> >But if the binding cache on the local server is working well
> enough to
> >maintain the binding, and the mobile node sends a binding-update
> >proactively with its SYN, then the home agent won't be needed. This
> >should be the common case when the mobile node is initiating the
> >communication.
>
> A-ha, now I get it. Yes, that would work, subject to the
> correspondent's
> willingness and ability to store the binding, as you mentioned before.
> But I would expect (not knowing for sure because I haven't
> read the spec)
> that there is no obligation on the correspondent to store
> all, or even any,
> of the bindings it receives. Obviously, we hope most
> correspondents will
> do the helpful thing, but particlarly dumb, lazy, or overworked ones
> would be within their rights to ignore binding updates and always just
> send via the home agent.
>
> But I guess the point you were making was: if most
> correspondents do the
> right thing, there will be many fewer occasions when a user-override
> would be necessary. That I can agree with.
Exactly.
The situations where having the app use a "this is short-lived" socket
option make the most sense are when the correspondent's binding cache is
likely to be ineffective, like non-persistent HTTP connections or DNS (UDP
or TCP) queries.
Rich
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------