RE: A question on RPR's service policy

--- Nader Vijeh <nader <at> lanterncom.com> wrote:

Raj responds......see embedded comments:

> The only way to control dx1/dt, dx2/dt versus
> d(x1+x2)/dt is to have a
> virtual or physical queue per x at ingress, all
> transit nodes as well as the
> egress port. CBQ by definition can not separate the
> flows and a flow
> approaching its peak rate will cause second flow to
> fall below its committed
> rate (unless network is provisioned for the peak
> rate). 

Nader, why involve calculus when we dont need it
(dx/dt - really?). CBQ by definition does not care 
about rat's *fine assets* about flows. i.e. this is 
the difference between intserv and diffserv:
One deals with aggregated flows and the other deals
on per flow basis - like ATM - yuch:( (I must have
stepped on somthing slimy)

> The other approach is Virtual WFQ at each congestion
> point and a closed loop
> flow control. This requires per flow queues at the
> ingress/egress of each
> node, but a queue-less transit. 

Queue-less transit: Sounds like you have been sitting
under the bodhi tree too long!

> With this approach
> only committed rate
> provisioning is required for the network. In other
> words Sigma(CIR) can
> approach link Capacity C, while sigma(PIR) >> C.

Suspicious math! Never encountered this in 
school, college or normal life (dreams maybe -
mea culpa). Anyway, what kind of kool aid is this ?

__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more
http://taxes.yahoo.com/