On 25 Nov 1998, Adam Atkinson wrote:
> I found out about this mailing list [pilc] today.
>
> It's uncannily similar to something I've been wondering about for the
> past few weeks... namely the behaviour of TCP/IP in extreme
> conditions.
This is one of the preoccupations of satellite types, who will happily
talk about congestion control algorithms and their susceptibility to
errors in long-pipe channels until the cows come home - hence the
tcp-over-satellite mailing list and recent bcp drafts.
> The meaning of "extreme" I've been playing with was at the
> level of having routers in orbit around the various planets and moons
> of the solar system, and other nodes moving about at various speeds,
> at distances from the Sun up to about the distance the Voyagers will
> reach in 20 to 30 years.
There was some lighthearted discussion of such issues on end2end in
July [hence my cc], back when the press were covering Vint Cerf
talking about what he dubbed the Interplanetary Internet Protocol
(hmm, IIP's already in use as a networking abbreviation; abbreviation
space not scaling for networking is already a problem, with multiple
reuse of e.g. PEP requiring context) and throwing 'internaut' around
as usual to get it into common usage (no chance!). See e.g.:
http://www.wired.com/news/news/technology/story/13909.html
Apart from routing tables scaling etc, and ad hoc network problems,
the 4.55 hours MSL limit in suggested clock implementations that I
kept coming across in the RFCs would be a real problem if you're out
mining the Oort.
[followups set to pilc as appropriate.]
L.
> On the face of it, this mailing list sounded
> like it might cover much of the same ground: low bandwidth, extreme
> variation in round-trip times, very high error rates and delays.
> (Stevens says TCP defaults to regarding 9 minutes as failure, which
> wouldn't even get you to Mars. Presumably you could recompile with
> upper limits in the multi-hour range, for the sake of argument.)
>
> I'd looked into "mobile IP" but it didn't QUITE seem the same sort of
> thing.
>
> Indeed, since it's based at NASA, this list might even be planning to
> consider interplanetary scales. Is it?
>
> --
> Adam Atkinson (ghira@mistral.co.uk)
> When we declare an alien species to be raman, it does not mean that
> _they_ have passed a threshold of moral maturity. It means that we have.
<L.Wood@surrey.ac.uk>PGP<http://www.ee.surrey.ac.uk/Personal/L.Wood/>
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 09:12:19 EST