IIRC they are offering service only to northern parts of North America to begin. The network is sparse right now and I think the plan is to start with Canada.
its increasingly feeling like we won't make it...cross-satellite link is out, so we're basically back to Iridium (read: DSL) speeds, meanwhile the company burns cash like crazy and needs to fundraise every 8 months. hearing it needs to be spun off to get funded separately
but then again I've been hearing that for awhile. Push come to shove moment seemed to be when the FCC said to stop bs'ing about theoratical network speeds in applications for rural broadband grants, I still don't understand how that got submitted..lawyers...
- To serve as an internet backbone they need to bounce off of ground stations
- In particularly busy areas where they might otherwise have routed traffic up, laterally, down they have to now go up, down resulting in 0.5x local available bandwidth.
- The latency is a few ms higher on long routes when using starlink as a backbone
None of this particularly effects their ability to serve as a consumer ISP. The advantages are still there. Also cross satellite laser links are still the plan as far as the public has been told, just somewhat delayed.
A day or two after that statement Elon said that the amount they were thinking of spinning it off right now is "0".
PS. Your use of "us" makes me think you might be an insider, I'm definitely not, but as an outsider it looks like SpaceX has a pretty good chance of pulling this off.
I would have thought that their backhaul ground station links would be many times faster than their user station bandwidth, since they can use large directional antennas.
Yes, the 0.5x multiplier is based on an over simplified model where everything is consuming the same resources... e.g. that might be onboard processing power. In reality it is probably much better than that, but considering what I was replying to I wanted to be conservative.
> FCC smacking us (article is more than a little generous towards us, thank Ajit Pai's reputation I guess)
This is the FCC protecting incumbent service providers on the basis that Starlink isn't already widely deployed. They are not claiming that they have any kind of technical analysis that Starlink will not in fact be low-latency (the threshold is a full 100ms, so not incredibly strict). They are claiming that simply because they do not already offer the service they will get lumped in with the high latency geo-stationary sat link providers;
"In the absence of a real world example of a non-geostationary orbit satellite network offering mass market fixed service to residential consumers that is able to meet our 100ms round trip latency requirements, Commission staff could not conclude that such an applicant is reasonably capable of meeting the Commission's low latency requirements, and so we foreclose such applications."
Frankly it's very sad, because the $16 billion in rural broadband grants are the perfect application for Starlink.
The funds are distributed by "multi-round, descending clock auction where bidders will indicate in each
round whether they will bid to provide service to an area at a given performance tier and latency. The auction will end after the aggregate support amount of all bids is less than or equal to the total budget and there is no longer competition for support in any area." [1]
I think if Starlink was allowed to bid as both a "low-latency" and "high performance" tiers, then at the pricing level they would likely have come in under, they would have wiped the floor with the competition and won a lion's share of that funding. So the FCC appears to be trying to keep them out of the most lucrative tiers to protect the incumbent approach of extremely expensive rural fiber deployments.
IMO a much smarter approach would have been to require some sort of performance bond from every applicant that they will successfully achieve the service levels they are bidding on. It's not like Comcast or Verizon haven't won massive subsidies in the past and then failed to deliver. The approach the FCC is taking (not surprisingly) just limits innovation and wastes taxpayer dollars. Actually in this case the dollars come from customer service fees, but they are effectively taxes.
I agree a performance bond would be best, with enough teeth that the worst case forces SpaceX to give every cent back or go bankrupt. Same with Comcast et al.
But you're unnecessarily ascribing malice. SpaceX has made some big promises, and if it delivers it deserves to hoover up basically all of that money.
But until the network is real, customers can sign up, and it's just a matter of scaling, a rational actor attempting to genuinely forward the interests of the country may well decide it poses too much risk.
I was under the impression that cross-satellite laser links would be coming with second-generation version of the satellites, as opposed to the first operational batches that've been launching. Has that changed?
That seems backwards. They should have much more ground station bandwidth to those giant tracking dishes than to little personal phased array antennas.
The problem with no sat-to-sat traffic is they'll need a crapton of downlink stations, which increases cost substantially. Also, they'll be geographically limited--no service out in the middle of the ocean.
> Also, they'll be geographically limited--no service out in the middle of the ocean.
There's a workaround for this, which is to stick a relay station out in the middle of the ocean on a ship that bounces the signal to a satellite closer to land. Expensive, but workable for the right price.
If there's a high enough density of commercial ships, you might be able to piggyback on the user terminals for this purpose. At the same time the military is likely one of the most interested customers and probably wouldn't be thrilled at relying on the typical distribution of commercial ships.
Last I heard, the current generation of sats do not yet have inter-sat comms hardware but it's planned for the next generation which should come in 2021
Is "global" not appropriate here?