Network-based Local Mobility Management (NetLMM) ------------------------------------------------ Monday, December 3rd, 2007 1300-1500 Afternoon Session I Chairs: Jonne Soininen Vidya Narayanan ------------------------------------------------- Meeting Minutes (thanks to Vijay, Gerardo): PRELIMINARIES Jonne went over the agenda The MN-AR draft has expired but will be updated during the week or right after the meeting. WG DRAFT DISCUSSIONS Proxy Mobile IPv6 open items after WGLC (Sri Gundavelli) --------------------------------------------------------- Sri presented some of the open issues on the base PMIPv6 draft Policy profile issue: Gerardo - There are many places where the policy profile is used to refer to the protocol operation. We are introducing a new entity called the "policy profile". Suggestion is to remove references to the policy profile and say whereever you need some configuratoin delivered to the MAG, it could be out of scope. Sri - Then what do we say about how the the MAG obtains configuration parameters Gerardo - Could be configuration on the MAG Ahmad - Disagrees with Gerardo. We had consensus on this issue, shouldn't be opened again. Hesham - Can we just call it "configuration"? Sri - This is already mentioned -It was decided to improve the definition of the policy profile. Multihoming issue: Gerardo - A comment about the multi-homing section. I really think we should have a single section for all processing rules Ahmad - support how it is now in the draft; some people do not wish to implement multihoming Sri - not an intention, we changed the rules; multihoming is not optional George: multihoming is not optional. Changes due to multihoming are applicable to all nodes. point is this comment made is that for an implementation point of virew it would be good to have one place where you implement BU; spreading across document is harder MNID: Alper - Why does the HNP assignment have to be in-band? Why does the MN-ID have to be there in the Proxy BU always.. Jari - Would like the MNID to be a stable identifier. George - Problem is that the protocol uses a lot of external parameters and profiles, like the the policy profile. You cannot use PMIP as standalone -Long discussion on this.... No conclusion. IPv4 support for PMIPv6 (Ryuji Wakikawa) ---------------------------------------- Hesham - How will the scenario 3 work (where DHCP server is located somewhere in the PMIPv6 domain with the relay on the MAG only) Ryuji - The DHCP exchange will be followed by the PBU/PBAck exchange. The MAG will provide the hint for the IPv4 home address in the PBU -Document needs more review by the WG ADDITIONAL PRESENTATIONS ------------------------ (Individual drafts that have had discussions on the list) PMIPv6 - MIPv6 interactions (Gerardo Giaretta) ---------------------------------------------- Jari - Since there are so many issues with scenarios B and C, do we need to support them? Gerardo - There are folks who want to use Scenario C Hesham - I have also always questioned the need for supporting Scenario C Vijay - There are some folks wanting to use scenario C. It is better if we actually document the issues and show some solutions based on some assumptions rather than leaving it to the SDOs to implement this on their own Gerardo - 3GPP uses scenario C. Jari - Agrees with documenting the issues GRE for PMIPv6 (Ahmad Muhanna) ------------------------------ Hui Deng - How do you secure GRE tunnels? Sri - IPsec can be used. Hui Deng - Any Security reasons for GRE tunneling? Sri - No, just flow separation (New problem statements) Problem Statement and Requirements for Route Optimization (Christian Vogt) -------------------------------------------------------------------------- Sri - The prefix is anchored on the LMA, how can you do RO Christian - The MAGs would talk to each other directly without going through the LMA Hesham - we need to think about: proxy RO? Is this about local or global mobility? Hesham - The interactions with a MIPv6 RO capable MN and CN should also be considered Christian - Agrees Multiple Interface Nodes Problem Statement (Mohana Jeyatharan) --------------------------------------------------------------- No discussion noted. PMIP6 and NEMO scenario and problem statement (John Zhao) --------------------------------------------------------- No discussion noted. (Other drafts - as many on the list as time permits) Multihoming Scenarios in NETLMM (Vijay Devarapalli) ---------------------------------------------------- No discussion noted. Inter access technology handovers (Marco Liebsch) ------------------------------------------------- Vijay - Why is the "preliminary PBU" required only for inter-access technology handovers and not for a regular PMIPv6 handover? Marco - It might take longer to bring up the new interface in an inter-access technology handover. In an intra-access technology handover, the interface is already configured with the necessary information Sri - Likes the idea Re-chartering discussion ------------------------ Hesham - Hard to discuss rechartering without knowing the topics. It makes no sense to just say we need to recharter Hesham - Need to discuss scope of PMIPv6 as a global mobility management protocol Gerardo - previous slide was about principles, next slide is what is out of scope; one could implement something like this with the current principles Jim - Thinks we should work towards making Mobile IPv6 happen end-to-end Sri - The statement about no changes to the host has been misinterpreted. It should have meant no mobility signalling from the mobile node. Not a layer 2 hint. Vijay - A couple of points. We do need to clarify what it means by no host changes. It should be interpreted as not requiring a the mobile node to do any mobility related signalling. It should not mean that the mobile node cannot provide hints or other information at the handover. The other point is that NETLMM is being used for mobility within a mobile operator whose network could cover a huge country for example. So it is no longer a local mobility management protocol. We should recognize this, instead of insisting that PMIPv6 should only be used for local mobility Jari - Hosts could be changed for optimization. However, the protocol needs to work also with unchanged hosts. Gerardo - Disagrees with calling PMIPv6 as anything other than a local mobiliy management protocol Gerardo - We need to be sure to keep the principles. The current charter has some principles and some items out of scope. We can include now the items which were considered out of scope but we should not change the principles Sri - I think the principles can be changed. The MN involvement can be made clear Jim - this group does not make sense and should not be rechartered Jonne - Principles are up to discussion Gerardo - They are not, let's stick to the principles of the charter for which this group was created Jari - I think there is interest for rechartering Jim - We need to work on improving MIPv6 and making that happen instead of all this stuff George - We are going against the principles. All customers are doing global mobility. IAB should be involved Gerardo - Prioritize what is needed to deploy the protocol: AAA support, LMA discovery, etc. Vijay - PMIP and MIP interactions should be included Rechartering discussions will be taken up on the mailing list after the current WG drafts are sent to the IESG. Meeting adjourned.