MIP4 WG Minutes - Wednesday, November 19th, 2008 Chairs: Henrik Levkowetz and Pete MaCann Scribe: Acee Lindem - acee@redback.com Agenda Bashing - Pete McCann Document Status - Pete McCann - RFC 3344 Bis - Revision posted. Need through document review before sending to ADs - Generic Notification - Hui Deng - Changes * Timstamp based Replay Protection Added (Auth-Ext now mandatory) * Result Codes Added * Binding Revocation removed since it has it own message. Pete - Need review by WG - Nemo FA Draft will be dropped due to lack of interest - Nemo4 Dynamic - Will go to WG Last Call * Pete advised to look at the IPR statement. - MIP4 Gen Ext - Need discussion - MIB Document - Latest revision covers comments. - DSMIPv4 - In IETF last coll. Comments received and new revision published. Service Selection - Antti Makela (for Jouni Korhonen) See Slides - First presented in Praque. Requirement is for service ID beyond user NAI. - Facilitates multiple services with independent policies - Index added in addition to service ID. 0 indicates index is not applicable. - Request to WG document in chartering discussion Jari Arkko - Is this exactly the same as the MIPv6 equivalent (RFC 5149)? Antti - I hope so :^) Jari - Is this needed for 3GPP R8? If so, this should be an AD sponsored document. Antti - Yes - 3GPP TS 24.304 requires it. See document for use cases. Sri - Is this referenced in 3GPP? Antti - Yes. Jari - Is this on 3GPP dependency list. Antti - Yes. Mobile Agent Discovery Proxy (MADP) in IPv4 - Telemaco Melia - Applies to xDSL networks for WiFi extensions - MADP acts as a proxy to limit Bandwidth used by Agent Advertisements from FA/HA - MADP periodically transmit AAs to MNs - MADP can work off cached information or statically provisioned information - Use case puts MADP in DSLAM2 Alper Yegin - What is the protocol work? Telemaco - Caching and sending AAs. Henrik - Why are AAs received on downlink dropped? Telemaco - Looping prevention. Ahmad - Why doesn't MN send AS? Telemaco - Unsolicited AA helps MN know it has moved sooner. Telemaco - Send questions to the authors. Re-Chartering Discussion - Service Selection draft-korhone-mip4-service.txt - Multiple CoAs - Multicast/Broadcast Delivery Style draft-chakrabarti-mip4-mcbc-03.txt - Make-before-break (handover between PMIP and CMIP) draft-mccann-mip4-mbb-00.txt - Antti - Home Assisted Route Optimizations - Antti - Hui Deng - Proposes Host Configuration draft as part multicast/broaadcast work. - Henrik - Drafts should not be combined. - Ahmad Muhanna - Host configuration draft can be used as an example of multicast/broadcast but as a separate draft. - Behcet Sarikaya - BOF on multicast mobility may be taking on the work of multicast/broadcast and mobility. - Henrik - Asserts that this is a correction to the MIP4 base protocol that needs to be fixed. - Behcet - Opposes unicast WG doing multicast work. - Jari - Doesn't feel MIP4 charter precludes this work. MULTIMOB needs work before it can be chartered as WG. - Ahmad - RFC 3024 does this in a non-optimized manner. This draft just corrects that situation. Host Configuration Draft - Hui Deng - Has not refreshed draft - One question - Should HA be considered as DHCP server or should this be removed? Pete - Not discussed enough on list. Henrik - Expressed preference for removing DHCP from HA. However, there may be other opinions. Hui - Will remove it. Pete - Take it to the list. Multiple CoA Support - Sri Gundavelli. - Already supported in Cisco implementation - Multiple Access Interfaces on MN - Support already in MIP6 - MN can use different links for different traffic types - Can be via different FAs or directly attached MN - Need specificaiton in RRQ of link type and bandwidth - Out-of-Band policies can be apply based on this information Antti - Supports this work. Ahmad - Are there nodes behind MN? Sri - Yes - there could be. Ahmad - Is it the same home address? Sri - Yes. There is one home address with reachabilty over multiple paths. Policy is used to select the downlink path in the HA and uplink path in the MN Henrik - Thinks it is good. Do you have different service types for the different links? Sri - We use link types rather than service types. Henrik - Think a different identifier is needed. Alpher - Is the policy signalling in or out-of-band? Sri - This is out-of-band? Alpher - How would you get interoperability? Sri - Thinks policy should stay out-of-band. Charlie Perkins - Think using the link type for determining the behavior is wrong. Jari - Link type bothers him as well. Tunneling is the easy part and policy is the hard part. Leaving it out may not be unacceptable. Sri - But can tunneling be decoupled from policy? Jari - Yes - but the draft may not be useful without it. Sri - Even without policy, we get load balancing out of this. Jari - How will this be done? Sri - Stacks will do it today with flows spread over the available paths. Henrik - Pretty much the same point as Jari about policy. Thinks policy can be decoupled but should not be ignored. Pete - Any further comments? If not, meeting adjourned.