RE: 3gpp-analysis-07: (semi-)editorial issues

["Karim El-Malki (HF/EAB)" <karim.el-malki@ericsson.com>]

 > On Mon, 17 Nov 2003, Karim El-Malki (HF/EAB) wrote:
 > > OK I see. Reread the original now. I think the meaning of the
 > > "peer address" part of the sentence was along the lines of: if you
 > > only have a v4 pdp context and you discover that the peer 
 > is v6-only
 > > then set up a v6 pdp ctxt. Is the issue that this was not clear or
 > > that this does not make sense?
 > 
 > Isn't it already way too late to activate the v6 PDP context 
 > if you wait 
 > until looking up the addresses, because PDP context 
 > activation takes a lot 
 > of time?

As you say it would be better if the pdp context was already there
but in the above case you have little other choice? (also some
other considerations on "in advance" activation below)

 > 
 > And actually, if an application gets e.g. both AAAA and A 
 > records back
 > from the DNS, you don't know which will end up being used.

True, if both are returned then we don't know since it's up
to the app. I was taking the case where there is only a AAAA
record and you only have a v4 pdp ctxt.

 > 
 > The point of the reword was to state that activating based 
 > on the peer 
 > address basis doesn't usually make sense, and 
 > application-based activation 
 > may not either..

You could activate pdp contexts in advance or you may do
that based on app.s and in special cases (above) peer
addresses.  The "in advance" option is attractive but there
are cases where it doesn't work well: if the application needs
a pdp context with certain QoS and APN (e.g. corporate)
characteristics that is not currently active. So I am not
convinced that we should be making the above point.

/Karim