IntArea WG Agenda

IETF 122 - Hybrid meeting, Bangkok (Thailand) + Online

17:00-18:30 Monday, March 17, 2025, Afternoon Session IV (Bangkok,
Thailand
(GMT+7))

Chairs:
Juan Carlos Zuniga
Wassim Haddad

1. Agenda Bashing, WG & Document Status Updates (Chairs)
   (5 mins)

Juan Carlos requested WG to review draft-ietf-intarea-tunnels as the
authors have requested a WGLC.

1. TipTop WG Intro, P. Pillay-Esnault (5 mins)

Taking IP to Other Planets: Padma requested people to participate in the
tiptop meeting scheduled on Thursday

1. Communicating Proxy Configurations in Provisioning Domains - T.
   Pauly, D. Damjanovic, Y. Rosomakho
   draft-ietf-intarea-proxy-config-04 (5 mins)

   • new co-author (Yaroslav)
   • Yaroslav presented changes from the latest version
   • Tommy Pauly mentioned that there will be interop testing between
     senders and receivers and he would like to wrap this up before
     the Madrid IETF (also potentially a hackathon project there)
   • Hopefully WGLC in Madrid

2. Adding Extensions to ICMP Errors for Originating Node Identification

   • B. Fenner, R. Thomas draft-ietf-intarea-extended-icmp-nodeid-01

Bill mentioned there was a lot of interest to add new data types but he
recommends going forward with the document as is and extend as and when
needed.

- Suggestion to do WGLC. Thumbs up!

PROBE: A Utility for Probing Interfaces - B. Fenner, R. Bonica, R.
Thomas, J.Linkova, C. Lenart, M. Boucadair
draft-ietf-intarea-rfc8335bis-00 (10 mins)
- Some Existing RFC 8335 implementations are not interoperable
- There is some warning text added to the effect of saying the design
of this mechanism is very specific (extension handling) and should not
be replicated
- WGLC? Will be discussed on the list

1. ICMP Extension Header Length Field - R. Bonica, X. He, X. Min, T.
   Mizrahi draft-ietf-intarea-icmp-exten-hdr-len-00 (5 mins)

   • (projector issue)
   • Ron is asking for the WGLC, despite the draft has been just
     adopted

2. The Multicast Application Port - N. Karstens, S. Cheshire, M.
   McBride draft-karstens-intarea-multicast-application-port-00 (10
   mins)

   • [Gorry Fairhurst] did you talked to the IANA ports team
     regarding the policy. If not, Gorry would connect the authors
     with them.
   • [Lorenzo Colitti] mDNS and other protocols are using unicast
     for a reason (faster, more reliable, etc). What are our guidance
     to the protocol designers?
   • [Nico Cvitak] Good idea. Question about discarding unicast.
     Would it be benefitial to accept unicast from those sources?

   • [David Schinazi] Are you solving port exhaustion on machines,
     or on the IANA registry?

     • Yes, on the IANA registry.
     • In this case recommendations are needed for new protocols so
       they do not reserve ports, otherwise problem not solved.

   • [Stuart Cheshire] (answering David's question) Not actually
     port exhaustion problem. Each user of multiple instances of the
     same application (e.g., same mcast chat app) needs a port. To
     Lorenzo's point this document does not encourage use of
     multicast instead of unicast but rather describes a preferred
     mechanism for use with multicast. Stuart has already had a
     conversation with Joe Touch regarding the port allocation.

3. IPv4 Routes with an IPv6 Next Hop - J. Chroboczek, W. Kumari,
   T.Hoiland-Jorgensen draft-chroboczek-intarea-v4-via-v6-03 (10 mins)

   • [Warren presenting] Adopt?
   • [David Lamparter] Minor concern: it should work on hosts. Need
     to be explicitly included (and there is an DHCP draft). Pls ship
     it!
   • [Tobias Fiebig] Please go ahead! It works on hosts. I have an
     IPv4-free AS. Just works.
   • [David Schinazi] Great! Support, happy to review.
   • [Maria Matejka] More guidance on cases when routing needs to
     be different depending on the nexthop? (audio was unclear)
   • [???] wants to check on implications for savnet for source
     address verification
   • [Krishnaswamy] Cisco NX-OS has it for more that 8 years. How
     multihop would work? Pls add more details.
   • [Eric Vyncke] IPv6 via IPv4 would work? (Warren: like we need
     more IPv4??) Should it be standard track or informational?

4. Safe( r ) Limited Domains - W. Kumari, A. Altson, E. Vyncke, S.
   Krishnan, D. Eastlake draft-wkumari-intarea-safe-limited-domains-04
   (10 mins)

   • Adoption?
   • [Warren] This doesn't change protocols. More like a
     guidance/tool for people who design protocols.

5. EVN6: Mapping of Ethernet Virtual Network to IPv6 Underlay for
   Transmission - C. Xie, J. Sun, C. Bao, M. Smith
   draft-xls-intarea-evn6-03 (10 mins)

   • Adoption?
   • [David Lamparter] At the very least you are missing a referce
     for the group ID. The document doesn't explain it
   • [Bob Hinden] I'm not a big fan of encoding semantics into IPv6
     inteface IDs/addresses.
   • [Eric Vyncke] As the responsible AD: I'm wondering if this
     work is best fit in intarea, and not, for example, NVO3. It
     might be a better fit for it. I.e., there is some work to be
     done to check other WG charters.

6. Enhancing ICMPv6 Error Message Authentication Using
   Challenge-Confirm Mechanism - K. Xu, X. Feng, A. Wang
   draft-xu-intarea-challenge-icmpv6-00 (10 mins)

   • [Eric Vyncke] (no hats) Interesting goal. What happens if the
     packet with challenge (or reply) is lost? Secondly, where is the
     state installed? What about DoS/amplification? This needs to be
     addressed in the draft.
   • [Gory Fairhurst] This is a real problem space, but I'm
     concerned...Can't you use the port information to verify the
     ICMP message?
     • [presenter] can't be done for ICMP....

7. Proposal for Updates to Guidance on Packet Reordering - G. White, I.
   Johansson, D. Das draft-white-intarea-reordering-00 (10 mins)

   • Interest in contributing? Is Intarea a right group for this?
   • [Lars Eggert] Yes, there are QUIC implementations (but the
     Mozilla implementation does not support this yet). Fast moving
     stacks are not your problem. Long-tail of old devices is the
     issue. It takes a while to fix old implentations
   • [Antoine F.] how do you balance between keeping a pakcet for
     too long vs loss?
   • [Stuart Cheshire] Good work, important. Within a TCP flow (or
     QUIC) decoding out-or-order is hard. But packets aren't like
     that. WiFi imposing some global ordering. Stuart believes that
     there is value in reordering instead of holding up the whole
     aggregate in WiFi networks (i.e. a missing DNS fragment
     affecting a video flow)
   • [Lorenzo Colitti] IPSec used to fail if re0sequencing was
     disabled. Modern TCO doesn't like jitter. It should be
     per-traffic type. TCP is robust, but some random protocols might
     not be
   • [Eric Vyncke] Is it IETF or IRTF? Maybe it's research

If time permits: [didn't have time]

1. Enhancing ICMP Error Message Authentication Using Challenge-Confirm

   Mechanism - K. Xu, X. Feng, Y. Yang, Q. Li
   draft-xu-intarea-challenge-icmpv4-00
   (10 mins)

2. Extending ICMPv6 for SRv6-related Information Validation - Y. Liu,
   Y. Liu
   draft-liu-6man-icmp-verification-07
   (5 mins)

3. IKEv2 Link Maximum Atomic Packet and Packet Too Big Notification
   Extension

   • D. Liu, D. Migault, R. Liu, C. Zhang
     draft-liu-ipsecme-ikev2-mtu-dect-09 (5
     mins)

4. DHCPv4 Option for IPv4 routes with IPv6 nexthops - D. Lamparter, T.
   Fiebig
   draft-equinox-intarea-dhcpv4-route4via6-00
   (10 mins)