Brian E Carpenter <brc(_at_)zurich(_dot_)ibm(_dot_)com> wrote on 11/14/2006
11:36:40 AM:
...
This illustrates some of my concerns about this requirements work being
done outside the IETF.
...
2. The notion that solutions such as precedence and preemption
are (a) requirements and (b) applicable to all applications just
doesn't compute for me. We'd actually need to understand at a more
basic level what the functional requirements are, in terms that are
meaningful for a datagram network. I don't believe that will
happen in ATIS or ITU-T.
Brian (personal opinion)
Exactly.
In the circuit switched world, a circuit is either up or down, and
"preemption" means taking the circuit down.
But in the IP world, there is a full continuum of states in between. Some
of these are candidates for a useful service, and some of which aren't.
The understanding of this continuum, and of the (intended and unintended)
consequences is much stronger in the IETF than in the
historically-circuit-switched world.
Some of the possibilities in that continuum include (in no particular
order):
- Allowing extra sessions in, and permitting degradation in QoS across all
sessions.
- Allowing a higher packet drop rate across all the "lower priority" calls.
- Negotiating a lower bandwidth allocation, possibly accompanied by a
changing to a lower rate bandwidth codec when a higher priority session
needs to "preempt".
- Negotiating (or arbitrarily imposing) a different PHB (e.g. AF or BE
rather than EF) for lower priority sessions when a higher priority session
needs to "preempt".
- Different Capacity Admission Control mechanisms for different priority
sessions.
The analysis/understanding of these (and other) alternatives is much better
done in the IETF than in the historically-circuit-swiched SDOs.
Janet
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf