Author Topic: Amazon Project Kuiper Broadband Constellation  (Read 194991 times)

Offline Cheapchips

  • Full Member
  • ****
  • Posts: 1044
  • UK
  • Liked: 902
  • Likes Given: 1973
Re: Amazon Project Kuiper Broadband Constellation
« Reply #100 on: 07/30/2019 08:10 pm »
Found few articles about power usage of AWS data centres. We are talks 100MWs for some centres. Once their satellite constellation is in place I can see AWS starting to placing data centres in space. They could well end up being biggest user of space stations.
I don't know specifically about AWS data centres, Azure ones will have at least 10's of staff on site daily to fix hardware and software.  Blue would need to get their manned flights up and running for orbital data centres have parity with an Earthbound one.

(Unless AWS centres are much better than Azure ones)

Offline Lar

  • Fan boy at large
  • Global Moderator
  • Senior Member
  • *****
  • Posts: 13469
  • Saw Gemini live on TV
  • A large LEGO storage facility ... in Michigan
  • Liked: 11869
  • Likes Given: 11115
Re: Amazon Project Kuiper Broadband Constellation
« Reply #101 on: 07/31/2019 01:59 am »

https://www.cnbc.com/2019/07/15/morgan-stanley-amazon-project-kuiper-could-be-a-100-billion-business.html

Amazon hired lot of management team from Starlink that Musk fired.
That's not necessarily a good idea since a lot of them were supposedly fired for not moving fast enough... oh wait.... No problem! They will fit right in.
"I think it would be great to be born on Earth and to die on Mars. Just hopefully not at the point of impact." -Elon Musk
"We're a little bit like the dog who caught the bus" - Musk after CRS-8 S1 successfully landed on ASDS OCISLY

Offline lonestriker

  • Full Member
  • ****
  • Posts: 417
  • Houston We've Had A Problem
  • Liked: 820
  • Likes Given: 5155
Re: Amazon Project Kuiper Broadband Constellation
« Reply #102 on: 07/31/2019 02:14 am »
Found few articles about power usage of AWS data centres. We are talks 100MWs for some centres. Once their satellite constellation is in place I can see AWS starting to placing data centres in space. They could well end up being biggest user of space stations.
I don't know specifically about AWS data centres, Azure ones will have at least 10's of staff on site daily to fix hardware and software.  Blue would need to get their manned flights up and running for orbital data centres have parity with an Earthbound one.

(Unless AWS centres are much better than Azure ones)

I don't understand the justification for "datacenters in space".  High energy usage demands huge radiators to handle the thermal load.  Look at what the ISS needs for their very modest power budget.

Just the launch cost would dwarf the value of servers, storage and network devices that would populate your space-based datacenter, so what's the actual value of putting the datacenters in space?  It can't be to save money on power or cooling as those costs would increase exponentially.

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.


Offline russianhalo117

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 8818
  • Liked: 4748
  • Likes Given: 768
Re: Amazon Project Kuiper Broadband Constellation
« Reply #103 on: 07/31/2019 02:47 am »
Found few articles about power usage of AWS data centres. We are talks 100MWs for some centres. Once their satellite constellation is in place I can see AWS starting to placing data centres in space. They could well end up being biggest user of space stations.
I don't know specifically about AWS data centres, Azure ones will have at least 10's of staff on site daily to fix hardware and software.  Blue would need to get their manned flights up and running for orbital data centres have parity with an Earthbound one.

(Unless AWS centres are much better than Azure ones)

I don't understand the justification for "datacenters in space".  High energy usage demands huge radiators to handle the thermal load.  Look at what the ISS needs for their very modest power budget.

Just the launch cost would dwarf the value of servers, storage and network devices that would populate your space-based datacenter, so what's the actual value of putting the datacenters in space?  It can't be to save money on power or cooling as those costs would increase exponentially.

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.
ISS as a reference is outdated tech wise. The move to self processing sats per Mars sat plans and TDRS-4G is a step in that direction.

Offline Eerie

  • Member
  • Full Member
  • ****
  • Posts: 858
  • Liked: 209
  • Likes Given: 25
Re: Amazon Project Kuiper Broadband Constellation
« Reply #104 on: 07/31/2019 08:50 am »
I would guess a single data center weights more than all things humanity launched into space so far.

Offline Cheapchips

  • Full Member
  • ****
  • Posts: 1044
  • UK
  • Liked: 902
  • Likes Given: 1973
Re: Amazon Project Kuiper Broadband Constellation
« Reply #105 on: 07/31/2019 09:41 am »
I would guess a single data center weights more than all things humanity launched into space so far.
And the maintenance staff would be the most ambitious human space flight program so far. :)

Online meekGee

  • Senior Member
  • *****
  • Posts: 14667
  • N. California
  • Liked: 14670
  • Likes Given: 1420
Re: Amazon Project Kuiper Broadband Constellation
« Reply #106 on: 07/31/2019 04:19 pm »
Found few articles about power usage of AWS data centres. We are talks 100MWs for some centres. Once their satellite constellation is in place I can see AWS starting to placing data centres in space. They could well end up being biggest user of space stations.
I don't know specifically about AWS data centres, Azure ones will have at least 10's of staff on site daily to fix hardware and software.  Blue would need to get their manned flights up and running for orbital data centres have parity with an Earthbound one.

(Unless AWS centres are much better than Azure ones)

I don't understand the justification for "datacenters in space".  High energy usage demands huge radiators to handle the thermal load.  Look at what the ISS needs for their very modest power budget.

Just the launch cost would dwarf the value of servers, storage and network devices that would populate your space-based datacenter, so what's the actual value of putting the datacenters in space?  It can't be to save money on power or cooling as those costs would increase exponentially.

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.
Also, data centers are locale-optimized.  If orbital, you'd need to replicate a lot of data, and to what end?
ABCD - Always Be Counting Down

Offline niwax

  • Full Member
  • ****
  • Posts: 1428
  • Germany
    • SpaceX Booster List
  • Liked: 2045
  • Likes Given: 166
Re: Amazon Project Kuiper Broadband Constellation
« Reply #107 on: 07/31/2019 05:58 pm »
Found few articles about power usage of AWS data centres. We are talks 100MWs for some centres. Once their satellite constellation is in place I can see AWS starting to placing data centres in space. They could well end up being biggest user of space stations.
I don't know specifically about AWS data centres, Azure ones will have at least 10's of staff on site daily to fix hardware and software.  Blue would need to get their manned flights up and running for orbital data centres have parity with an Earthbound one.

(Unless AWS centres are much better than Azure ones)

I don't understand the justification for "datacenters in space".  High energy usage demands huge radiators to handle the thermal load.  Look at what the ISS needs for their very modest power budget.

Just the launch cost would dwarf the value of servers, storage and network devices that would populate your space-based datacenter, so what's the actual value of putting the datacenters in space?  It can't be to save money on power or cooling as those costs would increase exponentially.

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.
Also, data centers are locale-optimized.  If orbital, you'd need to replicate a lot of data, and to what end?

The bigger thing would be to offer high-bandwidth satellite uplink to cloud companies. Suddenly, their edge datacenter covers a 500km radius in one hop on the backbone.
Which booster has the most soot? SpaceX booster launch history! (discussion)

Offline octavo

  • Full Member
  • **
  • Posts: 247
  • Liked: 186
  • Likes Given: 740
Re: Amazon Project Kuiper Broadband Constellation
« Reply #108 on: 08/11/2019 08:50 pm »
I would guess a single data center weights more than all things humanity launched into space so far.
And the maintenance staff would be the most ambitious human space flight program so far. :)
Microsoft have been experimenting with submarine data centres. If you don't need oxygen in the air and you keep them cool, electronics are far less failure prone.

https://www.bbc.com/news/technology-44368813

Offline Tywin

Re: Amazon Project Kuiper Broadband Constellation
« Reply #109 on: 10/02/2019 02:01 pm »
A little more info about this project:

Quote
“Amazon is designing syb-systems to withstand impact from non-trackable debris without major fragmentation risk,” Keisner said. “For example, Amazon intends to use a propulsion system with a chemically inert propellant and is exploring innovative propellant solutions that can be unpressurized.”


Keisner said Project Kuiper’s engineers were still weighing exactly what kind of propulsion system to use. The options range from “conventional designs using pressurized noble gas propellant to designs using liquid ionic propellant that can be non-pressurized,” he wrote.


https://www.geekwire.com/2019/amazons-project-kuiper-oneweb-raise-curtain-higher-satellite-plans/
The knowledge is power...Everything is connected...
The Turtle continues at a steady pace ...

Offline Rondaz

  • Senior Member
  • *****
  • Posts: 27059
  • Liked: 5301
  • Likes Given: 169
Re: Amazon Project Kuiper Broadband Constellation
« Reply #110 on: 10/02/2019 02:08 pm »
In a job posting for #ProjectKuiper from 20 Sep @amazon seek an "Electrical Engineer – Satellite Payload Systems" responsible for "low cost satellite payload system including integration of phased array mmWave RF systems". Do they plan use of V band, too?

https://twitter.com/Megaconstellati/status/1179393902773706752

Offline russianhalo117

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 8818
  • Liked: 4748
  • Likes Given: 768
Re: Amazon Project Kuiper Broadband Constellation
« Reply #111 on: 10/02/2019 02:46 pm »
In a job posting for #ProjectKuiper from 20 Sep @amazon seek an "Electrical Engineer – Satellite Payload Systems" responsible for "low cost satellite payload system including integration of phased array mmWave RF systems". Do they plan use of V band, too?

https://twitter.com/Megaconstellati/status/1179393902773706752
They have said in the past that Amazon plans to serve many commercial, civil and government allocated bands. They have not stated if Kuiper will serve all of those bands or with other sats. With SDR based phased arrays and enough power and diameter they can serve many bands via pushing an update or provide listening detection and relay on an as needed basis while prohiding it's primary services.

Online sanman

  • Senior Member
  • *****
  • Posts: 6085
  • Liked: 1365
  • Likes Given: 8
Re: Amazon Project Kuiper Broadband Constellation
« Reply #112 on: 10/02/2019 06:05 pm »
So is this the Amazon Starlink?

(makes sense, since Amazon already sells underpants)

Offline ncb1397

  • Senior Member
  • *****
  • Posts: 3497
  • Liked: 2310
  • Likes Given: 29
Re: Amazon Project Kuiper Broadband Constellation
« Reply #113 on: 10/02/2019 06:16 pm »
Found few articles about power usage of AWS data centres. We are talks 100MWs for some centres. Once their satellite constellation is in place I can see AWS starting to placing data centres in space. They could well end up being biggest user of space stations.
I don't know specifically about AWS data centres, Azure ones will have at least 10's of staff on site daily to fix hardware and software.  Blue would need to get their manned flights up and running for orbital data centres have parity with an Earthbound one.

(Unless AWS centres are much better than Azure ones)

I don't understand the justification for "datacenters in space".  High energy usage demands huge radiators to handle the thermal load.  Look at what the ISS needs for their very modest power budget.

Just the launch cost would dwarf the value of servers, storage and network devices that would populate your space-based datacenter, so what's the actual value of putting the datacenters in space?  It can't be to save money on power or cooling as those costs would increase exponentially.

It is entirely possible that some computing systems would work better in microgravity. For instance, Bose Einstein condensate has been evaluated for quantum computers, but it is more stable in microgravity.

see:
Quote
Quantum computers promise to perform certain types of operations much more quickly than conventional digital computers. But many challenges must be addressed before these ultra-fast machines become available, among them, the loss of order in the systems – a problem known as quantum decoherence – which worsens as the number of bits in a quantum computer increases.

One proposed solution is to divide the computing among multiple small quantum computers that would work together much as today's multi-core supercomputers team up to tackle big digital operations. The individual computers in such a system could communicate quantum information using Bose-Einstein condensates (BECs) – clouds of ultra-cold atoms that all exist in exactly the same quantum state. The approach could address the decoherence problem by reducing the number of bits necessary for a single computer.
https://phys.org/news/2013-04-bose-einstein-condensates-quantum.html

Quote
The Cold Atom Laboratory (CAL) is a fundament physics user facility that will operate on the International Space Station (ISS). CAL will produce clouds of ultra-cooled atoms called Bose-Einstein condensates. Chilled to a fraction of a degree above absolute zero -- even colder than the average temperature of deep space -- the atoms in a BEC demonstrate quantum characteristics at relatively large size scales, allowing researchers to explore this strange domain.

On Earth, freely evolving BEC’s are dragged down by the pull of gravity, and can typically only be observed for a fraction of a second. But in the microgravity environment of the space station, each freely evolving BEC can be observed for up to 10 seconds, which is longer than what’s possible with any other existing BEC experiment. CAL is a multi-user facility and researchers will be able to conduct experiments remotely, with no astronaut assistance, with up to 6.5 hours of experimentation time available each day.
https://coldatomlab.jpl.nasa.gov/

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.

You basically just described a data center.
« Last Edit: 10/02/2019 06:21 pm by ncb1397 »

Online sanman

  • Senior Member
  • *****
  • Posts: 6085
  • Liked: 1365
  • Likes Given: 8
Re: Amazon Project Kuiper Broadband Constellation
« Reply #114 on: 10/02/2019 08:11 pm »
It is entirely possible that some computing systems would work better in microgravity. For instance, Bose Einstein condensate has been evaluated for quantum computers, but it is more stable in microgravity.

But gravity is a weaker longer distance force, whereas quantum decoherence seems to mainly be caused by stronger local influences, like electromagnetic forces. Plus individual particles have lower mass by definition, in contrast to large astrophysical objects which are dominated by gravitational forces due to their high mass.

It does seem like atomic interferometry could be very useful in space for gravity measurements and mass detection. Imagine being able to find some elusive Planet-9 hiding in the Kuiper belt by measuring anomalous changes in gravity.

But it doesn't seem like their choice of naming it "Project Kuiper" really has anything to do with the Kuiper Belt. It just seems to be a snazzy name.

Offline lonestriker

  • Full Member
  • ****
  • Posts: 417
  • Houston We've Had A Problem
  • Liked: 820
  • Likes Given: 5155
Re: Amazon Project Kuiper Broadband Constellation
« Reply #115 on: 10/02/2019 10:11 pm »

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.

You basically just described a data center.

I work with our own data centers across the globe daily, so this is actually in my wheelhouse.  What I've described is more like a stripped down CDN and not a data center.  Data centers provide compute and storage among other services.  If Starlink provides basic CDN features to cache content closer to customer downlinks, it would be a huge advantage and prevent a lot of duplicate data from having to traverse their constellation as well as very extremely low latency.  Whether they'll get into that game or not and solve all the technical issues with delivering content directly from their birds and all the associated hand-off issues is unknown.  I'd put the chance of that as pretty unlikely for the first few iterations of Starlink as it really is an optimization they can tackle after they've got the basics and ISLs working.

Offline ncb1397

  • Senior Member
  • *****
  • Posts: 3497
  • Liked: 2310
  • Likes Given: 29
Re: Amazon Project Kuiper Broadband Constellation
« Reply #116 on: 10/02/2019 10:28 pm »

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.

You basically just described a data center.

I work with our own data centers across the globe daily, so this is actually in my wheelhouse.  What I've described is more like a stripped down CDN and not a data center.  Data centers provide compute and storage among other services.  If Starlink provides basic CDN features to cache content closer to customer downlinks, it would be a huge advantage and prevent a lot of duplicate data from having to traverse their constellation as well as very extremely low latency.  Whether they'll get into that game or not and solve all the technical issues with delivering content directly from their birds and all the associated hand-off issues is unknown.  I'd put the chance of that as pretty unlikely for the first few iterations of Starlink as it really is an optimization they can tackle after they've got the basics and ISLs working.

Anyways, data repository in space has already been done. Do you think that when you order video on demand from Dish or DirectTV, they upload the file to the satellite, let you watch it, and then delete it only for the next user to request it shortly thereafter? No, that would be silly.

Offline lonestriker

  • Full Member
  • ****
  • Posts: 417
  • Houston We've Had A Problem
  • Liked: 820
  • Likes Given: 5155
Re: Amazon Project Kuiper Broadband Constellation
« Reply #117 on: 10/02/2019 10:45 pm »

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.

You basically just described a data center.

I work with our own data centers across the globe daily, so this is actually in my wheelhouse.  What I've described is more like a stripped down CDN and not a data center.  Data centers provide compute and storage among other services.  If Starlink provides basic CDN features to cache content closer to customer downlinks, it would be a huge advantage and prevent a lot of duplicate data from having to traverse their constellation as well as very extremely low latency.  Whether they'll get into that game or not and solve all the technical issues with delivering content directly from their birds and all the associated hand-off issues is unknown.  I'd put the chance of that as pretty unlikely for the first few iterations of Starlink as it really is an optimization they can tackle after they've got the basics and ISLs working.

Anyways, data repository in space has already been done. Do you think that when you order video on demand from Dish or DirectTV, they upload the file to the satellite, let you watch it, and then delete it only for the next user to request it shortly thereafter? No, that would be silly.

Satellite TV is a bad analogy here.  When you order a video-on-demand from Dish or DirectTV, they send you the video by bouncing it off their satellites AFAIK from their data centers on the ground, no storage done at the satellite.  They have channels with movies running all the time every N minutes that you can order and watch, but those are just like other channels and it's just bouncing a stream again and authorizing you to watch that channel.  They used to do some creative stuff with their home DVRs to cache content there and unlock it when you've paid.

If a content provider uses a CDN like Cloudflare, when a user requests a web page, it sends the user cached copies of some of the content (the non-dynamic bits.)  Netflix runs their own CDN to cache content closer to the user as well.  That's what I'm describing as moving to the satellite.  I'm not aware of any satellite service that caches content like the CDNs do nor any data repositories in space.  If you know of any, I'd be interested to get details.

Offline ncb1397

  • Senior Member
  • *****
  • Posts: 3497
  • Liked: 2310
  • Likes Given: 29
Re: Amazon Project Kuiper Broadband Constellation
« Reply #118 on: 10/02/2019 11:37 pm »

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.

You basically just described a data center.

I work with our own data centers across the globe daily, so this is actually in my wheelhouse.  What I've described is more like a stripped down CDN and not a data center.  Data centers provide compute and storage among other services.  If Starlink provides basic CDN features to cache content closer to customer downlinks, it would be a huge advantage and prevent a lot of duplicate data from having to traverse their constellation as well as very extremely low latency.  Whether they'll get into that game or not and solve all the technical issues with delivering content directly from their birds and all the associated hand-off issues is unknown.  I'd put the chance of that as pretty unlikely for the first few iterations of Starlink as it really is an optimization they can tackle after they've got the basics and ISLs working.

Anyways, data repository in space has already been done. Do you think that when you order video on demand from Dish or DirectTV, they upload the file to the satellite, let you watch it, and then delete it only for the next user to request it shortly thereafter? No, that would be silly.

Satellite TV is a bad analogy here.  When you order a video-on-demand from Dish or DirectTV, they send you the video by bouncing it off their satellites AFAIK from their data centers on the ground, no storage done at the satellite.  They have channels with movies running all the time every N minutes that you can order and watch, but those are just like other channels and it's just bouncing a stream again and authorizing you to watch that channel.  They used to do some creative stuff with their home DVRs to cache content there and unlock it when you've paid.

After some more research, it appears that dish requires a seperate internet connection for video on demand. My limited experience with it was through a hughes satellite internet and dish package assuming that the video was coming from the dish satellite but was apparently coming through the internet satellite operated by Hughes connecting to Dish's ground based servers. So, yeah, there wouldn't be any caching of data on the Dish or Hughes satellite, but it begs that question if it would be more efficient to do so and not have the same information transferred up to the satellite over and over again. We will see where Amazon goes with this, whether they bundle TV or Amazon Prime video with it and if they do any on satellite caching and if so, whether they open it up to other providers like Netflix through something like AWS.
« Last Edit: 10/02/2019 11:42 pm by ncb1397 »

Online meekGee

  • Senior Member
  • *****
  • Posts: 14667
  • N. California
  • Liked: 14670
  • Likes Given: 1420
Re: Amazon Project Kuiper Broadband Constellation
« Reply #119 on: 10/03/2019 12:36 am »

There's been discussions of putting some sort of cache for content in the constellation satellites themselves to move things like movies and other streaming content closer to the users.  I could see how something like having a bunch of SSDs to cache content in satellites could be useful.  But not much beyond that.

You basically just described a data center.

I work with our own data centers across the globe daily, so this is actually in my wheelhouse.  What I've described is more like a stripped down CDN and not a data center.  Data centers provide compute and storage among other services.  If Starlink provides basic CDN features to cache content closer to customer downlinks, it would be a huge advantage and prevent a lot of duplicate data from having to traverse their constellation as well as very extremely low latency.  Whether they'll get into that game or not and solve all the technical issues with delivering content directly from their birds and all the associated hand-off issues is unknown.  I'd put the chance of that as pretty unlikely for the first few iterations of Starlink as it really is an optimization they can tackle after they've got the basics and ISLs working.

Anyways, data repository in space has already been done. Do you think that when you order video on demand from Dish or DirectTV, they upload the file to the satellite, let you watch it, and then delete it only for the next user to request it shortly thereafter? No, that would be silly.
When you say "has already been done", is this based on anything you actually know?

Because what you're describing doesn't sound to me like something that actually exists or is practical.
ABCD - Always Be Counting Down

Tags: kuiper 
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
0