[e2e] TCP, retransmissions, and ISPs with byte caps

[Colin Percival]

At 18 Jun 2002 00:10:41 +1000, gja at ureach.com wrote:
>On a different tangent, has anyone got pointers to
>studies of TCP retransmission rates in the context of
>'typical' consumer broadband access?

   I haven't seen any specific studies on that topic, but I have heard 
anecdotal reports that packet loss due to corruption can reach relatively 
high levels when decades-old, possibly corroded, underground cable is used.

>(And, e.g.,
>the (dis?)incentive for a dodgy ISP to reduce packet loss
>rates if they're also counting retransmitted packets against
>a customer's weekly/monthly usage cap.)

   Studies done in the early days of broadband access suggested that people 
tended to spend roughly the same time downloading material regardless of 
their access rate.  (Presumably, this behaviour would change as speeds tend 
to infinity!)  In that case, forcing retransmissions would only result in a 
reduction of traffic.

Colin Percival