[address-policy-wg] 2016-04 New Policy Proposal (IPv6 PI Sub-assignment Clarification)

Anno domini 2016 Kai 'wusel' Siering scripsit:

Hi Kai,

> am 21.10.2016 um 10:32 schrieb David Croft:
> > Strong support in principle. We have been denied IPv6 temporary
> > assignments due to the NCC's interpretation that a single DHCP lease
> > on wifi is a "subassignment" to another entity, which was clearly not
> > the intention.

> I'm new to this, so please bear with my simple-mindedness here,
> but to me it looks like »the NCC's interpretation that a single
> DHCP lease on wifi is a "subassignment" to another entity« iswhat
> should be addressed, instead of special-caseing something intoan
> already complex policy document.

> Reading through ripe-655 multiple times, I can't find a basisfor
> counting an temporal, RA/DHCP-based, lease of PI space by an
> organisation holding Provider Independent Resources as a sub-
> assignment of a/128 prefix.
 [...]
> An »assignment« therefore is something that – to prevent the word
> »assign« here – dedicates an address space to someone for a specifc
> purpose and this act needs to be registered (see 5, and esp. 5.5)
> in an RIR database. But PI assignments cannot be assigned further,
> as clearly stated.
> 
> Operating a WiFi network for employees, relatives, event-visitors or
> even the general public (i. e. open WiFi, no WPA2) as an End User
> of Provider Independent Resources does not constitute an
> »assignment«, neither in terms of ripe-655 nor in real life.
> 
> As far as I understand the process, this WG suggests the policies
> which the NCC implements? So, unless there was a previous call from
> this WG to the NCC to interpret things as it is reportedly done –
> which, from the comment, wasn't the case? –, why not just vote on
> a statement that NCC's interpretation is outside of the scope or
> intention of ripe-655?

There is no such thing as a "vote on a statement hat NCC's
interpretation is outside of the scope". See below.

> I mean, it's not the policy that's at fault here; there exists
> an _interpretation_, used by the RIPE NCC during evaluation of PI
> space requests, which at least to me is not even remotely covered
> by ripe-655. Don't mess with what's not broken, fix what is broken ;)

As previously discussed via IRC the means of the community to "tell
the RIPE NCC how to interpret things and do it's job" are the
RIPE policies, such as ripe-655. Obviously the current policy is
"broken" - to use your wording - as there have been multiple
interpretations within the NCC.

The proposed policy change tries to close this room for interpretation
by defining "an interpretaton" for the point in question following the
Policy Development Process. See https://www.ripe.net/participate/policies
for details about this.

Best
Max