Keith Moore <[EMAIL PROTECTED]> wrote:
|On Tue, 14 Oct 2003 18:11:24 -0400 (EDT)
|Dan Lanciani <[EMAIL PROTECTED]> wrote:
|
|> IPv6 offers absolutely nothing as a
|> replacement for NAT's primary function: isolation of the customer network
|> from the (typically business-driven) address policies of the service
|> provider(e.g., cost, limits on number and stability, etc.).
|
|NAT isolates customer networks from upstream address policies only by severely
|limiting their functionality.
Clearly many users care a lot about the isolation and little about the
functionality that you believe is being limited. Rather than trying to
convince them that they are wrong for wanting to keep their networks
running, how about proposing a way to achieve that isolation without
limiting the functionality that you want to preserve? If we cannot
implement such a solution (and I doubt that we can since it changes the
economics of address rentals), do you really think it is reasonable to
demand that everyone give up what they need to have what you think they
should want?
|The functionality of an IPv6 network is better
|than that of a NATted IPv4 network even if the IPv6 address changes
|frequently.
To paraphrase your comments on host-based source routing, the functionality of
an IPv6 network whose addresses are changing frequently enough to make its
users' applications unusable is nil. Conjecture: ``frequently enough'' likely
isn't all that frequently. We just don't have the tools and models for dealing
with addresses changing out from under applications. Even if we did, I've yet
to see any commitment from apps folks that they would be willing to deal with
such changing addresses. And I think that dealing with changing addresses is
a much harder problem for an application than dealing with scoped addresses.
It's also something that will have to be handled by all applications, not just
those using referals.
|However I agree that we need to define some expectations about address
|stability - both to serve as models for service agreements between customers
|and ISPs, and to serve as guidance for applications developers.
Ultimately, I believe that stability is a binary property. Either someone
other than the owner of the network can change the address or they can't.
Anything that depends on the provider (and its provider and so on) to neither
screw up nor decide to change terms for economic reasons is not stability.
Dan Lanciani
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
