Date: Thu, Jul 29th 2010
Meeting Minutes: MIPv4 Working Group
Attendees: MIP4 WG Chairs, Presenters, AD, Attendees (No one in Jabber + In-room)

1. Status of current WG documents

Chairs gave the status of the current working documents. The read-out is what is in the slides for each of the documents.

Nemo-v4-dynamic, new version upload requested, Kent unaware who has latest version.

Sri:Is charter coming to an end?
Pete: Recharter or shutting down coming after 3 documents get WGLCd.

2. Tunnel Type Change for Mobile IPv4 (TTC)
Alex's presentation. Not related to any WG item or on charter. Personal draft, draft-petrescu-mip4-tuntype-change-00.txt.

Slide 3,
Charles P: What's 64b?
Alex: Bytes, ping payload for latency measurement?
Charles P: What's the 3G technology?
Alex: It's HSDPA + HSUPA I believe.

Slide 5,
Sri: So each time you hand off, you update same tunnel or create a new tunnel?
Alex: ATM we have to create new tunnel when updating tunnel type
Sri: Isn't it just an implementation issue then, to update tunnel instead of creating new?
Pete: What actually kills the session?
Alex: If the update does not happen in right order, if we handover from IP-in-IP to IP-in-UDP the registration gets dropped?
Pete: Why?
Alex: HA doesn't accept the tunnel type change. I don't remember exact error message.
Charlie: There's no requirement in spec to drop the req request.
Alex: The RFC doesn't say that it should accept the request, either. Could be implementation issue.

Slide 6,

Kent: We already have a deployment using your scenario. Seems implementation-specific issue to me.
Alex: Maybe difference is that we try to keep same tunnel interface.

Slide 8,
Sri: Nothing in spec forbids you to change type mid-session.
Alex: Nothing allows it, either.
Sri: F-flag simply means that client always wants UDP tunneling regardless of NAT detection
Alex: Yes, but spec doesn't say when F flag should be used.
Henrik: Declining UDP encapsulation does not mean rejecting registration request!
Alex: Ok.

Slide 12, feedback

Kent: Implementations can do all this. I don't see need to change any specs.
Antti: How to detect in-transit packets changing encapsulation if nothing else changes?
Kent: Taking care of in-transit packets is a separate issue, you've chosen some tunnel type for a good reason, why would you change it without a handover?
Pete: Sense of room is that there's no ambiguity in specs. 
Alex: Will ask on the mailing list.

3. Closing comments

Alex: Nemo-v4-dynamic? What's the status?
Pete: Will be on WGLC when new version is up.

No comments.