Hi Suresh,
On 9/21/11 3:32 PM, "Suresh Krishnan" <suresh.krish...@ericsson.com> wrote: > Hi Sri, > > On 11-09-19 01:29 PM, Sri Gundavelli wrote: >> Hi Jari: >> >> In case of PMIPv6, we need the interface ID allocation for PMIv6 >> domain-wide usage. We may not be able tie this to a specific EUI-64 >> identifier derived from a MAC identifier of any individual MAG hosting this >> configuration. >> >> But, if your recommendation is to tie the IPv6 interface identifier to the >> reserved link-layer identifier (the other IANA action), that should be fine. >> But, the reserved block that Suresh has in his spec is not tied to any >> EUI-64 block ? That implies, we need an interface ID allocation from a new >> block ? Or, recommend the node to generate the interface ID based on the >> reserved Mac address and not allocate from the block Suresh created ? > > There are two ways by which you can get a stable reserved IID. One by > reserving an IID block using an IANA registry and the other by getting a > stable MAC48 that can be used to create an IID. The only danger with > getting a stable MAC48 is that if you happen to have 2 MAGs in the same > broadcast domain, it will lead to issues. > Two MAG's in the same broadcast domain can be ruled out, as there is no mechanism for the nodes to decide on who should support the attached MN. More over, we still have the P2P link assumption and so multiple MAG's in the same broadcast domain is not possible. So, we are fine on this aspect. Stepping back a bit, IMO, 5453 issue is not applicable for the P2P link and given that its a managed mobility domain. So, I wonder, if the concern is valid in PMIPv6 context ? Regards Sri > Thanks > Suresh -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
