Tuesday, November 3, 2015< ^ >
Room Configuration
Room Occupants

[00:01:46] Alex Petrescu joins the room
[00:01:56] <Alex Petrescu> Hello, I will do some jabber scribe.
[00:02:32] <Alex Petrescu> The slides are at
[00:02:54] <Alex Petrescu> Now showing slide: Next Step
[00:05:02] <Alex Petrescu> Fred Templin is FT is going to present
[00:05:30] <Alex Petrescu> Slide: AERO DHCPv6 Prefix Delegation Implications
[00:05:37] <Alex Petrescu> Slide: AERO Client Operation
[00:06:01] Simon Romano joins the room
[00:06:02] <Alex Petrescu> Slide: AERO Client Model
[00:06:37] <Alex Petrescu> Slide: Client as Host
[00:07:31] <Alex Petrescu> Slide: Client as Router
[00:08:03] <Alex Petrescu> Slide: Client as a Hubrid Host/Router
[00:08:35] <Alex Petrescu> Slide: Summary
[00:10:07] <Alex Petrescu> Seil Jeon is SJ is going to present
[00:10:48] <Alex Petrescu> Slide: Use Cases and API Extension for Source IP Address Selection
[00:11:02] <Alex Petrescu> Slide: Status
[00:11:59] <Alex Petrescu> Slide: Overview
[00:12:46] <Alex Petrescu> Slide: In more details...
[00:13:46] <Alex Petrescu> Slide: Prior RFC
[00:15:04] <Alex Petrescu> Slide: Example
[00:16:28] Simon Romano leaves the room
[00:17:00] <Alex Petrescu> Slide: ON_NET Property
[00:18:00] <Alex Petrescu> Slide: Way Forward
[00:18:17] <Alex Petrescu> Dapeng Liu is DL
[00:18:37] <Alex Petrescu> DL: you want to define this new src IP address to act here?  But how can you know what is the IP addres in the current subnetwork
[00:18:51] <Alex Petrescu> SJ: application may know it, preferences, deliver it to IP stack
[00:18:58] <Alex Petrescu> SJ: if multi-PHY then you dont follow
[00:19:09] <Alex Petrescu> DL: how the MN know what is in the current subnet, when moving
[00:19:39] <Alex Petrescu> SJ: missing part, but I dont think we havet o put it here.  In other draft maybe.  Terminal may get that information, prefix coloring or cost, may be addressed in other draft
[00:19:49] <Alex Petrescu> SJ: application this draft is about, deliver app preferences
[00:19:52] <Alex Petrescu> Sri Gundavelli is SG
[00:19:59] <Alex Petrescu> SG: I agree with concept, but need to think more generic.
[00:20:11] <Alex Petrescu> SG: app may need to specify certain properties, qith reqs a, b and c
[00:20:27] <Alex Petrescu> SG: if you include certain prefix properties, mix and match, app that best matches these reqs.
[00:20:33] <Alex Petrescu> SG: this is just example,
[00:20:39] <Alex Petrescu> SG: I think we should maybe take it forward.
[00:20:56] <Alex Petrescu> SJ: I agree, but if you can generalize it is good, but a good example is good that we'll have to address
[00:21:04] <Alex Petrescu> SG: general framework and focus on one example?
[00:21:18] <Alex Petrescu> SG: give general framework and mayebe in same draft do this property
[00:21:30] <Alex Petrescu> SJ: makes sense you think?  then why we have to care if it is one or many.
[00:21:40] <Alex Petrescu> SG: if you have proper framework then ytou dont reinvent
[00:21:45] <Alex Petrescu> Danny Mosses is DM
[00:21:56] <Alex Petrescu> DM: didnt read latest version, but we had discussions about this in the past.
[00:22:38] <Alex Petrescu> DM: I understand that in general yoiur concern is that after MN did handoff it might own addresses which might not be the most optimal ones, therefore makes sense to request a new sustained address, then app may enjoy the most optimal sustained address after handoff
[00:22:53] <Alex Petrescu> DM: assume we have a flag enabling app to request a new IP address or not request.
[00:22:59] <Alex Petrescu> SJ: flag we have to do this
[00:23:20] <Alex Petrescu> DM: question is can you give an example of use-case where an app would prefer to _not_ request a new sustained addr.
[00:23:34] <Alex Petrescu> DM: an app _allways_ rprefers such a n address.
[00:23:52] <Alex Petrescu> SJ:can you make sure no app will not require this kind of situation?
[00:24:15] <Alex Petrescu> SJ: I believe we need to think more about usecases, yes, but I think we make and we make more architecture?  should not be our work.
[00:24:24] <Alex Petrescu> SJ: not sure qhere we have to go, maybe a compromise.
[00:24:32] <Alex Petrescu> SJ: an example may help make sesne.
[00:24:37] Tae You joins the room
[00:24:40] <Alex Petrescu> SJ: we can not imagine all the things that never happend
[00:25:07] <Alex Petrescu> DM: cant give the opposite example, but I m still concerned about creating more and more auhtoritative ways for apps to use different hthings.
[00:25:35] <Alex Petrescu> DM: if an app writer needs to learn so much in order to use socket interface, may lead to bad interpretations.  We should not flood the iface, unless we sure we need it.
[00:25:55] <Alex Petrescu> SJ: once it was proposed in ..., the 3 type of src address has changed some ways of use Internet for app.
[00:26:05] <Alex Petrescu> SJ: correspondingly, some after work should be follow.
[00:26:10] <Alex Petrescu> Brian Haberman is BH
[00:26:16] <Alex Petrescu> BH: rule 8 in address selection process?
[00:26:18] <Alex Petrescu> SJ: yes
[00:26:41] <Alex Petrescu> BH: is it possible to the actual way is to employ rule 6 - use labels, and assign priorities to prefixes.
[00:27:12] <Alex Petrescu> BH: not sure I see an easy to use the address selection policy distribution (DHCP).  But if you use labels, then you can find a way to change policy models, rather than playing with prefix length rules.
[00:27:17] <Alex Petrescu> SJ: for this I ...ject.
[00:27:28] <Alex Petrescu> SJ: can not say anything, but I have to check rule number 6.
[00:27:32] <Alex Petrescu> BH: policy table,
[00:27:46] <Alex Petrescu> SJ: you say using that rule 6 w/o API then we can think?
[00:27:52] <Alex Petrescu> BH: no, not w/o API.
[00:28:01] <Alex Petrescu> BH: I want label X.
[00:28:18] <Alex Petrescu> SG: we can tight it to coloring, deliver properties, either in RA, o n SLAAC, or DHCPv6.
[00:28:30] <Alex Petrescu> BH: something else I come later.
[00:28:40] <Alex Petrescu> DL; no other comments, can we ask adoption?
[00:28:42] <Alex Petrescu> SJ: I ask adoption.,
[00:28:52] <Alex Petrescu> DL: poll: how many people read the doc:
[00:28:59] <Alex Petrescu> DL: about 4-5?
[00:29:08] <Alex Petrescu> DL: it depends how many people read.
[00:29:19] <Alex Petrescu> DL: trigger some discussion on the email list.  Today not enough feedback.
[00:29:31] <Alex Petrescu> DL: more people read the doc, more energy, then we can get interest work on this.
[00:29:34] <Alex Petrescu> SJ: in email list?
[00:29:43] <Alex Petrescu> DL: we can do that on email list, or next meeting.
[00:29:46] <Alex Petrescu> SJ: thank you.
[00:30:10] <Alex Petrescu> Ha Anthony Chan is HAC is going to present
[00:30:38] <Alex Petrescu> Slide: Enhanced mobility anchor wt
[00:31:38] <Alex Petrescu> sSlide: LM and FM in RFC7429
[00:31:51] <Alex Petrescu> HAC is looking for slides on anothe rcomputer
[00:33:28] <Alex Petrescu> HAC and DL look fpr slides
[00:33:38] <Alex Petrescu> DL: next agenda item.
[00:34:00] <Alex Petrescu> Danny Moses is DM is going to present
[00:34:38] <Alex Petrescu> Slide: Mobility Exposure and Selection WT Status and Next Steps
[00:35:12] <Alex Petrescu> Room: turn the mic
[00:35:57] <Alex Petrescu> Slide: Item #1
[00:37:28] <Alex Petrescu> Slide: Item #2 and #3
[00:39:55] <Alex Petrescu> SG: Jouni's draft is ADP
[00:40:04] <Alex Petrescu> SG: NDP is the draft...
[00:40:17] <Alex Petrescu> SG: We plan to do something of draft-korhonen
[00:40:17] Simon Romano joins the room
[00:40:29] <Alex Petrescu> SG: an option for DHCP another for ND
[00:40:36] <Alex Petrescu> DM: for NA you want the WG to adopt that?
[00:40:40] <Alex Petrescu> SG: we need to make a call
[00:40:46] <Alex Petrescu> Slide: Item #4
[00:42:07] <Alex Petrescu> DL: for the 1st item you new draft, with reference to...
[00:42:15] <Alex Petrescu> DL: do you plan a new draft, is it the WT1?
[00:42:17] <Alex Petrescu> DM: yes
[00:42:27] <Alex Petrescu> DL: Item 2 and 3, your request the WG to adopti?
[00:42:42] <Alex Petrescu> DM: I am asking to initiate a WG LG for ... and adoption on the DHCP draft.
[00:42:59] <Alex Petrescu> DL: we can initiate the WG LC for this one, I want to make a poll
[00:43:05] <Alex Petrescu> DL: how many people read this document?
[00:43:14] <Alex Petrescu> Room: zero
[00:43:43] <Alex Petrescu> DL: suggest people have review the document and then allow comments on this, and ..shment on this, before LC w need sufficient reviewing of this document
[00:43:50] <Alex Petrescu> DL: I suggest you to talk off the line
[00:44:05] <Alex Petrescu> DM: this draft has nothing new compared to the last individual draft.
[00:44:09] <Alex Petrescu> Marco Liebsch is ML
[00:44:17] <Alex Petrescu> ML: you request more infoput for item 4?
[00:44:23] <Alex Petrescu> ML: what's missing?  interest or details?
[00:44:43] <Alex Petrescu> DM: regarding Alper's interest... we dont expect each topic one draft, but several drafts.
[00:44:50] <Alex Petrescu> DM: seems nobody interested in that.
[00:44:56] <Alex Petrescu> DM: if so then we may need to drop it.
[00:45:04] <Alex Petrescu> DL: for the DHCP one,
[00:45:14] <Alex Petrescu> DL: how many people read this DHCP draft, raise hand
[00:45:18] <Alex Petrescu> Room: zero
[00:45:24] <Alex Petrescu> DM: but I receive input from several people.
[00:45:37] <Alex Petrescu> BH: as an AD possibly you got caught with ... draft...
[00:45:52] <Alex Petrescu> BH: q to the room, anybody read any other drafts around?
[00:45:54] <Alex Petrescu> Room: some
[00:45:59] <Alex Petrescu> BH: just checking.
[00:46:11] <Alex Petrescu> DM: ...
[00:46:31] <Alex Petrescu> SG: part of problem, not only with these WT items, but also FCDP, but in general , ...
[00:46:43] <Alex Petrescu> DM: I'll prepare a presentation for the next f2f and then ask for adoption.
[00:47:10] <Alex Petrescu> HAC found slides
[00:47:17] <Alex Petrescu> H. Anthony Chan is HAC is going to present
[00:47:32] <Alex Petrescu> Slide: Enhanced movilit..
[00:47:39] <Alex Petrescu> Slide: Work since IETF93
[00:48:04] <Alex Petrescu> Slide: IP prefix anchor switching.
[00:48:14] <Alex Petrescu> test
[00:49:06] <Alex Petrescu> SlidE: IP prefix anchor in new network
[00:49:49] <Alex Petrescu> Slide: Channging anchor with new prefix
[00:50:49] <Alex Petrescu> Slide: Switchning anchor in mid-session
[00:51:59] <Alex Petrescu> Slide: Relations to work of other wt
[00:52:49] <Alex Petrescu> SG at mic
[00:52:58] <Alex Petrescu> Carlos Bernardos is CB queueing.
[00:53:11] <Alex Petrescu> Slide: IP prefix anchor in new network
[00:53:31] <Alex Petrescu> SG: back to call flow
[00:53:48] <Alex Petrescu> Slide: Switching anchor in mid-session
[00:53:54] <Alex Petrescu> SG: you move prefix from p-AR to n-AR
[00:53:59] <Alex Petrescu> SG: then BGP route updates?
[00:54:03] <Alex Petrescu> SG: for /64 prefix?
[00:54:16] <Alex Petrescu> HAC: this router will be sending bgp updates.
[00:54:22] <Alex Petrescu> SG: allocation is for /64?
[00:54:37] <Alex Petrescu> HAC: yes /64
[00:54:45] <Alex Petrescu> SG: two MNs attached to same p-AR
[00:54:56] <Alex Petrescu> SG: 2 dfifferent prefixes /64s, but same block.
[00:55:06] <Alex Petrescu> HAC: two different mobivles, different mobiles
[00:55:13] <Alex Petrescu> SG: but from different blocks.
[00:55:21] <Alex Petrescu> SG: so you do bgp update from /64?
[00:55:26] <Alex Petrescu> SG: it;s like moving host routes
[00:55:30] <Alex Petrescu> HAC: it's a router.
[00:55:43] <Alex Petrescu> HAC: whenever prefix moves it is moved, it is basically moving a prefix
[00:55:47] <Alex Petrescu> SG: so it is host routes.
[00:55:55] <Alex Petrescu> HAC: if it is a router it may move...
[00:56:14] <Alex Petrescu> SG: whatever prefix you give, it is hosted in AR link.
[00:56:26] <Alex Petrescu> SG: the router is hosting prefix (Dnot matter mn or mr).
[00:56:39] <Alex Petrescu> SG: MN's egress link and access router.
[00:56:44] <Alex Petrescu> HAC: owned by the MN, right?
[00:56:51] <Alex Petrescu> SG: but it is not delegated prefix, it is a hosted prefix.
[00:56:55] <Alex Petrescu> CB:
[00:57:10] <Alex Petrescu> CB: I share the comment from Sri, you are making host route udpates with BGP, scalability concerns.
[00:57:17] <Alex Petrescu> CB: some discussion was for Pete's draf.t
[00:57:28] <Alex Petrescu> CB: analysis discuss this scalability issues you have in the draft?
[00:57:34] <Alex Petrescu> CB: not sure host-route updadtes will work.
[00:57:43] <Alex Petrescu> HAC: centralized or distributed routing.
[00:58:01] <Alex Petrescu> HAC: when centralized routing the someone will take care.  There will be time to converge, it could be slower.
[00:58:23] <Alex Petrescu> HAC: all this bgp, and may not be ... a control plane talk to differennt switches.
[00:58:38] <Alex Petrescu> HAC: in virtualization, in bgp deployment, it will be everybody centralized control plane.
[00:58:45] <Alex Petrescu> HAC: not sure qwhat these commands are.
[00:58:55] <Alex Petrescu> HAC: once you have these commands then you realize in control plane.
[00:59:09] <Alex Petrescu> HAC: the issues you mention will be overcome because centralizesd is faster.
[00:59:15] <Alex Petrescu> CB: scalability issues will e tehre.
[00:59:28] <Alex Petrescu> CB: untill you provide figures with details, then this is not feasible.
[00:59:33] <Alex Petrescu> CB: it will take time, a lot.
[00:59:37] <Alex Petrescu> HAC: yes.
[00:59:50] <Alex Petrescu> HAC: I can give an analysis, it will take time but not that long.  I have some experiment I can show.
[00:59:59] <Alex Petrescu> HAC: instead of analysis, use this in SDN network.
[01:00:11] <Alex Petrescu> CB: we published paper where we analyze performance, a routing
[01:00:17] <Alex Petrescu> CB: a BGP, not centralized.
[01:00:23] <Alex Petrescu> CB: we also have BGP and PMIP
[01:00:29] <Alex Petrescu> HAC: Ill lookk at that.
[01:00:41] <Alex Petrescu> HAC: but SDN everyone talks about, and this will scope, will be actually faster.
[01:01:01] <Alex Petrescu> FT: we dont talk the global BGPO, but a private BGP instance in the private domain.  IT's not dumping bgp routes in the global.
[01:01:24] <Alex Petrescu> FT: it's for macro-mobility, not micro-mobiliyt so not subject to these issues
[01:01:31] <Alex Petrescu> Jurgen(?)
[01:01:48] <Alex Petrescu> J:I have same comment, can you ensure that the mechnaisms is going to be only utilized within this  network.
[01:02:09] <Alex Petrescu> J: Connexion for boeing, they had a long list of issues, when you fly with BGP for network mobility on internet scale.
[01:02:24] <Alex Petrescu> J: how do you prevent someone use a tool that is useful.
[01:02:30] <Alex Petrescu> J: this would be useful
[01:02:37] <Alex Petrescu> HAC: in the whole iInternet it will be impossible.
[01:02:44] <Alex Petrescu> HAC: this is only within the same datacenter.
[01:02:58] <Alex Petrescu> HAC:"move from one bs to another bs, then you dont need that.
[01:03:06] <Alex Petrescu> HAC: only when you move from one network to another.
[01:03:13] <Alex Petrescu> HAC: if move from one syubbnet to another.
[01:03:25] <Alex Petrescu> J: may include significant changes if moving from one subnet to another.
[01:03:31] <Alex Petrescu> HAC: just move the tunnel when...
[01:03:41] <Alex Petrescu> J: if they belong to same operator.
[01:03:57] <Alex Petrescu> SG: within a local domain?
[01:04:10] <Alex Petrescu> SG: you didnt say migrate prefix from West coast to east?
[01:04:17] <Alex Petrescu> SG: any consideration wbout this scope?
[01:04:25] <Alex Petrescu> HAC: edge router, you must stop this somewhere.
[01:04:36] <Alex Petrescu> HAC: not BGPO publicate to whole world.
[01:04:45] <Alex Petrescu> HAC: publicate only to the level.
[01:04:54] <Alex Petrescu> HAC: only up to some aggregation point.
[01:04:58] <Alex Petrescu> SG: how do you restrict that?
[01:05:08] <Alex Petrescu> SG: move from AR1 to AR100 to a completely different domain.
[01:05:14] <Alex Petrescu> HAC: another solution is that.
[01:05:23] <Alex Petrescu> HAC: you'd use a tunnel or some other mechnanimsm.
[01:05:31] <Alex Petrescu> HAC: we tried some PMIP draft but it was too complex.
[01:05:47] <Alex Petrescu> DL: you need to clear the scope of the BGP routing approach.
[01:06:07] <Alex Petrescu> ML: different aspects of that, the bgp discussion draft. Someone it scale because it applies to certain range of the network.
[01:06:16] <Alex Petrescu> ML: IP address continuity may use it.
[01:06:26] <Alex Petrescu> ML: how long it takes to propagate is a different story.
[01:06:48] <Alex Petrescu> ML: for that usecase we need to take care of migrating the context from par to nar, and the second is to adopt route, make packets, ...
[01:07:03] <Alex Petrescu> ML: is the use of bgp part of that solution?
[01:07:08] <Alex Petrescu> HAC: bgp is an example.
[01:07:20] <Alex Petrescu> HAC: dealing with what we have.
[01:07:35] <Alex Petrescu> ML: using bgp as an example is good, but binding it to solution.
[01:07:42] <Alex Petrescu> DL: I think many people are interested in this topic.
[01:07:54] <Alex Petrescu> DL: loiioooook for some opportunity make this draft more matuer.
[01:07:59] <Alex Petrescu> DL: align with others.
[01:08:08] <Alex Petrescu> DL: PDP signalling alsoi should be considered.
[01:08:16] <Alex Petrescu> HAC: we have a ref to another good job.
[01:08:46] <Alex Petrescu> Slide: Communiccating PRefix Cost to Mobile Hosts
[01:08:51] <Alex Petrescu> Slide: Updates from version 1
[01:09:17] KJ Sun joins the room
[01:09:34] <Alex Petrescu> J. Kaippallimallil from Huawi presenting
[01:09:38] <Alex Petrescu> (John)
[01:09:42] <Alex Petrescu> JK is John
[01:11:25] <Alex Petrescu> Slide: Introduction (1)
[01:12:58] <Alex Petrescu> Slide: Introductin (2)
[01:13:37] <Alex Petrescu> DM:
[01:13:52] <Alex Petrescu> DM: I undesrand the benefot of being aware of cost., but
[01:14:05] <Alex Petrescu> DM: not sure whether we'd want to inflect such knowledge to app developpers.
[01:14:18] <Alex Petrescu> DM: according to this noition, app devs who want to benefit from these things.
[01:14:25] <Alex Petrescu> JK: we dont want that complexity either
[01:15:00] <Alex Petrescu> DM: I suggest rather that, maybe you could twiset a bit, rather than providing info to app about cost, you could expect apps indicate what they';d want to do, and let the network figure it out.
[01:15:14] <Alex Petrescu> DM: example, download, x amount of gb, and network decides.
[01:15:32] <Alex Petrescu> DM: if app is real time app, jiutter, indicate that to socket, and comm it to network and network provides ressources.
[01:15:41] <Alex Petrescu> JK: this is an additional method, I will
[01:15:51] <Alex Petrescu> JK: not make app dev understand these complexities.
[01:16:01] <Alex Petrescu> JK: in 3gpp specs there is something about this.
[01:16:16] <Alex Petrescu> SG: we say this in not path characterization
[01:16:27] <Alex Petrescu> SG: but still we say there is something abstract which is the cost,
[01:16:39] <Alex Petrescu> SG: not tell exactly how is computed, then that's a problem.
[01:16:51] <Alex Petrescu> JK: next slides le'ts do, then we see, I take the q at that point.
[01:17:00] <Alex Petrescu> JK: I understand, and it is worth clarifying, but.
[01:17:09] <Alex Petrescu> Slide: Motivation (1)
[01:18:11] <Alex Petrescu> Slide: Motivation (2)
[01:18:53] <Alex Petrescu> SG:look at Jouni's draft, prefix as a home address, we indicate 'wrong prefix'.
[01:18:59] <Alex Petrescu> SG: local or remote prefix.
[01:19:04] <Alex Petrescu> JK: both are in the home network.
[01:19:12] <Alex Petrescu> JK: still end up as...
[01:19:21] <Alex Petrescu> SG: again goes back to the q of cost.
[01:19:26] <Alex Petrescu> JK: I agree with that.
[01:21:38] <Alex Petrescu> DL: running of of time, quick
[01:22:08] <Alex Petrescu> ML: work item 1, proposing to the UE the ability fixed/nomadic/sustained, is this a mechanism complementing?  Or alternative?
[01:22:22] <Alex Petrescu> ML:offer an IP address and the MN just assumes?
[01:22:39] <Alex Petrescu> JK: it's complimentary.  You can assume request for a nomadic address, and then cost changes as you move.
[01:22:49] <Alex Petrescu> Slide: Policy
[01:23:50] <Alex Petrescu> Slide: Prefix cost suboption
[01:23:58] <Alex Petrescu> Slide: IETF next steps.
[01:24:05] <Alex Petrescu> DL: next item.
[01:24:16] <Alex Petrescu> Marco Liebsch is ML is going to present.
[01:24:30] <Alex Petrescu> Slide: DMM Work Item: F...
[01:24:33] <Alex Petrescu> Slide: Outline
[01:24:39] <Alex Petrescu> Slide: Functional Architecture
[01:24:43] <Alex Petrescu> Slide: Current Status
[01:24:55] <Alex Petrescu> Slide scrolling
[01:25:12] <Alex Petrescu> Slide: Functional Architecture
[01:26:15] <Alex Petrescu> Slide Current Status/Mature information
[01:27:20] <Alex Petrescu> Slide: Current Status/Needs more discussion/details
[01:27:51] <Alex Petrescu> Slide: FPC Concept Discussions At a Glance
[01:29:26] <Alex Petrescu> Slide: Current analysis - Impact of operational aspects
[01:30:19] ling xu joins the room
[01:30:48] <Alex Petrescu> JK: commennting
[01:30:53] <Alex Petrescu> JK: I like the idea of abstraction
[01:31:02] <Alex Petrescu> JK: that work be one of the things I 'd like to see
[01:31:10] <Alex Petrescu> JK: the more abstractions related to...
[01:31:20] <Alex Petrescu> ML: abstractions could also apply to the right model, autonomous.
[01:31:32] <Alex Petrescu> ML: more control and more semantics required on that interface.
[01:31:45] <Alex Petrescu> ML: that's currently standardized, see whether there is overlap of same types of attributes.
[01:35:03] <Alex Petrescu> Slide: Feedback to the current draft
[01:35:39] <Alex Petrescu> Slide: Related work/Synergies with other SDOs
[01:36:24] <Alex Petrescu> Slide: Next steps
[01:37:14] <Alex Petrescu> Tienshun is TS
[01:37:29] <Alex Petrescu> TS: macro-cell, in the combination macro cell and ...
[01:37:44] <Alex Petrescu> TS: the two scenarios can be used in two cases: controller function.  Maybe then _and_ the right.
[01:37:53] <Alex Petrescu> ML: both models deployed at the same time?
[01:37:55] <Alex Petrescu> TS: yes.
[01:38:00] <Alex Petrescu> TS: the core network will combine.
[01:38:07] <Alex Petrescu> ML: you recommend this combination?
[01:38:16] <Alex Petrescu> TS: combine the deployments.
[01:38:19] <Alex Petrescu> ML: thanks.
[01:38:37] <Alex Petrescu> JK: mobility control function: includes session management?
[01:38:48] <Alex Petrescu> JK: context related to auth policy , or is this just mobility?
[01:38:56] <Alex Petrescu> ML: the mobility control function has session.
[01:39:06] <Alex Petrescu> ML: we want to not reflect that in the data plane.
[01:39:20] <Alex Petrescu> ML: in the mobile communication architecture, that is deployment specific sol.
[01:39:27] <Alex Petrescu> ML if the mcf has more features, then.
[01:39:41] <Alex Petrescu> ML: I'd hesitate to overload that interface.  IT should remain extensible.
[01:39:59] <Alex Petrescu> DL: you have additional slides?
[01:40:05] <Alex Petrescu> ML: next step slides is last
[01:40:09] <Alex Petrescu> DL: other comments
[01:40:17] <Alex Petrescu> Satoru from softbank
[01:40:24] <Alex Petrescu> S: good discussion we had.
[01:40:46] <Alex Petrescu> S: we have a discussion also on the problem forward centralized or distributed.  All should be covered in the bcp document.
[01:40:58] <Alex Petrescu> S: if we deploy that combined scenario we need to take care of scalability.
[01:41:07] <Alex Petrescu> S: the principal will affect the future of scalability.
[01:41:29] <Alex Petrescu> S: if only the interactive model it would scale, but difficult to migrate to future.
[01:41:42] <Alex Petrescu> S: makes react the limit between cps
[01:41:52] <Alex Petrescu> .S: because there is no tight messages.
[01:42:03] <Alex Petrescu> ML: both models are valid.
[01:42:15] <Alex Petrescu> ML: I apperciated the comments, and the others I received in private.
[01:42:29] <Alex Petrescu> ML: if you can support us by comments on the email list that would be appreciated.
[01:42:43] <Alex Petrescu> S: we already have a model that can be embpdyu the declarative model (decorative?)
[01:43:01] <Alex Petrescu> S: the terminology of dmm, in that fundamentally declarative model.
[01:43:04] <Alex Petrescu> S: do you catch?
[01:43:08] <Alex Petrescu> ML: not the last one.
[01:43:31] <Alex Petrescu> S: let us have a basement model.  on that declarative model.  IT would be very useful for dmm.
[01:43:38] <Alex Petrescu> ML: the declarative is a good base?
[01:43:54] <Alex Petrescu> S: yes, if we have such a base model then it is easy to make something on top of that.
[01:44:12] <Alex Petrescu> ML: in the interface attributes' wise it could be made more.
[01:44:21] <Alex Petrescu> ML: command-wise we may need new commands.
[01:44:32] <Alex Petrescu> S: on the specific attribute or parameter it is easy very easy business.
[01:44:40] <Alex Petrescu> S: but important is once the base of the model.
[01:44:44] <Alex Petrescu> DL: not enough time.
[01:44:53] <Alex Petrescu> SG: I agree with Satoru-san.
[01:45:08] <Alex Petrescu> SG: focus on standardization of attribtues?  Then any new operation we can come up with.
[01:45:20] <Alex Petrescu> DL: security considerations still need some work.
[01:45:24] <Alex Petrescu> ML: yes we do.
[01:46:00] <Alex Petrescu> Slide: Distributed mobilit management deployment scenario and architecture
[01:46:05] <Alex Petrescu> V. Liu is VL is presenting
[01:46:12] <Alex Petrescu> Slide: Update from -04
[01:46:40] <Alex Petrescu> Slide: NFV mScenario (Core Network)
[01:47:05] <Alex Petrescu> Slide: DMM in NFV system architecture
[01:49:16] <Alex Petrescu> Slide: Next Step
[01:49:39] <Alex Petrescu> Slide: DMM in NFV system architecture
[01:49:44] <Alex Petrescu> Slide: Next Step
[01:49:51] <Alex Petrescu> DL: Sri first q goes to you
[01:50:06] <Alex Petrescu> SG: I think probably post some calls, you can then, and we can review this and make feedbakc.
[01:50:11] <Alex Petrescu> DL: any other comments for this?
[01:50:26] <Alex Petrescu> DL: Mobile IP maintenance topic.
[01:50:33] <Alex Petrescu> DL: PMIP extensionsl
[01:50:41] <Alex Petrescu> Sri Gundavelli is SG is going to present.
[01:50:49] <Alex Petrescu> Slide: Motivation
[01:51:19] <Alex Petrescu> SLide: Community Wi-Fi Deployments
[01:54:26] <Alex Petrescu> Slide: LMA Comtrolled MAg Parameter Option
[01:54:41] <Alex Petrescu> Slide Binding de-registration control suboption
[01:54:58] <Alex Petrescu> Slide: Heartbeat Control sub-option
[01:55:01] <Alex Petrescu> Slide Conclusion
[01:55:50] <Alex Petrescu> JK: I read the draft it looks good
[01:56:00] HS YANG joins the room
[01:56:06] <Alex Petrescu> JK: miss syntax error.
[01:56:42] <Alex Petrescu> Slide: Conclusion
[01:57:16] <Alex Petrescu> DL: how many people read draft?
[01:57:18] <Alex Petrescu> DL: several.
[01:57:27] <Alex Petrescu> DL: how many people want this draft as basis for this work?
[01:57:30] <Alex Petrescu> DL: several.
[01:57:41] <Alex Petrescu> DL: we have some good basis, but we need to confirm adoption.
[01:58:05] <Alex Petrescu> DL: any update om the hnprenum draft?
[01:58:14] <Alex Petrescu> DL: the author is in the room.
[01:58:25] <Alex Petrescu> DL: just a short summary, what changes are from prior version.
[01:58:29] <Alex Petrescu> DL: you can just list.
[01:58:35] <Alex Petrescu> Slide: ...
[01:59:01] <Alex Petrescu> zhiwei is presenting
[01:59:11] <Alex Petrescu> zhiwei : we are here for one or couple of options.
[01:59:27] <Alex Petrescu> DL: how many people read the draft?
[01:59:30] <Alex Petrescu> DL: several
[01:59:47] <Alex Petrescu> DL: how many people think it is good as base solution?
[01:59:54] <Alex Petrescu> DL: confirm this good number on the mailing list.
[02:00:13] <Alex Petrescu> DL: next doc is the multihoming draft (draft-seite-rg-multihoming)
[02:00:16] <Alex Petrescu> SG: thanks Dapeng
[02:00:27] <Alex Petrescu> SG: we presented this at least 3-4 meetings.
[02:00:34] <Alex Petrescu> SG: earlier, document was a use-case.
[02:00:39] <Alex Petrescu> SG: what is the extension?
[02:00:45] <Alex Petrescu> SG: MAG register 2 CoAs.
[02:01:01] <Alex Petrescu> SG: currently PMIP only does one CoA, but this draft allows multiple.
[02:01:30] <Alex Petrescu> SG: BBF use-case induced some controversy ebcbecause hybrid access, then I talked to Chairs, to AD, so better not touch the use-case of banana.
[02:01:42] <Alex Petrescu> SG: we remoived the references to use-case of BBF.
[02:02:00] <Alex Petrescu> SG: the current revision reflectts that: no BBF ref, and just PMIP minor semantic update.
[02:02:12] <Alex Petrescu> X from Alcatel:
[02:02:19] <Alex Petrescu> X: but you mention DSL?
[02:02:24] <Alex Petrescu> SG: we removed,
[02:02:29] <Alex Petrescu> X: but still there.
[02:02:33] <Alex Petrescu> SG: I'll remove.
[02:02:42] <Alex Petrescu> X: remove ...
[02:02:50] <Alex Petrescu> X: also hybrid access?
[02:02:59] <Alex Petrescu> X: for a single PDN you also have 2 addresses?
[02:03:05] <Alex Petrescu> SG: in the PMIP context
[02:03:19] <Alex Petrescu> X: PDN connection - two IP address fro same PDN, or two PDN connections?
[02:03:26] <Alex Petrescu> SG: no PDN connection, just MAG.
[02:03:34] <Alex Petrescu> X: ;which link?  A link, a connection?
[02:03:38] <Alex Petrescu> SG: it is an interface
[02:03:44] <Alex Petrescu> SG: for same egress interface?
[02:03:47] <Alex Petrescu> X: same.
[02:03:57] <Alex Petrescu> X: once you talk LTE you talk hybrid access.
[02:04:01] <Alex Petrescu> SG: hybdir is bbf.
[02:04:08] <Alex Petrescu> X: you ohave that particular think we have to review
[02:04:15] <Alex Petrescu> SG: m-tcp group do the same.
[02:04:22] <Alex Petrescu> SG: is a problem we gave ref to hynrid access.
[02:04:40] <Alex Petrescu> X: nothing I have agains the approach.e It would be much better if you referred the hynrid access awork we do.
[02:04:49] <Alex Petrescu> SG: separate track, we dont want to object that.
[02:05:00] <Alex Petrescu> SG: we have a protocol, a semantinc missing, we want to document it.
[02:05:10] <Alex Petrescu> X: repsin this draft, and epxplain what scenario
[02:05:21] <Alex Petrescu> SG: what change?
[02:05:25] <Alex Petrescu> X: reference.
[02:05:31] <Alex Petrescu> SG: PMIP has no multiple CoAs.
[02:05:46] <Alex Petrescu> X: q is isimple: is it for single multi or single pdn connection?
[02:06:04] <Alex Petrescu> X: wifi, lte, cellular ... leads to hybdir access.
[02:06:06] HS YANG leaves the room
[02:06:15] HS YANG joins the room
[02:06:20] <Alex Petrescu> SG: already have the semantics for mip4, nemo, mip6, nemov6.
[02:06:28] <Alex Petrescu> X: not saying ...
[02:06:32] <Alex Petrescu> X: what is the sccenario?
[02:06:39] <Alex Petrescu> SG: related for MAG to use the multiple CoAs.
[02:06:48] <Alex Petrescu> X: what is
[02:07:08] <Alex Petrescu> DM:Sri says tis to be able to reflect a functionality mip provides to pmip.
[02:07:11] <Alex Petrescu> SG: yes.
[02:07:15] <Alex Petrescu> X: is the use-case there?
[02:07:17] <Alex Petrescu> DM: no use-case.
[02:07:25] <Alex Petrescu> DM: PMIP should be able to provide.
[02:07:33] <Alex Petrescu> X: there is agap is fine, there is no usecase to back it up.
[02:07:41] <Alex Petrescu> SG: every single protocol extension you dont want a usecase.
[02:07:49] <Alex Petrescu> SG: no need for business use case.
[02:07:58] <Alex Petrescu> X: there are many problems you can look at like that.
[02:08:15] <Alex Petrescu> SG: last comment, these comments are coming from only one thing.  There's hybrid access. We dont block that.
[02:08:21] <Alex Petrescu> X: you make that problem statement.
[02:08:41] <Alex Petrescu> SG: this work dont want to use it as problem for that work.
[02:08:45] <Alex Petrescu> DL: summary I make
[02:08:55] <Alex Petrescu> DL: we made a presentation the use-case.
[02:09:08] <Alex Petrescu> DL: the motiation is a simple extension.
[02:09:19] <Alex Petrescu> DL: is that extension... ?
[02:09:25] <Alex Petrescu> DL: that is how?
[02:09:31] <Alex Petrescu> Behcet Sarkiaya is BS
[02:09:39] <Alex Petrescu> BS: we discussed this in prefiovuos meetings.
[02:09:45] <Alex Petrescu> BS: this is not the mobility draft.
[02:09:52] <Alex Petrescu> BS: is this draft is out of scope?
[02:09:57] <Alex Petrescu> DL: this is extension to PMIP.
[02:10:06] <Alex Petrescu> BS: for a fixed problem, not for mobility issues.
[02:10:21] <Alex Petrescu> BS: this is artificial PMIP use of signalling to do somehting fixed network problem.
[02:10:30] <Alex Petrescu> BS: all the mobility procedures for PMIP...
[02:10:35] <Alex Petrescu> BS: this is out osc scope.
[02:10:42] <Alex Petrescu> DL: draft is itself to extend mobility protocol.
[02:10:46] <Alex Petrescu> Brian Haberman BH
[02:11:17] <Alex Petrescu> BH: DMM has responsibility for protocol maintenance of PMIP.  Does not mean that these changes have to do strictly for mobility.  It is protocol update and extensions.  
[02:11:29] <Alex Petrescu> BH: also, these are voluntary standards, if they want to use it they can.
[02:11:35] <Alex Petrescu> DLL out of time.
[02:11:58] <Alex Petrescu> DL: how many prople read the doc.
[02:12:09] <Alex Petrescu> DL: how many peoplke think it's good base.
[02:12:10] <Lyle Bertz> hummmmm
[02:12:23] <Alex Petrescu> Lyle: do you want that relayed?
[02:12:33] <Lyle Bertz> no need to
[02:12:35] <Alex Petrescu> DL: good number
[02:12:47] <Alex Petrescu> Lyle: the numbers were good, and thanks for the hum.
[02:13:22] <Alex Petrescu> Jong-Hyouk Lee is JHL is presenting
[02:13:27] <Alex Petrescu> Slide: Motivation (1/2)
[02:13:44] Hiro Inoshita joins the room
[02:14:13] <Alex Petrescu> Slide: Motivation (2/2)
[02:14:27] Hiro Inoshita leaves the room
[02:14:52] <Alex Petrescu> Slide: Proposal
[02:15:16] <Alex Petrescu> Sliude: NExt Steps
[02:16:19] <Alex Petrescu> Slide:
[02:16:28] <Alex Petrescu> Brian HAberman: not sure you mena DEPRECATED?
[02:16:50] <Alex Petrescu> JHL: in IP mobility, after handover, MN moved to new network and confiured new address.
[02:17:05] <Alex Petrescu> JHL: if the mobile wants to keep previous CoA/...
[02:17:11] <Alex Petrescu> JHL: in terms of Mobile IP.
[02:17:28] <Alex Petrescu> BH: you talk about the HoA or CoA.
[02:17:42] <Alex Petrescu> BH: maybe you want to talk to 6man if you want to do this RS/RA option.
[02:17:52] <Alex Petrescu> DL: PMIP maintenance, Charlie's document?
[02:18:09] <Alex Petrescu> Charlie Perkins is CP heads to presneter box.
[02:18:19] <Alex Petrescu> The draft is about mnids.
[02:18:23] <Alex Petrescu> Slide: Recent changes
[02:20:00] <Alex Petrescu> Slide: Recent issues raised
[02:21:24] <Alex Petrescu> DL: document is still in LC
[02:21:28] <Alex Petrescu> DL: you need some reviewers
[02:21:40] <Alex Petrescu> DL: sugest the members of WG to review this dcoument and send comments to the email list.
[02:21:46] <Alex Petrescu> DL: 9 minutes.
[02:21:57] <Alex Petrescu> Huii Deng is HD is going to present
[02:22:00] HS YANG leaves the room
[02:22:11] <Alex Petrescu> HD: Motivations
[02:22:21] <Alex Petrescu> HD: (usecase of community wifi)
[02:23:34] <Alex Petrescu> Slide: Modelsof VCPE Deployment--1
[02:24:04] <Alex Petrescu> Slide: Models of VCPE Deployment--2
[02:25:15] <Alex Petrescu> Slide: Models of VCPE Deployment--3
[02:26:07] <Alex Petrescu> Slide: VCPE deployment in the Community Wi-Fi Scenario
[02:26:54] <Alex Petrescu> Slide: VCPE deployment in the Community Wi-Fi Scenario
[02:26:58] <Alex Petrescu> Slide: Conclusion
[02:27:05] <Alex Petrescu> Slide: Conclusion and Needs for the WG
[02:27:56] <Alex Petrescu> DL: 3 minutes left.
[02:28:03] <Alex Petrescu> SG: this is valid
[02:28:12] <Alex Petrescu> SG: should be part of this document, if there is support.
[02:28:15] <Alex Petrescu> DL: other comments?
[02:28:21] <Alex Petrescu> DL : other topics ?
[02:28:23] <Alex Petrescu> DL: we done.
[02:28:26] <Alex Petrescu> DL: close, thanks.
[02:28:31] Tae You leaves the room
[02:28:34] <Alex Petrescu> HD: ask for support?
[02:29:02] KJ Sun leaves the room
[02:29:23] Lyle Bertz leaves the room
[02:29:26] Meetecho leaves the room
[02:35:55] Simon Romano leaves the room
[02:45:00] ling xu leaves the room
[02:51:15] Alex Petrescu leaves the room