[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ih] vm vs. memory

    > From: Paul Vixie

    > every other attempt to add rapid renumbering and transparent
    > multihoming has been rejected. NAT, by not trying to do those things
    > and by not saying it would do those things, snuck under the defenses.

Bearing in mind that this is the Internet-_History_ list, and not the
Internet-_Architecture_ list, I will say just a little bit.

If you want your system to be able to do 'rapid re-numbering' and 'transparent
multi-homing' (and I'm not sure quite what you mean by that latter - having
two interfaces, and being able to sustain a connection from either one, is a
whole different kettle of fish from being able to use either interface for a
single connection, which is _much_ harder), you need to list those as
requirements when you sit down and do the basic architecure - including the
routing and addressing architecture (well, the whole naming architecture,
actually; of which addresses - topological names - are just one class).

IPv4 can't do either (those not having been goals when it was designed), and
since v6 is (basically) 'v4 with a few more bits', it's no surprise that it
can't do them either.

I've thought extensively about how to do 'hard multi-homing' (as I'll name
that particular flavour), and it's doable, but it needs to be built in up

As to 'rapid re-numbering', I've given that less thought, although I know (I
think!) how it has to be approached. Again, it has to be built in up front.

NAT may do these things (as a side effect; that wasn't its original goal when
Paul and Van came up with it), but NAT has the issue that it puts a lot of
state in the network, and does so invisibly.

This may or may not be a problem: in theory it should make the network more
'brittle' (by having state out there, state that you can't see, and thus
can't manage), but in practise it seems to work OK - actual problems seem to
be more likely to be caused by DNS issues, etc.