AB
The Moon is arguably a satellite of Earth, but the distances are different from
LEO and GEO. That affects TCP performance.
TCP was designed with geostationary satellite hops in mind - see eg RFC829. I
use TCP across geostationary satellites daily.
I back up my statements with experience and references. What do you back yours
with? Why is UDP inefficient?
Lloyd Wood
http://about.me/lloydwood
________________________________________
From: Abdussalam Baryun [abdussalambaryun(_at_)gmail(_dot_)com]
Sent: 27 December 2013 06:01
To: Wood L Dr (Electronic Eng)
Cc: mcr+ietf(_at_)sandelman(_dot_)ca; ietf(_at_)ietf(_dot_)org
Subject: Re: Chang'E 3
So if TCP has problems then MPTCP has problems then they may not be preferred
for that satellite networking. For LEO SATNETs also there are problems for
using TCP. The UDP is a good solution in general but not efficient, but always
I think the SATNETs may need some special internet transport protocols because
they are having more dynamics in the net's lower layers.
AB
On Friday, December 27, 2013, wrote:
TCP won't work well from the Moon to Earth - slow start, congestion backoff as
a reaction to lost/errored frames, makes for inefficient use of the space link.
(And TCP may be a larger codebase than you want on an embedded system running
on a constrained radhard processor)
We explored TCP's distance limitations in
http://personal.ee.surrey.ac.uk/Personal/L.Wood/publications/index.html#protocol-radius
Lloyd Wood, Cathryn Peoples, Gerard Parr, Bryan Scotney, Adrian Moore, TCP's
protocol radius: the distance where timers prevent communication, Proceedings
of the International Workshop on Space and Satellite Communications (IWSSC
'07), Salzburg, Austria, 13-14 September 2007, pp. 163-167.
AFAIK vanilla ftp/TCP has (only?) been used on CHIPsat from LEO, but that did
not have much data to transfer.
We designed http://saratoga.sf.net with the idea of a continuous UDP stream of
data from Pluto in mind...
Lloyd Wood
http://sat-net.com/L.Wood/
________________________________________
From: ietf [ietf-bounces(_at_)ietf(_dot_)org<javascript:;>] On Behalf Of
Abdussalam Baryun [abdussalambaryun(_at_)gmail(_dot_)com<javascript:;>]
Sent: 26 December 2013 16:40
To: Michael Richardson
Cc: ietf(_at_)ietf(_dot_)org<javascript:;>
Subject: Chang'E 3
Transport protocol for satellites depend on the applications, the network
dynamics, links and transmission channels. IMO the use of MPTCP for Back2Earth
may not be the best option making the communication complex, not sure why did
you hope to use it with this project.
AB
On Sunday, December 15, 2013, Michael Richardson wrote:
The rover landing is rather exciting.
I'm wondering what the communication(s) protocol back to earth is.
(I'm hoping the answer will be MPTCP over IPv6 w/IPsec... )
--
Michael Richardson <mcr+IETF(_at_)sandelman(_dot_)ca<javascript:;>>, Sandelman
Software Works
-= IPv6 IoT consulting for hire =-