Yoav Nir writes:
> But if it's different IP addresses for different applications, then
> there are probably also different traffic selectors for these
> applications, and then possibly different IPsec SAs.

Most likely, but it is still most likely going to be few -> many
mapping, i.e. you have few (1-3? or so IP-addresses), but you can have
many SAs each using few of those addresses.

I do not really see any need to use configuration payloads along with
create child SA exchanges. The reason we do configuration payloads in
the IKE_AUTH is to avoid extra round trips for the most common case
when creating IKE SA, getting IP address, and creating IPsec SA.

If some implementation needs more IP-addresses, then it can first use
INFORMATIONAL exchange with configuration payloads to get the
IP-addresses it needs, and then use create child SA to create the
IPsec SA using that address. For the GW it is enough to check that
traffic selectors for the source address is one of the addresses
allocated for the client (or otherwise address allowed by policy).

Draft-ietf-ispecme-ikev2-ipv6-config might need configuration payloads
in information exchanges also, especially if you use the sharing of
VPN access (section 3.4), as in that case you most likely need one new
perfix per interface where you share the address, and all of your
interfaces might not be up and in use when you first create the IKE
SA. 
-- 
kivi...@iki.fi
_______________________________________________
IPsec mailing list
IPsec@ietf.org
https://www.ietf.org/mailman/listinfo/ipsec

Reply via email to