> > > It is an important discussion. There is a very critical architectural
> > > point here, and it is being glossed over by 'the sky is falling'  claims
> > > that apps might fail, or have to do some work.
> > 
> > It's become clear that you don't know what you're talking about.
> 
> On the contrary, I know exactly what I am talking about. 

No you don't.  You've never written a multiparty app that has to cope with the
conditions you are trying to impose on apps.  If you had, you wouldn't be
dismissing the extra complexity required to do this as "some work" and you
wouldn't be dismissing the problems as Chicken Little claims.

For better or worse, the current TCP/IP architecture uses IP addresses as 
both attachment names and endpoint names.  That might be unfortunate, but
you can't fix this by dismissing the use of IP addresses as endpoint names as
inappropriate, declaring that the network no longer has to provide a coherent,
uniform view of addressing, and pretending that any lack of coherency in the
network is now the apps' problem to solve.   Moving the problem from layer 3 
to layer 7 doesn't solve it; it makes it more difficult to solve.

If you want to really fix this you need to create real endpoint
names, provide a fast and reliable layer of indirection between these names
and addresses, and change the transport protocols to allow them to use those
endpoint names instead of addresses.

Until those changes are made and deployed, it will continue to be perfectly
reasonable for apps to pass IP addresses around, and it's the network's job to
maintain a coherent view of addressing for hosts to use.  Even when those
changes are made, much of the responsibility for maintaining that mapping
layer must still lie with the network.

--------------------------------------------------------------------
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]
--------------------------------------------------------------------

Reply via email to