[e2e] TCP in outer space

Mark Allman mallman at grc.nasa.gov
Mon Apr 9 07:18:26 PDT 2001

Regardless of my email address, I have no first-hand knowledge of
what kinds of networking is (or, is not!) happening on ISS (not that
I haven't heard some horror stories!).

> Assuming the same TCP is being used for the emails , then
> how is the TCP still surviving in such long delays, as the
> problems in long latency especially in satellite links are already
> a known fact.

Yes, we know about the problems.  But, it does work.  And, it does
work well in many cases.  Could it work better?  Yes.  But, to
insinuate that TCP does not work over satellite links is wrong.

> If i am correct, the latencies from the space ship to earth would
> certainly be much higher and would that be the cause why the crew
> had problems with the email ?? 

Actually the delay between the ground and ISS is roughly the same as
the delay over a Earth->Earth connection using a GEO hop because ISS
uses a GEO satellite for its communications (I believe -- the space
shuttle does, and I am almost suer ISS is about the same on this
front, but it is Monday morning...).  (ISS is in a *much* lower
orbit than a geosyncronous satellite).

For an overview of the issues with TCP in satellite networks, please
see these two papers (and the references within):

    Mark Allman, Dan Glover, Luis Sanchez. Enhancing TCP Over
    Satellite Channels using Standard Mechanisms, January 1999. RFC
    2488, BCP 28.  

    Mark Allman, Hans Kruse, Shawn Ostermann. A History of the
    Improvement of Internet Protocols Over Satellites Using
    ACTS. Invited paper for ACTS Conference 2000, May 2000. 


Mark Allman -- BBN/NASA GRC -- http://roland.grc.nasa.gov/~mallman/

More information about the end2end-interest mailing list