RE: [DNA] Comments on draft-shim-dna-proactive-00.txt

[Zhigao Chen <zgchen@psl.com.sg>]

> > You are right. IPv6CP hint is not as strong as its counterpart in 
> > PPPv4.
> 
> Since it doesn't provide the prefix, it may not even qualify 
> as a "hint" at all.  For example if PPP authentication 
> provides no indication of the network to which the peer has 
> attached, I'm not sure on what basis the PPP peer could even 
> guess the prefix.
> 
In 3GPP/GPRS, AT commands and PPPv6 are used in between TE and MT to
support IPv6 packet transmission. My understanding is that MT sends the
IPv6CP Config Ack message to TE in response to the receipt of the
Activate PDP Context Accept message from GPRS network. The IPv6CP Config
Ack message contains a NSAPI, which identifies a PDP context, in
addition to PDP address (link-local IPv6 address formed from IID).
Consider later on the PDP context is going to be assigned a prefix. The
NSAPI could be a useful auxiliary information of the link-up trigger if
the prefix allocation by GGSN is on a one-prefix-mapped-to-one-NSAPI
basis. In that sense, IPv6CP might be considered as a hint, so to speak.

I am not sure whether IPv6CP options have been defined to carry NSAPI
and authentication information, which includes the network  infor of the
peer. It seems to me in RFC 2472 IPv6CP currently uses message types
(Config-Ack/Nak/Reject) to tell TE the authentication results only.

Regards,
Zhigao