Author Topic: Starlink : Markets and Marketing  (Read 314167 times)

Offline markbike528cbx

  • Full Member
  • *
  • Posts: 162
  • The Everbrown portion of the Evergreen State
  • Liked: 143
  • Likes Given: 88
Re: Starlink : Markets and Marketing
« Reply #680 on: 08/15/2022 08:34 pm »
Am I missing a thread? Or is this the correct/not correct place to discuss the Rural Digital Opportunity Fund  (RDOF) subsidy denial for Starlink?

https://www.businessinsider.com/elon-musk-starlink-subsidy-bid-rejected-fcc-2022-8

I would have thought that this community would be all over it. 800 million is a big chunk of change, even for Elon.

My thought is that perhaps Starlink is better off without the strings attached.   Now they can pick and choose customer locations, not being tied down to specific regions.

On the other hand, I think the decision smacks of inappropriate politics.  Just like the case of Tesla not being invited to an EV conference.
A FCC commissioner agrees:
"Carr went on to add: "I will have more to say because we should be making it easier for unserved communities to get service, not rejecting a proven satellite technology that is delivering robust, high-speed service today. "

https://www.lightreading.com/regulatorypolitics/fcc-commissioner-surprised-to-learn-of-starlink-rdof-rejection/d/d-id/779645

Offline AnalogMan

  • Member
  • Senior Member
  • *****
  • Posts: 3436
  • Cambridge, UK
  • Liked: 1612
  • Likes Given: 51
Re: Starlink : Markets and Marketing
« Reply #681 on: 08/15/2022 10:14 pm »
Am I missing a thread? Or is this the correct/not correct place to discuss the Rural Digital Opportunity Fund  (RDOF) subsidy denial for Starlink?

Discussion of recent RDOF news can be found in this thread:

Starlink : General Discussion - Thread 3
https://forum.nasaspaceflight.com/index.php?topic=55795.0

Offline su27k

  • Senior Member
  • *****
  • Posts: 6414
  • Liked: 9100
  • Likes Given: 885
Re: Starlink : Markets and Marketing
« Reply #682 on: 08/23/2022 05:35 am »
DARPA Kicks Off Program to Develop Low-Earth Orbit Satellite ‘Translator’

Teams begin work on novel optical intersatellite communications terminal to connect satellite networks

[email protected]

8/10/2022

DARPA has selected 11 teams for Phase 1 of the Space-Based Adaptive Communications Node program, known as Space-BACN. Space-BACN aims to create a low-cost, reconfigurable optical communications terminal that adapts to most optical intersatellite link standards, translating between diverse satellite constellations. Space-BACN would create an “internet” of low-Earth orbit (LEO) satellites, enabling seamless communication between military/government and commercial/civil satellite constellations that currently are unable to talk with each other.

The agency selected teams from academia and large and small commercial companies, including multiple performers awarded first-time contracts with the Department of Defense.

“We intentionally made making a proposal to our Space-BACN solicitations as easy as possible, because we wanted to tap into both established defense companies and the large pool of innovative small tech companies, many of which don’t have the time or resources to figure out complicated government contracting processes,” said Greg Kuperman, Space-BACN program manager in DARPA’s Strategic Technology Office. “We used other transactions and were very pleased with diversity of organizations that responded and quality of proposals. After a successful Phase 0 where we got to see the teams sprint to put together an initial architecture design for Space-BACN, I'm excited to get to work in Phase 1 building the actual system.”

In the first technical area, the following performers aim to develop a flexible, low size, weight, power and cost (SWaP-C) optical aperture that couples into single-mode fiber:
* CACI, Inc.
* MBRYONICS
* Mynaric

The following teams selected in the second technical area aim to develop a reconfigurable optical modem that supports up to 100 Gbps on a single wavelength:
* II-VI Aerospace and Defense
* Arizona State University
* Intel Federal, LLC

The performer teams listed above will also participate in a collaborative working group to define the interface between their respective system components.

In a third technical area, the agency selected five teams to identify critical command and control elements required to support cross-constellation optical intersatellite link communications and develop the schema necessary to interface between Space-BACN and commercial partner constellations:
* Space Exploration Technologies (SpaceX)
* Telesat
* SpaceLink
* Viasat
* Kuiper Government Solutions (KGS) LLC, an Amazon subsidiary

Phase 1 of Space- BACN spans 14 months and will conclude with a preliminary design review for the first two technical areas, as well as a fully defined interface between system components. The third technical area will develop the schema for cross-constellation command and control, and will conduct a connectivity demo in a simulated environment to test the schema for a baseline scenario.

At the completion of Phase 1, selected performers in the first two technical areas will participate in an 18-month Phase 2 to develop engineering design units of the optical terminal components, while performers in the third technical area will continue to evolve the schema to function in more challenging and dynamic scenarios.


Offline Asteroza

  • Senior Member
  • *****
  • Posts: 2853
  • Liked: 1092
  • Likes Given: 33
Re: Starlink : Markets and Marketing
« Reply #683 on: 08/23/2022 07:12 am »
DARPA Kicks Off Program to Develop Low-Earth Orbit Satellite ‘Translator’

Teams begin work on novel optical intersatellite communications terminal to connect satellite networks

[email protected]

8/10/2022

DARPA has selected 11 teams for Phase 1 of the Space-Based Adaptive Communications Node program, known as Space-BACN. Space-BACN aims to create a low-cost, reconfigurable optical communications terminal that adapts to most optical intersatellite link standards, translating between diverse satellite constellations. Space-BACN would create an “internet” of low-Earth orbit (LEO) satellites, enabling seamless communication between military/government and commercial/civil satellite constellations that currently are unable to talk with each other.

The agency selected teams from academia and large and small commercial companies, including multiple performers awarded first-time contracts with the Department of Defense.

“We intentionally made making a proposal to our Space-BACN solicitations as easy as possible, because we wanted to tap into both established defense companies and the large pool of innovative small tech companies, many of which don’t have the time or resources to figure out complicated government contracting processes,” said Greg Kuperman, Space-BACN program manager in DARPA’s Strategic Technology Office. “We used other transactions and were very pleased with diversity of organizations that responded and quality of proposals. After a successful Phase 0 where we got to see the teams sprint to put together an initial architecture design for Space-BACN, I'm excited to get to work in Phase 1 building the actual system.”

In the first technical area, the following performers aim to develop a flexible, low size, weight, power and cost (SWaP-C) optical aperture that couples into single-mode fiber:
* CACI, Inc.
* MBRYONICS
* Mynaric

The following teams selected in the second technical area aim to develop a reconfigurable optical modem that supports up to 100 Gbps on a single wavelength:
* II-VI Aerospace and Defense
* Arizona State University
* Intel Federal, LLC

The performer teams listed above will also participate in a collaborative working group to define the interface between their respective system components.

In a third technical area, the agency selected five teams to identify critical command and control elements required to support cross-constellation optical intersatellite link communications and develop the schema necessary to interface between Space-BACN and commercial partner constellations:
* Space Exploration Technologies (SpaceX)
* Telesat
* SpaceLink
* Viasat
* Kuiper Government Solutions (KGS) LLC, an Amazon subsidiary

Phase 1 of Space- BACN spans 14 months and will conclude with a preliminary design review for the first two technical areas, as well as a fully defined interface between system components. The third technical area will develop the schema for cross-constellation command and control, and will conduct a connectivity demo in a simulated environment to test the schema for a baseline scenario.

At the completion of Phase 1, selected performers in the first two technical areas will participate in an 18-month Phase 2 to develop engineering design units of the optical terminal components, while performers in the third technical area will continue to evolve the schema to function in more challenging and dynamic scenarios.

For a bit of nostalgia, try visualizing a dog yelling out "It's (Space-)BACN!!!"

As the pork must flow, so must the funds for increasing lasercomm work. That said, SpaceLink appear to be the only pure optical constellation, and ostensibly well positioned to be the mediator/gateway between multiple constellations for orbital links due to them being a MEO-LEO optical relay service by design due to their initial 4 sat MEO constellation configuration.

The probably additional usage as backhaul for BACN airborne tactical communication node pods (AKA tactical flying cell towers), also closely resembles the interplay between commercial HAPS cell tower relays with satellite backhaul. Think military private 5G networks as an analogue.

Which reminds me, the russians apparently had an encrypted tactical radio that was supposed to be used in ukraine, that hitches a ride on public cellular networks. Unfortunately due to the left hand not knowing what the right hand is doing, the russians also deliberately destroyed the cell towers those radios were supposed to piggyback on...


Offline oldAtlas_Eguy

  • Senior Member
  • *****
  • Posts: 5305
  • Florida
  • Liked: 5005
  • Likes Given: 1459
Re: Starlink : Markets and Marketing
« Reply #684 on: 08/23/2022 05:55 pm »
Think about the feasibility of an in-space common Internet backbone backhaul that different constellations can purchase connections and data transport through. Possibly making a large LEO interconnected Internet backbone that covers the globe. This could be in the long run a big positive due to not having a lot of data traffic to not get transported back and forth from space and Earth surface by going through an internet switching node on-orbit between in-orbit ISPs. By eliminating a single up-down bounce to the point to point data traffic you can greatly increase the effectiveness of the offered up-down troughput.

Online DanClemmensen

  • Senior Member
  • *****
  • Posts: 5583
  • Earth (currently)
  • Liked: 4390
  • Likes Given: 1791
Re: Starlink : Markets and Marketing
« Reply #685 on: 08/23/2022 06:17 pm »
Think about the feasibility of an in-space common Internet backbone backhaul that different constellations can purchase connections and data transport through. Possibly making a large LEO interconnected Internet backbone that covers the globe. This could be in the long run a big positive due to not having a lot of data traffic to not get transported back and forth from space and Earth surface by going through an internet switching node on-orbit between in-orbit ISPs. By eliminating a single up-down bounce to the point to point data traffic you can greatly increase the effectiveness of the offered up-down troughput.
You just need interconnects, not a separate core backbone net. Easier and more efficient. The sending net would forward to the receiving net at the earliest (i.e., shortest-path) opportunity because the sending net does not know the geolocation of the receiving user. The receiving net then does optimal routing based on its algorithms. Otherwise, you would need to exchange an unacceptably large amount of info about the current dynamic topology of each network. Back in the day, we used to do this in the core internet in locations known as IXPs, and I suspect we still do, although Peer-to-peer is now more prevalent. It seems to be selfish and greedy (get this traffic off of my net as quickly and cheaply as possible) but it is in fact more efficient. To implement this maybe use a small constellation of IXP satellites that each are members of all participating networks. These satellites don't in general need to talk directly to each other.

Offline oldAtlas_Eguy

  • Senior Member
  • *****
  • Posts: 5305
  • Florida
  • Liked: 5005
  • Likes Given: 1459
Re: Starlink : Markets and Marketing
« Reply #686 on: 08/23/2022 08:37 pm »
Think about the feasibility of an in-space common Internet backbone backhaul that different constellations can purchase connections and data transport through. Possibly making a large LEO interconnected Internet backbone that covers the globe. This could be in the long run a big positive due to not having a lot of data traffic to not get transported back and forth from space and Earth surface by going through an internet switching node on-orbit between in-orbit ISPs. By eliminating a single up-down bounce to the point to point data traffic you can greatly increase the effectiveness of the offered up-down troughput.
You just need interconnects, not a separate core backbone net. Easier and more efficient. The sending net would forward to the receiving net at the earliest (i.e., shortest-path) opportunity because the sending net does not know the geolocation of the receiving user. The receiving net then does optimal routing based on its algorithms. Otherwise, you would need to exchange an unacceptably large amount of info about the current dynamic topology of each network. Back in the day, we used to do this in the core internet in locations known as IXPs, and I suspect we still do, although Peer-to-peer is now more prevalent. It seems to be selfish and greedy (get this traffic off of my net as quickly and cheaply as possible) but it is in fact more efficient. To implement this maybe use a small constellation of IXP satellites that each are members of all participating networks. These satellites don't in general need to talk directly to each other.
The third party IXP sat Internet nodes would likely have a hosted Gateway of the attached network ISP constellation or the links in the constellation wold also implement the Gateway function as if the link it is transmitting through is a Internet backbone connection. There would be some level of negotiation in the routing data in the Internet node sat that can be read by a constellation to determine which node it needs to send the packets through in order to have reasonably short routes. As occurs even in terrestrial routing, short routes are not guaranteed. Even though the hardware links don't change like the constant change of the links occurring between two in-space connected constellations.

Routing would be significantly more dynamic than what it is now with fixed location Gateways. NOTED is that Starlink routing usage of a Gateway for a specific User connection changes even now. A specific Gateway usage change would just occur a little more often in the remain in-orbit situation routing. But something would have to accomplish the Gateway functions for each of the constellations even if there is a direct link between 2 constellations. The Gateways negotiate the routing between the routes used internal to the constellation and the routing used external to the constellation.

The idea that terrestrial Internet routing is a fixed route is a falsehood. The route can change multiple times during the interchange between the connected actors. This is one of the factors in why the round trip can time can change during the connection. It's dynamic. Just not as dynamic as what is being proposed for an in-orbit Internet Node functionality.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 4489
  • Tampa, FL
  • Liked: 3353
  • Likes Given: 641
Re: Starlink : Markets and Marketing
« Reply #687 on: 08/23/2022 09:08 pm »
The idea that terrestrial Internet routing is a fixed route is a falsehood. The route can change multiple times during the interchange between the connected actors. This is one of the factors in why the round trip can time can change during the connection. It's dynamic. Just not as dynamic as what is being proposed for an in-orbit Internet Node functionality.

Routes do indeed change, but excessive route "flaps" are a fairly big pain, because they generate a fair number of out-of-order packets, which TCP treats as congestion.  As a result, backbone ISPs try to keep their routes as stable as possible.

Route stability also keeps peering relationships simple.  Many ISPs like to use settlement-free peering, and a flap that forces them into a network requiring settlement (i.e., one where the sending party pays the receiving party to carry traffic) makes the business case complicated.

I don't see much use for a generic on-orbit backbone.  In almost all cases, terrestrial backbones have orders of magnitude more bandwidth, and the topology is soooooo much simpler.  However, I can think of a few special-purpose cases where a small backbone might make sense:

1) The SpaceX-touted case of a backbone with specific quality-of-service characteristics.  The high-runner here is obviously extremely low latency.  However, orbital backbone bandwidth will be limited, which will make it very expensive.  I really can't think of many applications beyond financial transactions and mission-critical videoconferencing.  I doubt that most gamers will be willing to pay the equilibrium price that would make it worthwhile.

2) Aggregating traffic from satellites to piggyback on the backbone's suite of ground stations.  If SpaceX makes the ISL hardware and protocols available to third parties, then it enables a huge number of applications that aren't worth dealing with the FCC for spectrum, but which can work quite well with modest bandwidth supported by SpaceX to get their data down and up over the Starlink-licensed spectrum, using Starlink gateways.

3) I think the idea of putting content distribution caches inside Starlinks is full-blown bat-guano crazy, but there are cases where over-the-top content distro networks need to reach back to content root providers to get rare content.  If the CDNs can satisfy those rare requests via low-latency backbones, that might give them a competitive advantage that would make it worthwhile.

I still expect that 99.9+% of SpaceX's traffic will get hot-potatoed to the nearest gateway and proceed via terrestrial backbone from there on. There will obviously be lots of cases where that nearest gateway will be several hops away (transcontinental air, maritime, and land areas that are so desolate that they don't warrant a gateway and a terrestrial link to service it).  But end-to-end traffic that never touches the ground will be exceedingly rare.

Online DanClemmensen

  • Senior Member
  • *****
  • Posts: 5583
  • Earth (currently)
  • Liked: 4390
  • Likes Given: 1791
Re: Starlink : Markets and Marketing
« Reply #688 on: 08/23/2022 10:24 pm »

I don't see much use for a generic on-orbit backbone.  In almost all cases, terrestrial backbones have orders of magnitude more bandwidth, and the topology is soooooo much simpler.  However, I can think of a few special-purpose cases where a small backbone might make sense:

Optical ISL has the potential to change this completely. (Whether or not is actually happens, and when, is another story.) If access to space is cheap you can use large satellites with lots of power for lasers and optical amplifiers. A single ISL in free space can carry a very large number of lambdas, not just the optical C-band and S-band of a fiber link. Speed of light in free space is 300,000 km/s, while speed of light in fiber is about 200,000 km/s. More importantly, ISL links are straight lines, not following fiber rights-of-way. It will ultimately be cheaper to use ISL than long-haul fiber, and this will be true from the beginning for any data that can avoid an extra uplink and downlink. You don't necessarily need a separate shared ISL backbone. You do need to interface each constellation's backbones to each other.

Offline Mandella

  • Full Member
  • ****
  • Posts: 524
  • Liked: 799
  • Likes Given: 2605
Re: Starlink : Markets and Marketing
« Reply #689 on: 08/23/2022 10:38 pm »
So, ah, just got this:

Quote
Your service area remains at capacity which means we are unable to convert you to Residential service at this time. Our best estimate for delivery of Residential service in your area is now mid-2023.   
 
Adding users before adding additional network capacity would result in delivery of a lesser service. However, we understand that some customers may have urgent connectivity needs with no other alternatives.   
 
For those with urgent connectivity needs, we are offering a Best Effort service option. 

What is Best Effort?
Best Effort enables typical internet activity with the understanding that Best Effort users will be deprioritized behind Residential users. 
 
The monthly service charge is the same but unlike Residential, Best Effort users will have the option to pause service on their account page. 

What will the internet experience be like with Best Effort?
During times of peak network congestion, Best Effort users will experience notably slower speeds than Residential users.  Best Effort users will be able to engage in typical internet activity like email, online shopping, or streaming an SD movie, but they won’t be able to engage in activities like online gaming, video calls or streaming 4K and HD movies.
 
Outside times of peak network congestion, Best Effort users can expect to engage in all of the above activities and more, but speeds will be slower than Residential users.   

If I choose Best Effort now, do I lose my place in line for Residential service?
No, if you choose Best Effort you will not lose your spot in line. When capacity increases in your area, your service will be upgraded automatically at no additional cost. 

You can opt-in to Best Effort from your account page at any time or by clicking the button below. 
OPT-IN TO BEST EFFORT
Best Effort is intended for pre-order customers with urgent connectivity needs.  If you can wait for Residential, we strongly encourage you to do so.   
 
In just 18 months, our teams have already enabled over 600,000 users across 40 markets.  By mid-2023, our rapid launch rate and higher performing satellites, coupled with the continuous deployment of software features to optimize our network, will enable a high quality connection for millions more across the globe.
 
As always, you cancel your pre-order at any time on your account page for a full refund of your deposit.  We appreciate your continued patience and we look forward to getting you online!
 
The Starlink Team 


Must be nice to be so in demand that you can beg your customers not to buy...

Offline M.E.T.

  • Senior Member
  • *****
  • Posts: 2318
  • Liked: 2921
  • Likes Given: 508
Re: Starlink : Markets and Marketing
« Reply #690 on: 08/24/2022 01:36 am »
So, ah, just got this:

Quote
Your service area remains at capacity which means we are unable to convert you to Residential service at this time. Our best estimate for delivery of Residential service in your area is now mid-2023.   
 
Adding users before adding additional network capacity would result in delivery of a lesser service. However, we understand that some customers may have urgent connectivity needs with no other alternatives.   
 
For those with urgent connectivity needs, we are offering a Best Effort service option. 

What is Best Effort?
Best Effort enables typical internet activity with the understanding that Best Effort users will be deprioritized behind Residential users. 
 
The monthly service charge is the same but unlike Residential, Best Effort users will have the option to pause service on their account page. 

What will the internet experience be like with Best Effort?
During times of peak network congestion, Best Effort users will experience notably slower speeds than Residential users.  Best Effort users will be able to engage in typical internet activity like email, online shopping, or streaming an SD movie, but they won’t be able to engage in activities like online gaming, video calls or streaming 4K and HD movies.
 
Outside times of peak network congestion, Best Effort users can expect to engage in all of the above activities and more, but speeds will be slower than Residential users.   

If I choose Best Effort now, do I lose my place in line for Residential service?
No, if you choose Best Effort you will not lose your spot in line. When capacity increases in your area, your service will be upgraded automatically at no additional cost. 

You can opt-in to Best Effort from your account page at any time or by clicking the button below. 
OPT-IN TO BEST EFFORT
Best Effort is intended for pre-order customers with urgent connectivity needs.  If you can wait for Residential, we strongly encourage you to do so.   
 
In just 18 months, our teams have already enabled over 600,000 users across 40 markets.  By mid-2023, our rapid launch rate and higher performing satellites, coupled with the continuous deployment of software features to optimize our network, will enable a high quality connection for millions more across the globe.
 
As always, you cancel your pre-order at any time on your account page for a full refund of your deposit.  We appreciate your continued patience and we look forward to getting you online!
 
The Starlink Team 


Must be nice to be so in demand that you can beg your customers not to buy...

So, they’ve passed the 600k user mark, and are aiming for at least 2.6 million by mid 2023 (“millions more” than 600k).

So about $800M annual Starlink revenue already, and aiming for more than $3B annual revenue by mid-2023.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 4489
  • Tampa, FL
  • Liked: 3353
  • Likes Given: 641
Re: Starlink : Markets and Marketing
« Reply #691 on: 08/24/2022 04:41 am »
Optical ISL has the potential to change this completely. (Whether or not is actually happens, and when, is another story.) If access to space is cheap you can use large satellites with lots of power for lasers and optical amplifiers. A single ISL in free space can carry a very large number of lambdas, not just the optical C-band and S-band of a fiber link. Speed of light in free space is 300,000 km/s, while speed of light in fiber is about 200,000 km/s. More importantly, ISL links are straight lines, not following fiber rights-of-way. It will ultimately be cheaper to use ISL than long-haul fiber, and this will be true from the beginning for any data that can avoid an extra uplink and downlink. You don't necessarily need a separate shared ISL backbone. You do need to interface each constellation's backbones to each other.

Don't confuse bandwidth and propagation speed.  They're not completely unrelated, but multimode fiber has vastly less dispersion and more chances to re-amplify the signal.

In theory, you can have a lot more links in an n-to-n mesh with satellites, but it's a horrid routing problem to keep things from flapping, and of course you can plunk down as many fibers as you like in a fiber bundle.

You might be right one day in the future, but I don't think we're talking in the Gen2 timeframe.  And I doubt that the ISL cost per bps will ever be lower than that of fiber.
« Last Edit: 08/24/2022 05:27 am by TheRadicalModerate »

Online DanClemmensen

  • Senior Member
  • *****
  • Posts: 5583
  • Earth (currently)
  • Liked: 4390
  • Likes Given: 1791
Re: Starlink : Markets and Marketing
« Reply #692 on: 08/24/2022 05:36 am »
Optical ISL has the potential to change this completely. (Whether or not is actually happens, and when, is another story.) If access to space is cheap you can use large satellites with lots of power for lasers and optical amplifiers. A single ISL in free space can carry a very large number of lambdas, not just the optical C-band and S-band of a fiber link. Speed of light in free space is 300,000 km/s, while speed of light in fiber is about 200,000 km/s. More importantly, ISL links are straight lines, not following fiber rights-of-way. It will ultimately be cheaper to use ISL than long-haul fiber, and this will be true from the beginning for any data that can avoid an extra uplink and downlink. You don't necessarily need a separate shared ISL backbone. You do need to interface each constellation's backbones to each other.

Don't confuse bandwidth and propagation speed.  They're not completely unrelated, but multimode fiber has vastly less dispersion and more chances to re-amplify the signal.

In theory, you can have a lot more links in an n-to-n mesh with satellites, but it's a horrid routing problem to keep things from flapping, and of course you can plunk down as many fibers as you like in a fiber bundle.

You might be right one day in the future, but I don't think we're talking in the Gen2 timeframe.  And I doubt that the ISL cost per bps will ever exceed that of fiber.
Sorry, no. A long-haul fiber needs optical amplifiers, usually EDFAs, at intervals of perhaps 150 km. An ISL link in vacuum using a 10 CM telescope can go at least 1000 km if you use enough power. Long-haul fiber is single mode, not multi mode. A single fiber carries multiple "lambdas", each of which carries one digital channel, using DWDM. A fiber almost always operates in only one optical band, and for long haul this is the optical C band. In 2000 a fiber carried 160 lambdas at a spacing that allowed 10 Gbit/s per lambda, for 1600 Gbit/s per fiber. The C band, centered on about 1550 nm, is in the transparency sweet spot for the long-haul fiber, where the sum of scattering loss and absorption loss is lowest. We now use higher bit rates but on more widely-spaced lambas resulting in maybe 3200 Gbit/sec per fiber. Of course, a cable has lots of separate fibers.

By contrast, we do not have absorption loss or scattering in free space, so we can use all wavelengths. Our losses occur due to dispersion, which depends on distance. We can use lasers from Moon to earth, but dispersion is so high that the signal is weak. but this effect is obviously much smaller at distances in LEO. We will eventually be able to use all optical freqencies: thousands of times more lambdas than a single fiber, but only by using sufficient power.

Routing is an entirely separate problem. I prefer not to call it routing in the ISL network because except at the IXPs the frames will be forwarded at the frame level, not the packet level, using a highly specialized frame forwarding scheme. The topology changes radically and continuously, but in a completely predictable way, so the problem is very different than either fixed networks or cell networks and any attempt to reason by analogy will fail.

Offline getitdoneinspace

  • Full Member
  • *
  • Posts: 140
  • Liked: 305
  • Likes Given: 224
Re: Starlink : Markets and Marketing
« Reply #693 on: 08/24/2022 04:34 pm »


At 7:00 p.m. CT on Thursday, August 25 at Starbase, SpaceX Chief Engineer Elon and T-Mobile CEO and President Mike Sievert will provide a public update announcing how T-Mobile and SpaceX will work together to increase connectivity.

Does anyone have any insight into this announcement?


Offline Tywin

Re: Starlink : Markets and Marketing
« Reply #694 on: 08/24/2022 05:12 pm »
Mobile connection?
The knowledge is power...Everything is connected...
The Turtle continues at a steady pace ...

Offline jimvela

  • Member
  • Full Member
  • ****
  • Posts: 1662
  • Liked: 901
  • Likes Given: 71
Re: Starlink : Markets and Marketing
« Reply #695 on: 08/24/2022 05:15 pm »
My uninformed speculation is that T-mobile will start rolling out terrestrial 5G/LTE cellular infrastructure in remote locations using Starlink backhaul instead of terrestrial backhaul. 

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 4489
  • Tampa, FL
  • Liked: 3353
  • Likes Given: 641
Re: Starlink : Markets and Marketing
« Reply #696 on: 08/24/2022 05:55 pm »
Sorry, no. A long-haul fiber needs optical amplifiers, usually EDFAs, at intervals of perhaps 150 km. An ISL link in vacuum using a 10 CM telescope can go at least 1000 km if you use enough power. Long-haul fiber is single mode, not multi mode. A single fiber carries multiple "lambdas", each of which carries one digital channel, using DWDM. A fiber almost always operates in only one optical band, and for long haul this is the optical C band. In 2000 a fiber carried 160 lambdas at a spacing that allowed 10 Gbit/s per lambda, for 1600 Gbit/s per fiber. The C band, centered on about 1550 nm, is in the transparency sweet spot for the long-haul fiber, where the sum of scattering loss and absorption loss is lowest. We now use higher bit rates but on more widely-spaced lambas resulting in maybe 3200 Gbit/sec per fiber. Of course, a cable has lots of separate fibers.

By contrast, we do not have absorption loss or scattering in free space, so we can use all wavelengths. Our losses occur due to dispersion, which depends on distance. We can use lasers from Moon to earth, but dispersion is so high that the signal is weak. but this effect is obviously much smaller at distances in LEO. We will eventually be able to use all optical freqencies: thousands of times more lambdas than a single fiber, but only by using sufficient power.

Yes, you need lots of amplifiers, but their cost to deploy isn't that big a deal, and once you have them, you have them.

OC-768 fibers have 16 frequencies, and each fiber carries about 34Gbps.  That's not thousands of frequencies, obviously.  But each ISL frequency has a separate laser associated with it, with separate optics, and the power budget associated with driving them all is non-trivial, as is the thermal management.  In contrast, terrestrial drivers have as much power as they need, and thermal management is (literally) a breeze (or a nice cold dip).

Quote
Routing is an entirely separate problem. I prefer not to call it routing in the ISL network because except at the IXPs the frames will be forwarded at the frame level, not the packet level, using a highly specialized frame forwarding scheme. The topology changes radically and continuously, but in a completely predictable way, so the problem is very different than either fixed networks or cell networks and any attempt to reason by analogy will fail.

The topology is only predictable at the geometric level--just like in a terrestrial network.  At the logical level, you still have unpredictable loads and therefore unpredictable congestion--unless you want to overdesign to the point where the route for any given flow is completely deterministic.  And if you do that, there goes any hypothetical bandwidth advantage you might have.

Again, I'm only arguing that a generic backbone is silly.  If you have a few classes of service for which customers are willing to pay through the nose, cool.  But I suspect that there will be ships laying new fiber for decades more, if not centuries.

Offline TheRadicalModerate

  • Senior Member
  • *****
  • Posts: 4489
  • Tampa, FL
  • Liked: 3353
  • Likes Given: 641
Re: Starlink : Markets and Marketing
« Reply #697 on: 08/24/2022 06:20 pm »
My uninformed speculation is that T-mobile will start rolling out terrestrial 5G/LTE cellular infrastructure in remote locations using Starlink backhaul instead of terrestrial backhaul.

Yeah, I think that backhaul will turn out to be a very nice market for SpaceX.  However, I'm betting that Verizon will be offering a private-labelled picocell solution that uses Starlink.

I looked a little bit at the economics of a pure backhaul solution and they're... not a slam-dunk.  It doesn't make much sense to roll out nano- or pico-cells in the middle of nowhere for general service, so you still need a big tower, with a lease on its footprint, and an adequate power supply.  About all you're really saving on is the microwave relays.

That's a non-trivial savings, but it's not like a cell deployment suddenly costs half of what it did before.  The relays aren't that expensive, but the towers are.  But there aren't that many places where towers are so sparse that you need to erect another one just for the relay.  Most of the time, remote areas get enabled with a single new tower that talks to a relay on an existing tower.  So satellite backhaul only makes sense if you leapfrog way out ahead of your current deployment footprint.

Two things that might be very interesting to SpaceX:

1) 5G has a much more seamless handoff between the mobile carriage and wifi, with the wifi stuff able to authenticate to use the rest of the 5G backhaul infrastructure.  That means that very small but colocated populations can be served with a picocell very efficiently, no tower involved.  More important for Verizon, they can offer a seamless, end-to-end service where no other ISP has to be involved.  That's quite a bit bigger deal than just backhaul.

2) Based on a superior picocell architecture, that lets you break into a lot of national markets where the government frowns on competitive carriers but where the cost of rolling out the infrastructure to cover every village is prohibitively high.  India would be a fine example.  So SpaceX will likely consider Verizon to be a nice proxy guinea pig for some of the more authoritarian national telcos.

Offline FutureSpaceTourist

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 48463
  • UK
    • Plan 28
  • Liked: 82199
  • Likes Given: 37085
Re: Starlink : Markets and Marketing
« Reply #698 on: 08/24/2022 07:50 pm »

Online DanClemmensen

  • Senior Member
  • *****
  • Posts: 5583
  • Earth (currently)
  • Liked: 4390
  • Likes Given: 1791
Re: Starlink : Markets and Marketing
« Reply #699 on: 08/24/2022 09:01 pm »

You can do wayyy more than 3.2Tbps now with DWDM --  modern line systems use C+L band and are capable of well over 32Tbps. The Ciena WaveServer does 12.8Tbps in 2RU. New submarine cables are doing >20Tbps per fiber pair.
Thanks. I've been out of the industry (retired) for awhile. What's the spacing between amplifiers now? For a 20 Tbps/pair cable with 24 pair, what is the expense of an amplifier? My (possibly outdated) mental model would be that the amplifier is more expensive than a satellite. Yes, it's an apples-to-bananas comparison, but it might give us a feel for the complexities.

Are we talking about 100 Ghz/lambda? if so, it looks like you are getting more than one bit/hz. Is that being done with higher-order modulations or with multiple polarizations or both? or something else?

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
1