stla registry db issue

[Brian E Carpenter]

Well, here is the key text from RFC 2374. (there is no reason it should
be different in subTLA space):

   ...It is recommended that
   organizations assigning NLA address space use "slow start" allocation
   procedures similar to [RFC2050].

   The design of an NLA ID allocation plan is a tradeoff between routing
   aggregation efficiency and flexibility.  Creating hierarchies allows
   for greater amount of aggregation and results in smaller routing
   tables.  Flat NLA ID assignment provides for easier allocation and
   attachment flexibility, but results in larger routing tables.

My concern is that the way Kazu asked his question, with the concern about
frequent updates, did not seem compatible with the idea of slow start and
hierarchical aggregation. If we don't start with habits that create aggressive
aggregation, IPv6 routing will be in deep trouble as it grows.

I also have a concern that if an operator is really an ISP, giving them an
NLA instead of a subTLA may be a problem until we have proved how to do
convenient renumbering. What happens when they want to migrate away from using 
WIDE as their aggregator? (I realise that this is a heretical thought, since
the current rules on subTLAs are more restrictive.)

However, I agree that Kazu is not describing a strict violation of the RFCs.

   Brian

Bob Fink wrote:
> 
> At 11:08 AM 12/22/99 +0900, =?iso-2022-jp?B?GyRCOzNLXE9CSScbKEI=?= wrote:
> >From: Brian E Carpenter <brian@hursley.ibm.com>
> >Subject: Re: stla registry db issue
> >
> > > Kazu, that is exactly what I am saying. It is a serious error to
> > > split a subTLA for subISPs. One ISP must get one subTLA. Never split
> > > a subTLA between ISPs.  (The same applies to exchange points.)  If
> > > you split subTLAs between ISPs, you create the IPv6 toxic waste
> > > dump.
> >
> >In my understanding, the difference between TLA and sTLA is only
> >scale. As TLA is to be split into NLAs, sTLA is to be split into NLAs.
> >
> >Nobody knows what is the best component/organization for TLAs and
> >NLAs. However, in the current Internet model, TLA is big ISPs and NLA
> >is medium ISPs.
> >
> >It seems to me that your opinion is incorrect.
> 
> In looking at the mail on this it seems that there is a confusion between
> you folk over what splitting the subTLA means, and you both have it right,
> but think the other is not understanding you. The /35 subTLA assigned by
> the RIR is intended to be split up to the right of the /35 (i.e., in the
> NLA space) for use by lower level providers, but not to the left of the /35
> (as this is part of the extended subTLA field the RIR is holding back for
> future use, unspecified for now).
> 
> Bob