I support this proposal,
Service Provider NAT as an example of a transition mechanism may need
to make use of IPv4 address space that is organisationally unique. I
do not agree that the use of Class E for IPv4 transition is
appropriate given the treatment of this as Martian addresses by most
equipment - this is limiting the usefulness.
While I believe the best approach is a NAT implementation in the
subscriber CPE that supports common address space on both inside and
outside (NAT != routing - RFC 4787 Section 4.4 is a must-read for
anyone considering SP-NAT) I accept that we need to support existing
CPE that simply cannot do this.
The current Class E draft recognises the need for transitional address
space, the problem is that Class E has been a no-go for so long it's
unlikely to be implemented within a useful timeframe - and especially
not for existing CPE.
-David Miles
* sig-policy: APNIC SIG on resource management
policy *
_______________________________________________
sig-policy mailing list
sig-policy at lists dot apnic dot net
http://mailman.apnic.net/mailman/listinfo/sig-policy