[governance] IPv[4,6, 4/6] was IGF delhi format

Thomas Narten narten at us.ibm.com
Tue Feb 26 12:00:03 EST 2008


> i would still like to see a real strategy for co-existence of the two  
> addressing architectures that had an actual chance of wide scale  
> deployment and success.  this is for some definition of success that  
> includes the ability to connect all of the world's peoples to a single  
> global Internet, with all that means about end to end reachability.
> 
> after over a decade of IPv6 'inevitability,'  i still don't know  
> exactly what IPv6  transition means, but if it means that there will  
> be no more global usage of IPv4, then i don't expect this to happen in  
> my life time (and I am expecting to live for a while yet).

One of the oft-repeated myths that seems to continue making the rounds
is that IPv6 is about "transition" and that it is necessary to move
away from IPv4 ASAP.

Transition is a poor term, it turns out, because people associate the
term transition with "must stop using IPv4" and that raises all sorts
of alarm bells (and rightly so).

It has been an assumption from the very beginning that there would be
a very long coexistance period where IPv4 and IPv6 would both be in
use. Many years. Decades more likely. This is not new thinking. It is
not some recent realization that wasn't thought about from the
beginning. (Though it is true that people have argued forever just how
long a coexistance period would be.)

I can imagine data centers and other parts of an enterprise
effectively NEVER turning off IPv4. Why should they? It would only
make sense to turn off IPv4 if it is no longer working or
necessary. Think about legacy apps and cobol. They still exist. :-)
You don't change things that are working unless you have a compelling
reason to.  The same will be the case with IPv4 deployments.

> this is for some definition of success that includes the ability to
> connect all of the world's peoples to a single global Internet, with
> all that means about end to end reachability.

The reality is that we don't have that today with IPv4. We have a
world in which some parts of the internet reach some other parts of
the world, that is, where the set of destinations I can reach may be
very different than the set of destinations you can reach. This has to
do with routing and how the routing infrastructure actually works as a
business (e.g., due to policy considerations, there may be no route to
me (or you) in some parts of the Internet). It also has to do with the
widespread use of NATs/Firewalls, where many machines do not have
direct connectivity to other machines.

So, I don't think its entirely useful to talk about "a single global
Internet" except at a very high level. Having IPv6 and IPv4 coexist
will add strains to this (e.g., one particular IPv4 device might not
be able to communicate with another particular IPv6 device). But the
reasons for this will be varied and may be just fine. E.g., consider
email. Email works today because mail is relayed from one part of the
Internet to another, allowing sites that are not really directly
connected to communicate. This sort of thing will also work for
IPv4/IPv6. E.g., an IPv6-only site can relay mail to gateway that does
dual stack, which in turns relays to IPv4 destinations. This is
already done today, and will surely also happen in an IPv4/IPv6 world.

My point here is that having IPv4 and IPv6 does add some complications
to the ideal of a single global internet, but it's not a black and
white kind of thing. We don't even have such an internet today (if one
looks closely), though most people don't notice.

And when you think of IPv6 deployment, think "coexistance". IPv4 will
not go away anytime soon. Or even within our lifetimes, most
likely. That is perfectly OK.

Thomas

____________________________________________________________
You received this message as a subscriber on the list:
     governance at lists.cpsr.org
To be removed from the list, send any message to:
     governance-unsubscribe at lists.cpsr.org

For all list information and functions, see:
     http://lists.cpsr.org/lists/info/governance



More information about the Governance mailing list