Document status: ---------------- Will be starting WGLC on GRE keys soon. Will proceed to WGLC after final reveiew by the group. Another document ready for WGLC is the PMIPv6-IPv4 document. All issues resolved on the MIP6/PMIP6 interactions document. Need reviews of the document on the ML and will proceed to WGLC after that. IPv4 support in PMIPv6 (presented by Sri): ------------------------------------------ Sri: LMA needs to provide IPv6 address irrespective of the transort network. Vijay from jabber: Lot of scenarios dicussed in mailing list and needs discussion. Raj: There are issues about DHCP server being in the previous MAG. No need to have DHCP server funtion at the MAG. It is a deployment option and can be considered as a farm of servers. Suresh relaying for Julien: Need a common way of configuration for interoperability. Suresh: Need to represent normative text from the section. Currently, text w.r.t to DHCP is not Normative. Alper: Fair to mention the issues in the document about what type of DHCP should be used so that SDOs know how to configure the options. All this should be outside in appendix and not in main text. Jonne: Inter operability issues need to be specified. Ahmad: If this is not normative we can move the text to an appendix. Sri: Within an access system there cannot be any change to where the DHCP server should be placed. Alper: Cannot mandate that DHCP should not be allowed. Raj: Possible deplyment models can be mentioned in the appendix. Vijay: LMA figures out the mode of operation of DHCP and tells the MAG how to operate. Vidya: Cannot force a deployment to adopt a certain DHCP model. Hesham: cannot force DHCP models based on access technology type. Vidya: Two domains may have different DHCP models and the two LMAs may be connected to same MAG. What happens to MAGs that belong to two seperate domains? Julien: If we specify too much about the archtecture, then it should be removed. Jari: Interop is not really a config issue and it should be handled in the document. Jonne: Whenever there is something that can break interoperability, then it needs to be specified. Alper: Normative text mandating a certain DHCP model doesn’t make sense. Jonne: Need proper definition for interoperability. Alper: If there is no impact, then it is informative. Sri: thinks it should be informative. Hesham: If this is an Internet specification, then this needs to be fixed. Alper: Wimax forum is using the DHCP server at the MAG. Hesham: Need to specify how it works and where this does not work. Sri: Turn flag on and then turn the flag off. No need for any special options. Sri to update the spec removing any mandatory text. There will be a config optin, with the DHCP server co-located at LMA being the default mechanism. Vijay recommends that Sri posts text to ML before updating the draft. Consensus call for whether two options should be specified in the draft at all ended with no consensus (3 hands for yes and 3 hands for no). GRE key option for PMIP (presented by Ahmad): --------------------------------------------- Sri: Encapsulation mode will be useful. Frank: I think it is possible to use the GRE option without the Key. IPv4 private network need GRE keys. IPv6 GRE tunnels can be used without the Key. Suresh: Need to have GRE keys. I dont really see GRE being used in PMIP. Some body should tell why GRE should be used without the keys. Needs GRE keys even in IPv6. Sri: The option is generic option and it should be allowed without the key. Frank: GRE without the key is another encapsulation and it shoiuld be supported. Another person(?): GRE without the key should not be allowed. Ahmad: Another error code to inform whether GRE without key should be allowed or not. Jonne: Update the draft. Will issue LC soon. Suresh: I think the draft is in good shape. RADIUS support for PMIPv6 (Presentation by Frank Xia): ------------------------------------------------------ Alper: You are sending PMIP6 attributes for RADIUS using options used for MIP6-RADIUS and that should not be the way. Having distinct LMA and HA are possible and different options or attributes need to be used for the PMIPv6 parameters. That may be a decision to be taken later. Julien thinks it is a good comment to consider. Frank: RADIUS has some attribute shortage and the saving is worh it. Vidya: Needs consensus call to whether to adopt the document. Not enough people have read the document – please review and comment on the list. Lets take it up on the ML. In the meanwhile, we will get the RADEXT WG people to take note of the document and provide comments. PMIPv6 MIB (presented by Sri): ------------------------------ No comments and Sri stresses it is important to get comments from WG since this is important for PMIPv6 deployments. LMA discovery (presented by Raj): --------------------------------- Glen: Needs to provide IP address and not FQDN during handoff. Also, should the AAA be involved during hand-off scenarios? The current mechanism is context transferred between MAGs when AAA is not involved Raj: There is a context transfering mechanism agreed to. Frank: LMA discovery using DHCP could also be considered Mohana: During inter technology hand-off in some cases, the MN giving the LMA address is better because AAA may not have been updated from LMA. Control plane and data plane seperation (Presented by Raj): ----------------------------------------------------------- No comments. MPLS Tunnel Support in PMIPv6 (presented by Frank) -------------------------------------------------- Raj: mpls can be used; any need to change the protocol? Frank: how to distribute mpls labels? It is the same problem as gre key distribution. Vidya: out of scope of current charter