[netext] IETF76 Meeting minutes

[<[e-mail]>]

Network-Based Mobility Extensions (netext) WG Minutes
IETF 76

WEDNESDAY, November 11, 2009, 1300-1500 Afternoon Session I

Chairs: Basavaraj Patil ([e-mail])
Rajeev Koodli ([e-mail])

Credit for these minutes to: Antti Makela(antti dot makela at tkk dot fi>

--------------------------------------------------



1. Agenda bashing, Blueseheets, Note takers, Jabber scribes      5 Mins
   Change in agenda: Flow Mobility and Inter-technology handovers moved
   to start due to conflicting schedule.

2. PMIP6 extensions for inter-access, presenters Rajeev Koodli and
   Jari Arkko

http://www.ietf.org/proceedings/09nov/slides/netext-12.ppt
and http://www.ietf.org/proceedings/09nov/slides/netext-6.pdf

Discussion on choices at
http://www.ietf.org/mail-archive/web/netext/current/msg00881.html and
slide 2 of JA's presentation

JA: Recharter or new WG (Netext2?)

Jonne Soininen/NSN: Everything is true, but Solution #1 is not needed
since IP does not notice anything, constrant to 2G/3G handover in 3GPP
JA: With different technologies where MTU changes IP level needs to be
aware
JS: Virtual links depend on operating systems, different on Symbian
and FreeBSD
JA: We shouldn't provide implementation advice, only conceptual
JS: Can it be done in high-level fashion?
Sri: We don't have triggering capability from UE -> MAG. How about
 moving only partial set of flows to a new interface? What is the
 evolution path? 
JA: Is triggering at L2 or L3? Discussion so far indicated that we
 don't want L3 solution, possible PMIP extensions would transfer
 triggers around the nextwork but they'd still be L2 triggers. With L3
 triggers you  can't have inter-access handovers and PMIP.
Sri: Can build everything on the network site, if triggering is out-of-band?
RK:  Regardless of choice with MAG interface signaling needed between
    LMA <-> MAG to allow for subset of flows being transferred

Juan-Carlos Zunlga: MAG <-> MN, Solution recommendations should be
 very specific Carlos J. Bernardos: I support writing BCP document
 Marcello: I think conclusions accurate.

Vote: Reasonable way forward, yes/no: ~ 17 yes, none no.

JA: Expecting text on the discussion as a basis for a minor recharter
Sri: Can triggers come from link-layer and PMIP support it on network side?
JA: Should be ok
Marcelo: There is no standard-track document on this.

3. Runtime LMA Assignment Support for Proxy Mobile IPv6, Jouni
Korhonen http://www.ietf.org/proceedings/09nov/slides/netext-8.ppt
I-D: draft-ietf-netext-redirect-00

About 5 people have read draft
No comments forthcoming

RK: What is the motivation for mid-session LMA assignment?
JK: Currently assignment only applies to initial assignment, input
  needed for motivation - do people find this useful? Applies to
  handover cases where you might cross a boundary where administrator
  has decided to switch LMAs. Lots of assumptions in such cases (sharing
  of mobility sessions etc)
RK: If session is anchored at specific LMA, is there session continuity.
JK: Have you read draft?
RK: No
JK: Open question at the moment, there is introduction to a discussion
  in the draft
RK: I don't think this should be done at all. Why both soft and hard
  support needed? Simpler approach needed, now three-level complexity.
JK: So, take handover exchange away, only apply assignment on initial
PBU/PBA?
RK: Yes

Hidetoshi Yokota: Mid-session redirection, where LMA sends back
 different LMA, session would break.
BP: When you a redirected to new MAG, the session should go to the
 anchored LMA
HY: It's just an example
BP: Bad example.

Kent Leung: Why are we doing LMA reassignment during handover? I don't
 see justification. Question: What is the option in PBU?
JK: If we apply only in initial, we don't use option. Draft doesn't
 comment on options yet.

4. Bulk Re-registration for Proxy Mobile IPv6, Basavaraj Patil
http://www.ietf.org/proceedings/09nov/slides/netext-1.pdf
I-D: draft-premec-netlmm-bulk-re-registration-03

Marcos: Is this same as group ID?
BP: Not to my understanding
Marcos: I propose merging group ID with Charlie's draft
RK: Charlie's draft tries to combine multiple ID's concerining single
 node
Marcos: So, there's a connection, in this case we are just doing it
 for multiple nodes. Better to have all in single draft.
BP: Charlie not here, but we can have different sets of IDs depending
 on policy. We can continue this on list.
Qin Wu: How do you define flag B, it has two purposes, what if both
 purposes happen at same time.
BP: We'll provide a clarification in future revision.

Localized routing - PS and solutions discussion

5. PMIPv6 Localized Routing Problem Statement
http://www.ietf.org/proceedings/09nov/slides/netext-2.ppt
draft-ietf-netext-pmip6-lr-ps-01.txt
BJ: Few people have read the document, please send comments, also on
 the roaming model.
Glen Zoren(?): I was under impression that PMIP domain was set of
 nodes which could set up SA's between each other - so you can have
 dynamic domains. 
Carlos: That's not incorrect as such, but can't make that assumption.
GZ: I didn't mean all PMIP entities in the world.

6. draft-oulai-netext-opt-local-routing-01.txt,  Suresh Krishnan
http://www.ietf.org/proceedings/09nov/slides/netext-9.ppt
>
Kent Leung: Is the scope limited to some cases only
SK: Yes.
BP: What do you mean by optimized handover?
SK: Not easy problem but if there's RO between MAGs and MN switches
 MAGs, we can use that
KL: Is it possible to map the problem spaces of different solutions to
 different drafts for comparison?
BP: Yes, at least we can talk now about them allow
Marcos: What does it mean "inter-MAG tunnel may be preconfigured"
SK: If some MAGs are close to one another, who can expect handovers
 between them, they can prepare for RO

7. draft-koodli-netext-local-forwarding, Rajeev Koodli
http://www.ietf.org/proceedings/09nov/slides/netext-11.ppt

 ?: What triggers the transaction?
RK: Trigger is still open.
[Several people did not state names]
?2: Two LMA's are not talking to each other, how's the authentication?
RK: If the LMAs trust each other, even if not talking, authentication is ok.
 Trust doesn't mean you have to talk. You don't want LMA <-> LMA signaling.
 What's is defined is MAG <-> MAG, with some LMA interaction. What's
 happening is dynamic delegation with fine granularity.
BP: Cutting off discussion

8. draft-wu-netext-local-ro Qin Wu
http://www.ietf.org/proceedings/09nov/slides/netext-3.ppt

Mohana Jeyatharan: LMA needs to know where prefix is from
QW: Ok, we'll address this in next revision

9. draft-loureiro-netext-pmipv6-ro-01.txt
http://www.ietf.org/proceedings/09nov/slides/netext-4.ppt

BP: That concludes discussion on localized routing

10. New Proposals:

Multihoming extensions for Proxy Mobile IPv6 Carlos J. Bernardos I-D:
 draft-bernardos-mif-pmip-01.txt 5 mins

Hui Deng: Clarifications on weak/strong host models?
 BP: Let's take this on mailing list

Service Flow Identifier in Proxy Mobile IPv6
Hui Deng - draft-hui-netext-service-flow-identifier-01.txt   5 Mins

HD: WG draft?
BP: We need to do the rechartering process, then we can consider
 taking this on.

Bulk PBU using Bitmaps
Mohana J. - draft-jeyatharan-netext-pmip-bulkpbu-bitmap-00  5 Mins

MJ: Comments on the way forward?
BP: We have adopted group ID approaches as WG items, for these we need
 to do more discussions on mailing lists.

Virtual interface for supporting multihoming and inter technology
 handover Tranh M. Trung -  draft-trung-netext-virtual-interface-01 5
 Mins

Sri: Can you statement the differences between this and other related
 document(s)?
TMT: I try to limit the modifications to mobile node.
Sri: If it's a simple change, does this require a new document? Better
 work with existing ones
BP: Moving on

ITHO support in MN for PMIP6
Hidetoshi Y./Sri  G. draft-yokota-netlmm-pmipv6-mn-itho-support-02.txt

Reflector Extension for Route Optimization Agent Xiansong Cui
 draft-cui-netext-route-optimization-agent-ext-01.txt 5 Mins

CJB: Why LMA, not MAG doing this? Some work done for RO in Nemo,
 please take a look
BP: Need to shut down and conclude

BP: Meeting adjourned, rest of the presentations not given due to lack
 of time.

Differentiated Services Support for Proxy Mobile IPv6 Behcet S./Frank
 Xia - draft-xia-netext-qos-00.txt

Gateway Initiated Dual-Stack Lite Deployment Sri G.
 draft-gundavelli-softwire-gateway-init-ds-lite-00

PMIPv6 and Network Mobility Problem Statement Carlos CJ.
 draft-bernardos-netext-pmipv6-nemo-ps-00