On 10-Apr-22 05:37, Toerless Eckert wrote:
On Sat, Apr 09, 2022 at 02:45:20PM +1200, Brian E Carpenter wrote:
Toerless askes for WG adoption of both drafts.
I haven't re-reviewed these recently but I did study them quite a while ago
and verified (by implementing it) that the GRASP/DNSSD interface works
quite well. I agree with Sheng's comment that more review is needed, but
I think that needs to come from a practising netops specialist. Is what
the drafts describe realistic in practice? I don't really know.
The services autoconfig functionality described was modelled pretty much
after the same functionality in the pre-standard ACP implementation
in Cisco IOS except that Cisco IOS ACP uses multihop mDNS forwarding.
without those services or experience in that implementation has been that
it is not possible to achieve zero touch rfc8366, aka: using the ACP with
existing networks to enable remote CLI/SDN configurability. Instead it is
exactly whats described in these drafts whats missing (vs. what we already
standardized in ANIMA). That pre-standard implementation functionality
was also valided by several customers and is publically documented.
The main difference is therefore really the replacement of mDNS
encoding/transport
of the service announcements with GRASP encoding/transport and we heard from
Stuart Cheshire that he agrees and supports our approach from the drafts.
Aka: i think we do have the netops experience.
Understood, but wouldn't we be happier with a review by a third party?
If nobody in anima is willing, maybe ask ops-dir to help?
Brian
fwiw my code sits at
https://github.com/becarpenter/graspy/blob/master/AskDNSSD2.py
https://github.com/becarpenter/graspy/blob/master/GetDNSSD2.py
Thanks!
Cheers
Toerless
Regards
Brian
_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima
