Re: [v6ops] PD-exclude was approved by the IESG -> should be included into 6204bis

[Anupam Kapoor <[e-mail]>]

 korhonen> Barbara,
 korhonen> On Feb 24, 2012, at 7:00 PM, STARK, BARBARA H wrote:

 >> I have a question as to whether this needs to be a MUST/SHALL or can it be a SHOULD.
 >> In my mind, the answer to this depends on how operators who intend to
 >> use PD-exclude will treat CE routers that do not support PD-exclude.
 >> Will they refuse to offer IA_PD (so the non-PD-exclude-supporting CE
 >> router is useless for attachment to that access network), or will they
 >> offer IA_PD and either also offer IA_NA (from a different prefix) or
 >> let the CE router operate in “unnumbered mode”?

 korhonen> I cannot speak on behalf of others but from 3GPP networks point
 korhonen> of view, unnumbered mode and/or using IA_NA are not options since
 korhonen> the system architecture does not support either of those.

,----
|  korhonen> If the CE router does not support pd-exclude and still, against
|  korhonen> 3GPP specifications, attempts prefix delegation the outcome is
|  korhonen> undefined. The 3GPP specs say today "The UE shall include
|  korhonen> OPTION_PD_EXCLUDE option code in an OPTION_ORO option.." etc.
`----
here are the relevant lines from sec:5.3.1.2.6 of 23.401 (ver:10.6.0)
      
      ,----
      | If the UE had indicated that it supports prefix exclusion and the
      | prefix to be delegated to the UE includes the /64 prefix that was
      | allocated to the PDN Connection, the PDN GW shall  
      | utilise the prefix exclusion feature as specified for DHCPv6 Prefix
      | Delegation in draft-ietf-dhc-pd-exclude-00 [70] 
      `----
from which it seems (to me at least) that ue *may* support
prefix-exclusion i.e. it is not mandatory for the ue to send
'PD-EXCLUDE' if supports prefix-delegation.

kind regards
anupam