Cell Relay Archive

[albert.e.manfredi@boeing.com]

In article <376E6371.EAAA99EC@lucent.com>,
  "ronald h. davis" <ronaldd@lucent.com> wrote:

> fair enough, but queueing is not necessarily qos.  for instance, fifo
> queueing isn't qos because it doesn't prioritize traffic to determine
> whether some packets should be delivered first regardless of order of
> arrival.  lacking a lan which implements some notion of qos (and i
> don't know if there were proposals out at the time that the cif spec
> came out 3 or so years ago) it is fair to say that cif does not
support
> end-to-end qos as was previously suggested.

I agree with Ron. The point being that when one says "QoS," what is
really meant is "QoS guarantees."

Suppose you have an Ethernet-connected CiF workstation, which has
simultaneously opened some CBR channels and perhaps ABR or UBR channels
for FTP. It's not clear to me how you can guarantee the CBR QoS
parameters in such a scenario, unless you have implemented multiple
queues with different priorities over the switched Ethernet.

--
Bert
manfredi@arl.bna.boeing.com


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.