[e2e] Question on assumption of congestion control algorithm

[Alexandru Andrei]

On Sat, 28 Apr 2001, Zhang Miao wrote:

> Hi, all
>       I have a question on the assumption of congestion control algorithm. In
> Jocobson's "Congestion Avoidance and Control"(1988), he models the network between
> two end systems as a pipe. After that, most research on this area uses this model
> as a basic assumption. This assumption requires that the path between the two ends
> is relatively stable. Well, in today's Internet, routing protocol also has a big
> impact on the behavior of the network. The routes in a router always change with
> time. To some extend, the route flap breaks the "pipe" model above. But I am not
> sure whether the routing behavior has much impact on the efficiency of congeston
> control algorithm. 
>       Is there any research related to this topic?
>       Thanks a lot!
> 
I think Vern Paxton has done some work in this area. Maybe an
interesting paper to read is "End-to-End Routing Behaviour in the
Internet" - IEEE/ACM Transaction on Networking vol.5 nr. 5, October 1997.

Regards,
Alexandru Andrei