Re: about Wedgelayer 3.5 / Fat IP approaches
marcelo bagnulo braun <[email protected]>
| Newsgroups | gmane.ietf.multi6 |
|---|---|
| Message-ID | <[email protected]> |
>> It is actually more constrained than that since other instances of >> the application might be running on non-multi6 aware hosts. >> So the conservative approach would be to always hand an IP >> address/locator >> to unmodified applications, that way the application can use it >> for callbacks and referrals without any constraints on which hosts the >> different instances of the app is running. > > For a transition scenario this will most likely more be the rule than > the exception. SIP comes to mind of an application where you would do > referrals and you can not trust that all "nodes" are of the same > "instance"/protocol version/awareness. > Does this implies that AID have to be routable IP addresses? regards, marcelo > - - kurtis - > > -----BEGIN PGP SIGNATURE----- > Version: PGP 8.0.3 > > iQA/AwUBQOPFuKarNKXTPFCVEQKT0QCg/eOifwbywPqBgcwcl0Kw2A5QZj8AoJFh > Dk+jyIfwZcadamZqSRMvpF4S > =y2oT > -----END PGP SIGNATURE----- >