[LARTC] FW: Some queueing disciplines that I wrote.

[Stephen Braithwaite]

> > The definition of a flow need not be the TCP definition of a
> > flow.  I am not sure if it will help, but any the queuing
> > discipline and ingress que filter are able to work with any
> > combination of protocol, source port number, source ip, dest
> > port, dest ip as the definition of a flow.  This may or may
> > not help.  
>   
> Ah, that's very interesting. So you could assign all traffic to/from a
> 'hog' ISP customer to the elephant category.

You cannot assign it as such, it has to happen automaically.

If you made the definition of a flow to be the
source/destination IP number then the flow consisting of
traffic to/from a 'hog' computer would find itself soon find
itself designated as an elephant.

If this is deployed on the router where the NAT occurs, then 
the queuing discipline sees the internal IP numbers.

The time scales over which a flow becomes/ceases to be an
elephant are configurable.  There is also a mechanism to have
the queuing discipline not purely mice and elephant and not
purely fair queueing, but somewhere in between.