[e2e] TCP in outer space

Jonathan Stone jonathan at DSG.Stanford.EDU
Mon Apr 16 13:15:08 PDT 2001


I buy the 4 RTT-> 3 RTT argument. Could you clarify what you said
about how this example reflects the rest of your samples-- that it's
both the median and the shortest?

I'm guessing that two metrics, RTT-delays-at-startup and connection
length (either data transferred or time) are involved, but I'm not of
that, or even (if it's right) which is which.

Or are all your samples equivalent, for some metric of RTT-quiet-time-
delays-during-startup? I'd buy that for HTTP-only connections, but
non-pipelined SMTP, etc, should have more application-level delays, surely?

--Jonathan Stone

More information about the end2end-interest mailing list