[e2e] High Packet Loss and TCP

[Craig Partridge]

In message <DAC3FCB50E31C54987CD10797DA511BA0301CE7F at WIN-MSG-10.wingroup.windep
loy.ntdev.microsoft.com>, "Christian Huitema" writes:

>One can argue that this can lead to congestion collapse: if too many
>connections try to use a too small link, then they will all pack-off to
>this one packet every 4 second behavior, which may well exceed the link
>capacity. Also, the slow progress will lead to many job aborted midway,
>which is a loss of resource. It has been shown that, in these
>circumstances, admission control helps: preferably drop the SYN packets,
>so that the remaining connections have some bandwidth and can actually
>complete.

Didn't Jon Crowcroft capture some traces from the trans-Atlantic link
c. 1990 when we had precisely this problem for a few months -- the
number of packets in the pipe divided by the number of connections was
less than 1?

As I recall, he found some interesting behavior.

Craig