[Starlink] First tests of the Starlink REV4 (aka gen3)

[Ulrich Speidel]

On 25/01/2024 1:37 am, Alexandre Petrescu via Starlink wrote:
> Thanks for the tests!
>
> The dl/ul speeds 300/15 mbit/s are impressive.

"Speeds" (observed data rates) in terms of Starlink hardware are 
actually fairly meaningless as they depend on:

  * Satellite(s) involved in the data transfer over the duration of the
    measurement(s).
  * Load on those satellites, which depends on the number of other
    current users whose traffic goes via those birds, and what these
    users are up to. Note this changes between handovers.
  * RTT to other end of transfer path.
  * Packet loss (including beyond Starlink's network)

Unless a Dishy is able to handle communication via multiple satellites 
in parallel, I would expect the rates to be the same more or less 
regardless of model used. In fact, I would expect a Dishy model that is 
able to align itself to do marginally better over time.

>
> At video pointer 5:53 the reported Ping ?/dl/ul 88/204/121 ms and Jitter
> 9.2 ms seem interesting.
>
>     ==> I am not sure which of the two (ping or jitter) you name 
> 'latency'?
All of them I guess.
>
>     ==> I am not sure why the dl (download) ping ms is higher than the ul.
Because that is where you have the longer queues.
>
>     ==> I don't know what is the first ('?') parameter reported as 88ms
> for Ping?

I presume unloaded ping RTT. The second is ping RTT during the download 
test (with inbound queues loaded), the last is ping RTT during the 
upload test (with outbound queues loaded).

I'd also note here that these are values obtained from Kyiv, which isn't 
a Starlink environment that is easy to assess. For one, we don't know 
whether there are Starlink gateways in Ukraine, or if there are, where 
they'd be. We know that there aren't any in Russia or Belarus. We know 
that there are gateways in Turkey, Poland and Lithuania and further 
afield that can reach satellites that users in Ukraine can reach by 
direct RF link. We don't know where else there may be gateways - not 
every jurisdiction publishes this - or whether Starlink may even be 
operating opportunistic fair weather optical gateways which they don't 
have to disclose to anyone. Plus we don't know what may be carted into 
the area or away from it via laser links to gateways much further away.

>
> I wonder whether the DHCPv6-PD is still supported by REV4 and whether
> the allocated prefix is still a non-/64 (i.e. a /56 delivered by
> DHCPv6-PD reported earlier on this email list by Steven on Dec. 12, 2023)?
>
> Alex
>
> Le 23/01/2024 à 10:07, Oleg Kutkov via Starlink a écrit :
> > I conducted the initial comparative tests of the new terminal in
> > Ukraine. I guess it's not a really "legal" because the new terminal is
> > not certified and not selling outside the US for the moment. But who
> > cares.
> >
> > Here's a video: https://youtu.be/hWPMpJrjd1g 
> <https://youtu.be/hWPMpJrjd1g>
> >
> > I will try to do more technical tests next week. There will be a new
> > video.
> >
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink 
> <https://lists.bufferbloat.net/listinfo/starlink>

-- 
****************************************************************
Dr. Ulrich Speidel

School of Computer Science

Room 303S.594 (City Campus)

The University of Auckland
u.speidel at auckland.ac.nz  
http://www.cs.auckland.ac.nz/~ulrich/
****************************************************************


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/starlink/attachments/20240125/862966ed/attachment.html>