Re: [hostmaster-staff] Re: MIR proposal / reservation revisited?
- Date: Tue, 2 Oct 2001 14:29:09 +0100
I have to say purley from a thechnical perspective this has merit.
> >PS To answer Wilfreds question, it is not reservation revisited it is
> >aggregation revisited ;)
>
> Stephen,
>
> in my opinion (from a technical persepctive only), there are two ways of
> attacking the problem:
>
> - 1) to clean up the legacy routes, you have to re-number (which I guess
> can be done within your address pool(s)
Ok this is the legacy IP stuff that has been there for upto 10 years and the
mess is partly from the ammalgamation of 17 registries (uunet) Worldcom
networks, MCI networks, AOL networks (legacy) and a number of other smaller
networks such as Pipex, NLnet and others. Renumbering this would be a
nightmare but not impossible, the biggest problem is getting customers to
renumber, as i am sure you are well aware. We could look at trying to
renumber by customer churn into the new space and hand it back once free.
The main reason is not to fix the current mess but to avoid more mess in the
future, that said we are trying to reaggregate teh current routing tables to
bew more efficient.
>
> - 2) to prevent it from happening again, someone, soemwhere, has to
> reserve a "big" block of addresses which would support
> a) additional demand by existing customers, and
> b) aggregation according to your network's internal structure
>
Ok so what you are describing is in simplistic terms the MIR concept. The
MIR would have a big block of addresses not reserved but allocated to it.
This would be based on networks size LIR number and other constraints.
Aggregation would be accross the 1 AS (AS702) and new assignments would be
made out of the new space. Though the allocation
(sub-allocation/partitioning) would be made by the MIR to the LIR the LIR
would talk to RIPE for assignment approval. When the LIR reached 80% and
only when then more space could be applied for to the MIR and get a block of
addresses from the same CIDR preferably applying the NCC method of use last
so natural holes are created next to the initial allocation.
> Item 2) is where my notion of "reservation" is coming from.
> Do I still get it (completely) wrong?
No i think you understand the issues just not the solution, reservation is a
bad word and bad practice in the current climate.
>
> Btw, doing only one of those two is probably not going to work...
>
> Wilfried.
