Transition

Peter Curran pcurran@ticl.co.uk
Wed, 10 Feb 1999 19:50:53 +0000


Tim

Hope I can answer a couple of your questions (they are good ones, by the
way :-)


>I have been reading through the relevent RFCs on migration and deployment
>methods for IPv6. I was wondering in the future when IPv4 addresses have
>completely run out, how will it be possible for a small organisation only
>using IPv6 (because they are unable to be assigned any IPv4 addresses) to
>communicate with IPv4 only nodes.

Good question!  The simple answer is that you will not be able to talk to
v4 nodes without at least one v4 address.  There are a number of
translation strategies around that will work with a single v4 address for
most common requirements:
	NAT-PT
	Socks
	Application Layer Gateways

Suggest you browse the ngtrans section of the i-d directory at ietf.org for
details on some of these.


>Also, why should IPv6 over IPv4 tunneling be used as opposed to using dual
>stack nodes, using IPv4 when packets are required to be sent through IPv4
>routers or networks ? Is tunneling used to maintain functionality of IPv6
>packets when they reach the end node ?

You question above answers this, to a degree.  To use dual-stack across the
board implies that all hosts have to have a v4 and a v6 address.  The
purpose of the tunneling is largely to provide connectivity between islands
of v6 across the v4 internet (the 6bone being an excellent example of this).  

As addresses become scarser, the simplistic dual-stack approach will simply
not scale.  Some sort of translation device that can service multiple v6
users from a small number of v4 addresses will be essential.  A lot of work
is going on in this area, and there are now a number of different internet
drafts out suggesting different approaches.  Many of these overlap in terms
of function and operation.  The NGTRANS WG agreed, last week, in Grenoble
to try and come up with some sort of roadmap to identify the way these
mechanisms can be deployed and what circumstances are appropriate for the
use of each mechanism.....hopefully a reasonably comprehensive set of
guidance will be forthcoming soon.

Hope this helps.

Cheers

Peter


==============================================================
Peter Curran			pcurran@ticl.co.uk    http://www.ticl.co.uk
Consultant and Author
PGP key available from http://wwwkeys.uk.pgp.net:11371 or
ldap://certserver.pgp.com
PubKey Fingerprint =  5F94 D9A9 45EC 40A7 FB24  18BE 9C2E 74D6 E051 7F1F
===============================================================