

It cuts both ways.

[ncb1397]

A small difference in the eyes of competitors might be that Amazon has more money than God while SpaceX can't close a $750 million loan. OneWeb signing a large order with SpaceX might mean the difference between having another competitor and not while with Amazon it is nearly a foregone conclusion (at least as far as funding goes). There also is technically a firewall between Amazon and Blue Origin while there is no legal seperation between SpaceX launch services and SpaceX Starlink.

Alternatively, OneWeb signing a large order with a company that has never put anything into orbit may mean the difference between SpaceX having a competitor or not.

It cuts both ways.

Soyuz?

[Lars-J]

A small difference in the eyes of competitors might be that Amazon has more money than God while SpaceX can't close a $750 million loan. OneWeb signing a large order with SpaceX might mean the difference between having another competitor and not while with Amazon it is nearly a foregone conclusion (at least as far as funding goes). There also is technically a firewall between Amazon and Blue Origin while there is no legal seperation between SpaceX launch services and SpaceX Starlink.

The bolded is an assumption on your part. It might not be the same separation as Blue Origin and Amazon/Kuiper, but there will likely be some legal separation where Starlink is a subsidiary to SpaceX - or spun off with the same owners.

[RonM]

A small difference in the eyes of competitors might be that Amazon has more money than God while SpaceX can't close a $750 million loan. OneWeb signing a large order with SpaceX might mean the difference between having another competitor and not while with Amazon it is nearly a foregone conclusion (at least as far as funding goes). There also is technically a firewall between Amazon and Blue Origin while there is no legal seperation between SpaceX launch services and SpaceX Starlink.

The bolded is an assumption on your part. It might not be the same separation as Blue Origin and Amazon/Kuiper, but there will likely be some legal separation where Starlink is a subsidiary to SpaceX - or spun off with the same owners.

Space Exploration Holdings, LLC is listed as a wholly owned subsidiary of Space Exploration Technologies Corp. That means SpaceX owns Starlink but it is considered a separate legal entity. Once you find the legal name of a company, not the DBA, it's easy to find info in a Google search.

edit/gonogra:  SpaceX Services, Inc. is another related subsidiary.

[meekGee]

Server farms in space makes no sense to me. A server farm may use tens of MW and getting power and especially cooling would be much harder than on Earth.

And server farms try to be more local to where the customers are.   An orbital server farm spends most of its time over the pacific or the poles and needs a lot of extra communication bandwidth to make up for it.

Plus as you say power and cooling, plus the cost of maintenance and upgrades...

I just don't see it either.

I'm betting on an alternate constellation.  OneWeb (if it ever gets off the ground) can't compete with StarLink, and NG needs customers.

To me, with how things are, I can't see how Bezos hasn't gone after a constellation already.

Bezos has no choice but to meToo on constellations as well, this time for a good reason - it touches on his AWS revenue source.

The problems is, as was demonstrated with Iridium, launch costs are fundamental to the economics of a constellation and by 2021-2022, SpaceX will be using a fully and rapidly reusable RTLS system, whereas Bezos will be using (at best) a partially reusable system.

Any of these constellations will be requiring a constant launch cadence.  1/week or more.

What Musk can do is push ahead with Starlink 2.0 that uses larger satellites, taking advantage of SS/SH capabilities.

What Bezos can do is keep pouring money in, hoping to catch up.

It'll get interesting

[ncb1397]

Any of these constellations will be requiring a constant launch cadence.  1/week or more.

Really? 250 kg satellites, 100,000 kg per launch, 400 satellites per launch. 10 launches for deployment. They get replaced every 2.5 months?

edit: even 500 kg satellites, 45,000 kg per launch, 90 satellites per launch, 3236 satellites would be 36 launches or replacing the constellation nearly twice a year at 1 per week.

[Asteroza]

Any of these constellations will be requiring a constant launch cadence.  1/week or more.

Really? 250 kg satellites, 100,000 kg per launch, 400 satellites per launch. 10 launches for deployment. They get replaced every 2.5 months?

edit: even 500 kg satellites, 45,000 kg per launch, 90 satellites per launch, 3236 satellites would be 36 launches or replacing the constellation nearly twice a year at 1 per week.

You aren't using NG to replace one sat a week though...

[Asteroza]

I'm a big subscriber to the cloud inflation concept by Francis McInerney (Information Surfaces, defined by app membrane, app enablement membrane, and app delivery membrane), where pretty much everything is hanging off of or enabled by cloud services. The effective end-to-end vertical cloud integration aspect here (Project Kuiper, New Glenn, AWS ground connect, maybe those self driving delivery vans) for Bezos is kinda fierce. At this rate, it wouldn't be surprising for Apple to go deep with Tesla/SpaceX/Starlink to maintain the i-hegemony (iPhone, iTunes, iCloud, iCar), which leaves Google with Android+Waymo and grabbing OneWeb, but who's really left for a RLV for Google to latch on to (Skylon...)? Remaining cloud-centric operators are Facebook (which has expressed interest in serving the next billion users via sat, but has increasingly turned into a radioactive mess from privacy scandals so that partners are getting scarce). We have various standalone sat networks like Telesat and LEOsat, and Boeing being involved in funding a sat network would have a launcher/network tieup via ULA, but no cloud. There's also the dark horse of a chinese vertical integration push between major cloud companies (Tencent, Baidu, and Alibaba), smartphone manufacturers (Hauwei, Xiaomi, OnePlus), and RLV (LinkSpace, Long March variant?)

Plus the recent stuff with space based LTE cellular communication startups suggests an end run (expanding Amazon's Fire tablet ecosystem/Whispernet to an Amazon smartphone with MVNO style operation with sat backup)(using satcomm LTE for low priority bulk data, particularly bulk mostly asynchronous data from AWS, like streaming music/video services?)(It would resemble current broadcast satTV or satellite internet with the upload terrestrial modem and bulk sat download, perhaps with IP multicast and TCP multipath?). Francis McInerney also suggests that the installation costs for 5G rollout is such a huge captial expense due to the number of terrestrial smallcell basestations needed that some other paradigm is needed (he seems to think some sort of crytpocurrency-esqe network access profit sharing scheme to subsidize consumer purchased hybrid wifi accesspoint/cellular microcell basestation hardware connected to terrestrial fiber optic networks is the answer to the capital tables issues for cellular network carriers, but that is sorta iffy, even if such hybrid equipment for LTE is available now from integrated cellular/home internet broadband providers)

There's also the aspect of feeding CDN/edge servers/datacenters as well (which is mostly push out for CDN, increasingly less so for edge services).

An interesting case with vertical integration here, is the very fat pipe desired from AWS datacenters to their constellation. Will we see lasercomm domes on AWS datacenters soon? I'm talking in the sense that not just intersat lasercomm, but having multiple ground pointing lasercomm terminals on the sats, with enough separation distance on booms to be able to target a specific terminal from the ground without overlap (freespace ground-to-space lasercomm has gone over 1Gbps, but isn't at 10Gbps level yet, and amazon will likely want more than 10 depending on how much of that data is end user and how much is data traveling between AWS datacenters to cut network transit costs or for very high priority network traffic). The separation distances involved would favor an archinaut/spiderfab approach to increase sat size.

[meekGee]

Any of these constellations will be requiring a constant launch cadence.  1/week or more.

Really? 250 kg satellites, 100,000 kg per launch, 400 satellites per launch. 10 launches for deployment. They get replaced every 2.5 months?

edit: even 500 kg satellites, 45,000 kg per launch, 90 satellites per launch, 3236 satellites would be 36 launches or replacing the constellation nearly twice a year at 1 per week.

Maybe I'm behind the times but I remember 45,000 kg per launch, so yeah, 90 at most.  Less the launch mount, which is not negligible.  You're also constrained by orbital planes, so OOM, you're delivering one orbital plane per launch, give or take.

For smaller constellations (like the proposed Amazon one), 20-30 planes?
But for Starlink, 3x as many?  (I don't have the official count)

If lifetime is 3 years, you have to replace 33% of the constellation every year, after a 3-year build-up.
Or, you try to accelerate the build-up by launching even more often.

That's where launch capacity and launch cost will really matter.

Bezos is doing all he can to make sure that if SpaceX slips, he'll be ready to take advantage of it.  But if SpaceX doesn't slip, I think he's going to have a really hard time keeping up.

[ncb1397]

Maybe I'm behind the times but I remember 45,000 kg per launch, so yeah, 90 at most.  Less the launch mount, which is not negligible.  You're also constrained by orbital planes, so OOM, you're delivering one orbital plane per launch, give or take.

I was previously under this misconception, but nodal precession makes the 1 orbital plane per launch not a hard limit.

Nodal precession is the precession of the orbital plane of a satellite around the rotational axis of an astronomical body such as Earth. This precession is due to the non-spherical nature of a rotating body, which creates a non-uniform gravitational field. The following discussion relates to low Earth orbit of artificial satellites, which have no measurable effect on the motion of Earth. The nodal precession of more massive, natural satellites like the Moon is more complex.

Around a spherical body, an orbital plane would remain fixed in space around the gravitational primary body. However, most bodies rotate, which causes an equatorial bulge. This bulge creates a gravitational effect that causes orbits to precess around the rotational axis of the primary body.

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

If lifetime is 3 years, you have to replace 33% of the constellation every year, after a 3-year build-up.
Or, you try to accelerate the build-up by launching even more often.

Seems like short lifespan satellites. Iridium lasted about a decade and a half.