Date: Wed, 04 Jun 2003 15:51:31 -0700
From: Alain Durand <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| My point is that I'd rather like the default case to be that apps that
| require global addresses don't have anything special to do to work.
As it happens, the way I do things generally fits with what you'd
want I believe, but the way you expressed it isn't the way I would.
That is, really, almost no apps "require global addresses" - there are
a small handful in that category.
On the other hand, it is also true, that there are really only small
handful of apps that really want to use local addresses in preference
to global, when both would work.
The vast majority of apps simply don't give a damn - all that matters
is that some address that works now, and is likely to keep working for
the next few minutes, be picked. Given that (for some networks)
sometimes only the local address will work, and sometimes (for
different networks, or destinations) only the global address will
work, there isn't really any way that a system can choose one or the
other as better, for these apps (most of them) and be correct. It
needs to be a network setting (more in RA packets??)
My personal preference is that every node should have a global address,
always, and that those addresses should always work (for most purposes)
so for apps that don't care, in the work I have been doing, global
addresses are what always gets used. On the other hand, for apps that
do care, there is an API that allows them to get local addresses
whenever they work (the API keeps changing as we play with it more
and make different demands upon it).
This scheme works for me, but won't work for everyone.
kre
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
