[Due to the part of the spectrum I am on, I do not have believes or opinions.]
The laser based inter-links still not working has been subject on various conferences like AngaCOM etc.
But in my case: I have simply tried it *). And every Starlink user can do it, too: Use traceroute. And if you think "they might be hiding the hop-to-hops between Sats!", you can dig deeper using MTR behind the modem or simply rooting the modem itself.
Last time I have connected to a v3 Sat however was ~6 months ago. Maybe an active user reading this can try today?
You're equating occasional dropouts (which can happen for all kinds of reasons even in bent-pipe topologies) with the absence of inter-satellite links. That makes no sense.
The empirical way to test for the existence of ISLs would be to go to the middle of an ocean, safely out of reach of any ground station, and see what happens. If you get a connection, that can only be due to ISLs.
It seems like your actual complaints are with network/routing stability, and you're drawing invalid conclusions from there.
Do you have a link to a blog or writeup regarding the inter-links not working? Hard to find it without getting lost in "Troubleshoot your starlink device" SEO hell.
> Do you have a link to a blog or writeup regarding the inter-links not working?
The simpler answer is intra-constellation communication is a bleeding-edge technology. It's an extraordinary challenge for which extraordinary proof is needed to show success, not the other way around. SpaceX has solved most of the gating technical problems. But getting it to work reliably enough that it becomes more economic than ground-based backhaul will take time.
"The range of the ground stations are under 1500 miles and I really don't know where people are getting the idea that the lasers don't work."
Maybe because v1 and v2 did not even have working lasers on the hardware level...?
The idea is coming from "reality", Starlinks own reporting, industry talks, tech press etc.
Anyway, to shorten this we can agree that we have different definitions of what one expects from having a dedicate backbone. I would expect seamless handover amongst other things, which I have never ever seen, and unless you show me a video recording of a 24h Starlink session with MTR running I simply will trust the data I have over a random claim.
As said elsewhere in this thread: It is extremely hard to find detailed benchmarks from happy Starlink users. Next to all positive content is paid content. And a quick look at trustpilot & co clearly hint that there a huge chunk of Starlink customers might be unhappy. And even if it's just because their online gaming sessions getting interrupted on every Sat hand-over, which exists in reality, but not in your mind :)
Seriously, if you have access to any benchmark data sources, please gimme. I'm not here for "winning" an argument. Data, Data, Data.
It's quite fascinating, there's people who's only (or first) experience with Starlink is via lasers and there's people on the Internet who'll tell you it doesn't work (I forgot to mention Georgia and Kazakhstan)
A simple way to verify that their inter-sat links are not working and/or are not used is to simply sit and wait: If you are switched from one Sat to the next, you get new "session" and previous NAT state is lost. If this would be a meshed backbone, that would not happen.
Erm. That person has posted a detailed explanation on how he has measured.
How can this be ridiculous? Is it ridiculous because the data does not match your believes...? Confirmation bias?
It's Data. And it hints, amongst other things that they have seen the same that I am seeing on every single Starlink installation I got my hands on so far: There is no active handover, and no shared state between Sats.
And you are referring to the wrong layer, talking about the ground station. Of course that does not move, and does not forget about your IP. Wrong layer.
It's about the Satellites (!) not doing an active handover and not sharing L2 state, like it would the case for any meshed network, no matter if cellular or WiFi. The analogy here would be a WiFi access point or a cell tower, and you roaming from one to the next while having a phone call, not having any drop-outs. That's the industry standard for Wireless. Starlink isn't there (yet).
If you don't think that true data is true, check ARP table of the MAC of your gateway IP changing after handover.
You appear to be a happy Starlink user - so do you care to share some 24h benchmark with us to prove your claims? I would highly appreciate that!
So far sadly none of the "But it works!" people has been able or willing to provide a benchmark on their own setup. `
Again: I am not here to win and argument. But to change my conclusions, I need data that hints at my conclusions potentially being wrong. As explained elsewhere in this thread, due to lack of serious benchmarks, most of this is based on anecdotical data points.
And this link form wired is about something completely unrelated - getting more stable coverage by using multiple different providers. It does not even mention Lasers.
You also clearly do not know what Layer 2 on the ISO/OSI model is.
But you are in total rage mode.
Triggered because the actual data invalidates what your cult says? :)
Sorry, will ignore you from now on. Again: Religion is not my cup of tea, bold claims on powerpoint presentations neither, I prefer to use data. We simply do not share a model of the world that is compatible to discuss these kind of things. No harm done, but no thank you :)
That person is measuring periodic packet drops (which can also be caused by e.g. having an incomplete view of the sky), and you're drawing unsupportable conclusions on "session drops" from that. (The word "NAT" does not even occur in the observation thread you've linked!)
The laser based inter-links still not working has been subject on various conferences like AngaCOM etc.
But in my case: I have simply tried it *). And every Starlink user can do it, too: Use traceroute. And if you think "they might be hiding the hop-to-hops between Sats!", you can dig deeper using MTR behind the modem or simply rooting the modem itself.
Last time I have connected to a v3 Sat however was ~6 months ago. Maybe an active user reading this can try today?