[Chapter-delegates] Internet Society information session on "New IP" | July 28, 07:00 UTC & 13:00 UTC

[Ted Hardie]

Hi Dave,

On Thu, Jul 22, 2021 at 1:25 PM Dave Burstein <daveb at dslprime.com> wrote:


> if the IETF draft is implemented the way the telcos intend, it would shift
> control to the telcos. It's about extending the QoS in 5G across networks,
> so the telcos can sell guaranteed QoS from headquarters in Brussels to a
> branch in Naples.
>

The interesting thing about DetNet is that its QoS focus is on
predictability, not necessarily performance.  As the RFC points out: "In
general, a trivial priority-based queuing scheme will give better average
latency to a data flow than DetNet; however, it may not be a suitable
option for DetNet because of its worst-case latency."     For real-time
systems, that can be critical; an industrial control loop may need input at
predictable times, where the delivery of a best-effort QoS would be
inappropriately jittery.  Handling that doesn't mean, however, that the
intelligence in the network takes over the basic principles. It means that
with access to good timing and a different sort of attention to  buffer
management you can optimize for predictability over average performance.
That's not the optimization most flows in the network want now, and
achieving their desired optimality has historically been cheaper and easier
to deploy via overprovisioning rather than QoS-heavy machinery.  Unless
customers are forced to accept QoS-heavy flows being required for all
flows, the end result is that the network flows that don't need that
predictability won't use it, since they can get their needs met better and
probably cheaper without.

One of the key worries about some of the proposals has been that they
presume that the whole system must change in order to satisfy asserted new
needs.  I don't, again speaking personally, see a reason to accept that as
a premise, and I've seen no evidence to support it as a conclusion.

best regards,

Ted
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://elists.isoc.org/mailman/private/chapter-delegates/attachments/20210722/1b5debb5/attachment.htm>