Iljitsch,
On Sep 13, 2007, at 3:01 PM, Iljitsch van Beijnum wrote:
Disconnect current session, reconnect.
Uh, not unless your application has some sort of retry or
checkpoint-restart capability.
I think the way IPv6 handles this by deprecating the current
address but keeping it alive for some time is entirely reasonable.
There is no real difference between how IPv4 handles this and how
IPv6 handles it.
Applications, regardless of whether they are IPv4 or IPv6, know the
address of the remote side of the conversation, it is cached in
application data space. If the address changes, the application
needs to be made aware of that somehow. This is usually done by a
connection reset, often causing the program to terminate. Unless
you're going to rewrite pretty much every Internet aware application
on the planet, IP addresses (both IPv4 and IPv6) are assumed to be
stable.
Assuming that you can keep a session alive for weeks or months
without the ability to recover from disconnects is not a smart
idea, to say the least.
Is it smart? Nope. However, most applications were designed with
this assumption and it works today if you have PI.
Like I said, EVERYTHING is a non-starter today so we don't start
anything anymore. The assumption that everything that works today
will continue to work forever is broken.
Yep. Question is, what's critical vs. important vs. nice-to-have?
(I consider long term connections to be a nice-to-have, btw)
Regards,
-drc
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf