[adkadskhj]

I don't know much about Starlink, but isn't one point of Starlink that eventually it could even beat wired connections for distance latency? Ie it's a shorter and more direct trip to use Starlink to get from US West to US East, for example.

Though this is quite a ways out i imagine.

Best of all this idea works in current tech, and can get even better with future tech when Starlink starts going Satellite <-> Satellite, avoiding unnecessary land hops.

I'm interested in Starlink for all use cases, once they get more satellites up.

[spookthesunset]

My understanding is that in addition to distance latency there is some non-trival amount of latency due to the use of TDMA. Unlike CDMA where everybody piles on the same spectrum at the same time, TDMA gives a timeslot to every station and you gotta wait your turn before TX/RX (this is a gross oversimplification of course).

While it doesn't add that much latency it is more than CDMA.

[syncsynchalt]

Not just TDMA - lasers in vacuum travel at the speed of light, but signals in optical fiber and coax copper are both 2/3 of that. This means that if (and that's still an if, not yet a when) starlink does laser between satellites it can beat terrestrial speeds.

I could see the end result being that we still do bent-pipe signaling over land and only laser interlinks for crossing oceans and for customers that pay for low ping to financial markets.

[dboreham]

Having the lowest latency is only useful to commodity traders. There's no way Starlink can ever compete with terrestrial service on QoS in general (due to Shannon theorem). Their USP is universal availability.