Hi Sri, On 12/13/11 3:29 PM, "ext Sri Gundavelli" <[email protected]> wrote:
>Hi Raj: > >Please see inline. > > >On 12/13/11 1:12 PM, "[email protected]" ><[email protected]> >wrote: > >> >> A few clarifying questions: >> >> 1. Would all the MAGs across different PMIP6 domains be required to use >> the same LLA and IID? >> As per the proposal a single LLA and IID are being reserved for use by >> PMIP6 MAGs. >> > >We already have this requirement of using a fixed link-layer Id and IPv6 >interface id on the MAG-AR interface, within a PMIPv6 domain. We did not >do >the allocation of the same values and deployments have to configure these >values across every MAG in a PMIPv6 domain. Given the scope is a localized >domain, we can only specify this requirement for a given domain, along >with >the other domain-wide parameters. Generally, requiring the configuration >of >a domain wide value is not an issue, but when the values are about >link-layer id, we need a standardized value, and in the absence of that, >administrator will configure some random addresses, as we are noticing, >which can be an issue. Agree. RFC5213 requires the MAGs in a PMIP6 domain be configured with the same LLA. Sec 9.3 of RFC5213 defines the parameter: "FixedMAGLinkLayerAddressOnAllAccessLinks". It would be good to explain that this I-D obsoletes this parameter or indicate the IANA assigned value to be now provisioned for the same. Do you see any concern from security or privacy perspective the use of a common LLA and IID for all MAGs across multiple PMIP6 domains? > > > >> 2. Is it mandatory for the MAGs in a PMIP6 domain to use this LLA and >>IID? >> Or is it simply a recommendation. >> Network configuration tools and protocols can ensure that the same LLA >>and >> IID is configured across all MAGs in a PMIP6 domain. Hence assigning a >> specific LLA and IID is unnecessary. >> > >Yes. With this draft, its requiring these assigned values to be used on >the >MN-AR interface, eliminating the need for static CLI-based configuration >across all the MAG's in a domain. Okay. But it is not just w.r.t all MAGs in a single domain; It applies to all MAGs across domains, right? > > >> 3. When a MAG is provisioned it needs to be associated with an LMA and a >> security association configured between the MAG and LMA. As a part of >>that >> process, the MAG could obtain the LLA and IID to be used from the LMA >> itself. Would this approach not be sufficient? >> > >Its not really tied to the home network. Its more about MAG to MAG >roaming, >within a PMIP domain. We required statically configured values in the base >spec, the same is now an IANA allocated fixed value. Sure, it is possible >to >extend the message interface and push it from the LMA as well, if we want >to >define new option for MAG-LLID on PMIP interface, but allocating fixed >values for the already defined domain-wide parameters is the simplest >option. Okay. I agree it is easier to have a static configuration value than to extend the protocol between MAG-LMA with an additional option. -Raj > > >Regards >Sri > > > > >> -Basavaraj >> >> On 12/13/11 2:50 AM, "ext Jari Arkko" <[email protected]> wrote: >> >>> The new version of this draft looks good to me: >>> >>>http://tools.ietf.org/rfcdiff?url2=draft-gundavelli-v6ops-pmipv6-address >>>-r >>> eservations-04.txt >>> >>> Ready to be approved? >>> >>> Jari >>> >>> -------------------------------------------------------------------- >>> IETF IPv6 working group mailing list >>> [email protected] >>> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 >>> -------------------------------------------------------------------- >> >> -------------------------------------------------------------------- >> IETF IPv6 working group mailing list >> [email protected] >> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 >> -------------------------------------------------------------------- > -------------------------------------------------------------------- IETF IPv6 working group mailing list [email protected] Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
