I reduced the number of lists this message is sent to. If someone
feels it is important to the missing lists, please forward it to them.
Christian Huitema wrote:
By the way, it should be noted that the only rationale, if any, for
TCP spoofing in the satellite relays is the inadequacy of the
end-to-end TCP implementation. The specificities of satellites and
their interaction with transport protocols have been known for more
than 15 years, and the cure is also very well known: use large
windows, use selective acknowledgments.
Correct me if I'm wrong. Using big windows is good if you know your
path includes a satellite (or big pipe) link. It is not necessarily
good in general. In particular, if the receiving host is behind a
slow line (say, a modem at home), the only effect of advertising a big
window is to build a big queue at the router where the slow line
connects to, with a consequent series of bad effects on throughout as
soon as the queue overflows and packets are dropped.
If what I said is true, the problem is not inadequacy of the TCP
implementation, but it is more complex.
--
Francesco Potorti` (researcher) Voice: +39-50-593 203 (op. 211)
Computer Networks Group Fax: +39-50-904052
CNUCE-CNR, Via Santa Maria 36 Email: [email protected]
56126 Pisa - Italy Web: http://fly.cnuce.cnr.it/
This archive was generated by hypermail 2b29 : Mon Feb 14 2000 - 16:14:38 EST