I'll combine two answers in message..

[me:]
> > Why exactly should we care if party X's internal applications break
> > because it hijacks a prefix?

On Mon, 3 Nov 2003, Hans Kruse wrote:
> We don't, and that is my point.  The draft in question improves on that 
> situation by creating a prefix that the rest of the network can easily deal 
> with.  Internal apps may still break, although I would argue that the local 
> addressing prefix opens some options to make that a little less likely...

No, the point is that when someone hijacks a prefix, they intentionally do 
something they know they should not do, and they "deserve" their 
applications to get broken.

If we specify a mechanism for local addressing that "just about works",
but still breaks apps, we've "blessed" a mechanism that doesn't work.  
That's even worse :-)

[Fred:]
> Because simply allowing internal apps to break is in clear violation of the
> robustness principle. (e.g., if two disconnected/intermittently-connected
> sites that have somehow hijacked the same prefix encounter one another
> we have what Data would call: "a simple matter/anti-matter reaction".)

Yes, but the site who hijacked a prefix is lost beyond redemption.  Again, 
why should we care if their apps break?  (We certainly care about apps not 
breaking in "valid" deployment scenarios..)

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings


--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to