IntArea WG meeting
1. Agenda Bashing, WG & Document Status Updates (Chairs)
No comments on agenda / bashing
One draft in discussion with 6man [To clarify]
ICMP exten len on AD desk, will go to IETF last call
(Video presentation)
Questions / comments
- Xiaohu Xu: Draft useful, need some clarifications about how to
design elegant UDP tunnel (e.g. GUE and design issues)
- Ron Bonica: Useful document on dynamic terminology, how will the
community approach this document that has been written over 18 years
- D. Schinazi: Some parts of the document look obsolete from MASQUE
perspective, not sure about document value.
- W. Kumari: Some value in document, informational maybe
3. Communicating Proxy Configurations in Provisioning Domains - T. Pauly - draft-ietf-intarea-proxy-config-06
(Slides in person)
- Standard alternative to .pac files
- Details on destination rules semantic
- Some implementations done and currently being tested, request to
start WG last call
Questions / remarks
- Jim Taft: Some missing elements in the current document (on taking
into consideration resolved IP addresses), Request for clarification
on load balancing, Encouraging session reuse for proxies.
- Tommy Jensen: Agree with first remark from previous speaker, on
second point soft disagreement.
- W. Kumari: Client should attempt to stick to same proxy once a proxy
has been chosen.
- D. Schinazi: Method to signal need to use another proxy
(Slides in person)
- Some updates following last IETF about support of mechanism in
routers
- One update needed: extension about ICMP error on different address
family
Question / remarks
- J. Linkova: More text on behavior of filtering associated with mixed
address family
- T. Fiebig: clarify behavior of V6 with v4 next hop
- M. Blanche: other document might be needed about end to end support
of mechanism
- E. Vyncke: position on intended status for document
- M. Matejka: It should be standard, it is the desired behavior
(Slides remotely)
Question / remarks
- S. Krishnan: Great idea, some missing elements about behavior on
used ports on unaware hosts. Port assignement should be sent to IANA
and split from document.
- L. Colitti: Question on behavior when traffic received on address on
assigned port. Question whether port 49149 already in use somewhere.
What is the behavior if someone uses that port somewhere.
- D. Schinazi: Minor editorial fixes. In theory, draft complies with
ephemeral port assignement policies, but OSes do not respect this
policy sometimes. Maybe use a user assigned port in lower part of
the range.
- D. Thaler: Remark on socket use for multicast / unicast traffic on
same port
(Slides in person)
Question / remarks ==> After next presentation
7. Enhancing ICMPv6 Error Message Authentication Using Challenge-Confirm Mechanism - Y. Yang - draft-xu-intarea-challenge-icmpv6-01
(Slides in person)
- Mechanism introduced in IETF122, with feedback => Improved design
inspired by SYN cookie to make it stateless
Question / remarks
- S. Krishnan: Asumption about 1 route between 2 hosts, falls in
multipath context
- T. Fiebig: Document should be more written as a specification.
- R. Bonica: [To clarify]
(Slides in person)
- Draft in 6man WG
==> Need to have people read the draft and review
Question / remarks
==> No time for presentation, discussion taken to the list, and
presentation requested for IETF 124