Re: [RRG] Consensus? End-user networks need their own portable address space

[jnc@mercury.lcs.mit.edu (Noel Chiappa)]

    > From: Brian E Carpenter <brian.e.carpenter@gmail.com>

    > This is where NAT goes wrong - it presents different tokens at the two
    > ends.

Yes, but that's an almost unavoidable consequence of the agglomeration of the
two main principles of NAT: i) no change to the hosts on either end, and ii)
multiple disjoint scopes for subsets (perhaps complete subsets) of the
address namespace.

I mean, if host X has address Ax in its local (source) scope, then unless
that host can somehow be assigned the same address in the foreign
(destination) scope, it is of necessity going to "present[] different tokens
at the two ends".

In other words, one can't take an N-bit namespace with local scope, and say
'this block is for local hosts, and foreign hosts get mapped (on demand) into
that block', and avoid hosts having different addresses at the two ends -
they have to have names out of different blocks of the namespace at the two
different ends.

(If one tries saying 'well, the local block at site P is this block, and that
other block at Q', and so on, so a host can have the same name (address) at
both ends, one has basically reinvented unique globally-scoped name...)


The only way to guarantee that hosts would have the same address at each end
would be to negotiation between the two ends, and find an address which is
free in both local scopes which both can then agree to assign to that host -
which probably breaks assumption i). And then if host X wants to talk to
hosts P and Q in different addressing scopes, one is almost for sure looking
at X having two different addresses at the same time - one to use when
talking to P, and the other for Q. Again, needing a change to the hosts.

If we'd been able to change the hosts, local naming scopes could have been
done a lot cleaner...

	Noel

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg