[e2e] High Packet Loss and TCP

[Jonathan M. Smith]

Hi - there was a group at U Mass that was looking at this
issue a few years back (Dan Bernstein's name comes to mind,
but I know the work was directed by Jim Kurose - Jim please
chime in if I have got this wrong) and I think the simplest characterization
is a bit error rate greater than ca. 10**(-6) makes things
really really bad. We (Ilija Hadzic and I, mainly Ilija) later confirmed
their model experimentally (with a couple of corrections/refinements)
using a cool system built by Hewlett-Packard
(HP75000 broadband analyser with a network impairment module).
Details are in Ilija Hadzic's Ph.D. thesis, which is at
http://www.cis.upenn.edu/~boosters/
                                                                 -JMS


At 03:40 PM 5/1/2003 -0400, Ross Callon wrote:
>My understanding is that there is some level of packet loss
>which causes TCP to back off to the point of stopping. My
>impression is that this is a sufficiently high loss rate that it
>shouldn't happen in a network which is behaving properly,
>and if it happens this should be considered a network failure
>rather than a TCP problem. (I am pretty sure that I saw this
>sort of behavior a few years ago when trying to access large
>files over a very bad link).
>
>Is there a paper which would describe what the appropriate
>loss rate is that would cause this problem? Is there any
>general understanding of what level of packet loss will cause
>serious problems?
>
>thanks, Ross
>
>(PS: Please CC me, since I am not on the Email list. thanks)