[arin-ppml] Draft Policy ARIN-2020-3: IPv6 Nano-allocations

[Owen DeLong]

> On Apr 18, 2020, at 15:20 , William Herrin <bill at herrin.us> wrote:
> 
> On Sat, Apr 18, 2020 at 2:44 PM Owen DeLong <owen at delong.com> wrote:
>> Handing out a /48 to each end site is a core engineering design that was put into IPv6 for many valid reasons.
> 
> Hi Owen,
> 
> If I understand correctly, the /32, /48 and /64 size recommendations
> were originally discussed on public, now-archived IETF mailing lists.
> At the time, dynamic dialups were the equivalent of what we today
> consider residential customers (as opposed to business and hobbyist
> end sites). Have you found some citations among those discussions
> which clarify that /48 was intended to apply to all ISP customers
> explicitly including individual end users of the ISP, not just network
> clients of the ISP.

While I don’t have citations, I can say that you aren’t entirely correct. While not as ubiquitous
as they are today, both CMTS and DSL systems were known residential ISP technologies
at the time those boundaries were established and the idea of automatic hierarchical
topologies through layered DHCP-PD was very much something that was considered as
applicable to a potential future residential context. It would take some time to find any
citations for this, but I personally recall participating in some of those discussions.

By the time we were establishing those boundaries, it was already assumed that an
always-on broadband connection via DSL, CMTS, or other future technology with
similar or even greater capabilities would become the norm for residential internet
services.

Not that all of the assumptions came true. It was assumed that IPv6 would be an enabling
technology helping to drive deployment of that infrastructure, rather than the other way
around.

There are others on this list who were there. I’m sure someone will correct any
misstatements I may have made or any incorrect details.


Owen