Re: [v6ops] Chair decision on WGLC for draft-ietf-v6ops-dhcp-pd-per-device-04

[Martin Huněk <[e-mail]>]

Hi David,

I've tried to gather some information about resources available to UMN to find out how it that you can accommodate this draft in your addressing plan. What I could find, you have a single /48 as we do. However, I didn't find any AAAA records for websites nor e-mail. So I got to ask:

With the number of students + staff, is it even possible for you to use this draft? I would say you can consider UMN to be a sort of large "enterprise" this draft is targeting.

It is certainly not for us, even we are quite smaller university also with /48, but our addressing plan would be an adult in the less than 2 years (we started around 2007) and we are fully dual-stack. In our case, only the Wi-Fi network with /64 per device would eat /49 by itself. 

Regards,
Martin

Dne neděle 5. listopadu 2023 22:20:55 CET, David Farmer napsal(a):
> On Sun, Nov 5, 2023 at 2:42 PM Gert Doering <[e-mail]> wrote:
> 
> > Hi,
> >
> > On Sun, Nov 05, 2023 at 01:40:01PM -0600, David Farmer wrote:
> > > Your argument above is hypocritical. If a /64 per phone is justified for
> > > cell carriers, it should be justice for end-user customers as well. If
> > it's
> > > not okay for end-users, it should not be okay for cell carriers either.
> >
> > Maybe you need to acquaint yourself to the realities of address
> > distribution hierarchies, outside large universities holding a /32?
> >
> 
> Maybe you need to acquaint yourself with the realities of other kinds of
> large-scale networks than just service providers. That is who this draft is
> aimed at, not residential or SOHO customers.
> 
> 
> > > If a cell carrier can get enough /64s to have one for every customer's
> > cell
> > > phone, then end-users should be able to do the same. If RIPE has to
> > change
> > > its policies to allow this, then so be it.
> >
> > You do not want everybody to become a RIR member, because of obvious
> > implications on routing table size.
> >
> 
> RIPE policy can allow larger than /56 or /48 for end-users without everyone
> becoming a member and impacting the route table size.
> 
> Furthermore, RIR policy is one of many factors in route table growth. The
> backbone transit ISP marketplace is another significant factor. You have to
> purchase transit for your RIR-assigned prefixes. This cost exceeds what
> most residential or SOHO customers are willing to pay, probably by a factor
> of 10 to 100.
> 
> You seem to be saying that because residential or SOHO users don't have
> enough address space, it is also invalid for larger enterprises with the
> address space necessary for this addressing model and that it is only
> allowed for carriers and service providers.
> 
>