Re: Hop-by-hop [not draft-han-6man-in-band-signaling-for-transport-qos-00.txt]

[Toerless Eckert <[e-mail]>]

On Fri, Oct 20, 2017 at 12:51:23AM +0000, Manfredi, Albert E wrote:
> Absolutely, and that's the rub, isn't it? In-band QoS, no matter really whether it's done using an IPv6 3-tuple, or IP-agnostic 5-tuple (without encryption), is ultimately practical only inside walled gardens.

IMHO only true for DiffServ QoS unmodified, not for other forms
of QoS. DiffServ QoS even has fundamental challenges inside walled
gardens:

- Way to complex to dynamically adjust queue share of classes
  in enterprise deployments (only useful in mission specific network
  with well defineable persistent traffic profiles).
- Unnecessary differentiation now that a lot more media traffic is elastic
  and a lot more best effort traffic not buffer-bloaty

> These discussions bring me back to ... ATM. Trying to re-create ATM with IP. But ATM died, in large part, because of this.

I don't know. There is this cool flying car from the 50th in the Seattle
Museum of Flight. I don't think its historic state proves that planes
or cars are bad ideas.  1280 is the the 53 and flying cars are also coming
back and so will QoS.

Anyhow. Even though DSCP is in IPv6 header and this WG claims that
it must be the ones deciding on onpath packet inspection, i fear
the mayority here is not interested in QoS, so lets move the discussion
to TSVWG.

Cheers
    Toerless

> Bert
> 
> 
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> [e-mail]
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------

-- 
---
[e-mail]