[6797] in www-talk@info.cern.ch
Tragedy of the commons: more on multiple-connections
daemon@ATHENA.MIT.EDU (Simon E Spero)
Wed Nov 23 11:49:53 1994
Date: Wed, 23 Nov 1994 17:38:55 +0100
Errors-To: listmaster@www0.cern.ch
Reply-To: ses@tipper.oit.unc.edu
From: Simon E Spero <ses@tipper.oit.unc.edu>
To: Multiple recipients of list <www-talk@www0.cern.ch>
I just did some packet traces comparing the effects of multiple connections
versus sending the same amount of data over a single connection. For the
test case I looked at, I discovered that the parallel connections did indeed
take less wall-clock time. However, they also caused an increase in the number
of data packets sent of over 22% due to retransmissions.
>From this we see that the mutli connect version increases the amount of
bandwidth available to a particular user *at the expense of reducing the total
amount of bandwidth available*.
The extra bandwidth for this user is coming at the expense of other users-
thus we have a classic instance of the tragedy of the commons.
Simon // Remember, caring means sharing.
