Re: [DNA] DNA link hints document

[Bernard Aboba <aboba@internaut.com>]

> Actually from an IP layer perspective on an MN, the GPRS link equals to
> a ppp connection. Therefore activation of a PDP context is seen by an IP
> node as the establishment of a ppp connection (because 3GPP specifies
> that an IP node interfaces to the GPRS mobile station via PPP).

Actually, activation will express itself via completion of the PPP NCP
negotiation.  Data may be sent after NCP negotiation is complete, so that
is when the link can be considered "up" for IPv4.  For IPv6, I think that
isn't true; RS/RA also needs to be  completed to enable address
assignment.  Similarly, an LCP-terminate would bring the link "down" for
any network protocol.

> the dna process would want to get more information on the real
> underlying link than just ppp parameters. Am i wrong?

I see no reason why DNA should care about the link technology operating
under PPP.  All that is relevant is the PPP state machine and IPv6 RS/RA
protocol.

> Today specification of GPRS/UMTS, such a mechanim does not exist and
> PPP(IPCP and LCP) messages are the only available link-layer events. The
> problem being that such messages do not identify the underlying link.

Correct.