[ppml] 2005-1:Multi-national Business Enablement

Stephen Sprunk stephen at sprunk.org
Thu Apr 28 07:56:04 EDT 2005


Thus spake "Daniel Roesen" <dr at cluenet.de>
> On Sun, Apr 24, 2005 at 09:13:48AM -0500, Stephen Sprunk wrote:
> > > Nope. They should get a /48 unless they can convincingly show that
> > > they'll never need more than a single subnet.
> >
> > It is ridiculous to think that ISPs are going to completely discard
their
> > current IPv4 topology to deploy IPv6.
>
> Why must they discard any topology?

IPv6 mandates a particular topology and disallows others which happen to be
in widespread use by IPv4 ISPs.  If a provider using the latter were to
offer IPv6, they would need to change their IPv4 topology as well, and I
have trouble seeing ARIN approving an ISP quadrupling (or more) their IPv4
requirements so they'd have an IPv6-compliant topology.

> They might have problems to charge for /48 instead of /64 if /48
> becomes "usual" at the competitors, yeah. But that's primarily a
> business thing.

Not what I was talking about.

> > Most "residential" ISPs I'm aware of use a single subnet for N
customers,
>
> Hm? I guess you are referring to cable modem stuff?

It's common for cable, DSL, wireless, and other technologies.  For instance,
my landlord provides a straight ethernet connection into my residence (which
is connected to a T1); with DHCP, I consume only one IP per PC.  For them to
offer me an IPv6 /48 or even /64, they'd need to change their IPv4
addressing to a /30 or shorter for each customer, wasting four addresses for
a customer with one PC.

> Over here (DE), almost all residential users use dial-up, be it real
> (analog, ISDN) or virtual (DSL, via PPPoE). So they are connected via
> virtual interfaces, and get their IP address usually via dynamic pools
> or static via RADIUS. No problem adapting this to assign /48s
> (especially via RADIUS).

If that's the topology, then that makes sense.  However, it's not the
dominant topology in the US today.

> > and that is perfectly reasonable to continue with IPv6 -- just assign
> > a /64. That allows customers to put as many hosts as they want on the
> > "bare" connection, and any who want their own subnet(s) can be issued
> > a /64 or /48 of their own
>
> And then charged additional install and monthly fee? Cool. We want to
> get rid of this as much as possible. Artificial scarcity sucks. In IPv4,
> there is the perceived scarcity excuse for ISPs. Don't recreate that in
> IPv6, thanks.

Compare to the artificial scarcity of IPv6 addresses created by (a) no PI
space being available, and (b) not allowing assignment of IPv6 addreses with
existing topologies.

Frankly, it costs less to deliver a single /64 to N customers than to
deliver N /64s or /48s to N customers.  If there's a price difference, so be
it -- as long as both are available.  Choice is good.

> The mantra is "/48, no questions asked, and by default".

When you consider how that affects the IPv4 topology, that doesn't make
sense in many cases.  If we're going to share subnets across customers in
IPv4, we need to do the same for IPv6.

S

Stephen Sprunk        "Stupid people surround themselves with smart
CCIE #3723           people.  Smart people surround themselves with
K5SSS         smart people who disagree with them."  --Aaron Sorkin





More information about the ARIN-PPML mailing list