Re: Proposal: valid-until time for an allocation

[Robert Kisteleki <robert@ripe.net>]

Rob Austein wrote:
> At Wed, 18 Apr 2007 21:13:43 +0200, Róbert Kisteleki wrote:
> > Rob Austein wrote:
> > > At Wed, 18 Apr 2007 12:15:36 +0200, Róbert Kisteleki wrote:
> > > > (list-resources :self-id 42		; issuer id
> > > > 		&optional		; If left off, we're asking
> > > >                                          ; about self rather than child
> > > > 		:child id)		; subject id
> > > > => (((:ipv4-address "10.0.0.44/32" "10.3.0.44/32")
> > > >       (:ipv6-address "fe80:dead:beef::/24")
> > > >       (:as-number "666")
> > > >       (:valid-until "date-in-some-form")
> > > >       (:class-name "experimental")
> > > >      )
> > > >      ...)
> > > The :valid-until extension sounds right.
> > >
> > > I don't see the need for :class-name.  What does the RE do with it?
> > If the above returns two ipv4-ipv6-validuntil sets, then it's two 
> > (sub)classes form the ISP's point of view.
>
> Yes.
>
> > We also know that the IR has to persistently name those two
> > classes.
>
> The IRBE doesn't need to name them.  The RE does, but it can make them
> up (X0001, X0002, ...).  I see no need for the IRBE to care.

The IRBE doesn't need to name them, sure. But how can the RE give the 
classes persistent names? Since the list-of-resource-sets can change 
(the number of them, the order of them, the contents of them, etc.) you 
have to have something coming them the BE that is stably associated to 
that set. That is what I call a 'name' for that set.

Robert