[ppml] Resurrecting ULA Central [was: Re: Policy Proposal 2006-2: Micro-allocations for Internal Infrastructure - to be revised ]

[william(at)elan.net]

On Thu, 20 Apr 2006, David Williamson wrote:

> On Thu, Apr 20, 2006 at 04:36:37PM -0400, Thomas Narten wrote:
>> Would there be interest in resurrecting the ula-central document?
>
> Speaking only for me, yes.
>
> It would seem mildly clever to put the AS number somewhere in the
> network number (pick an appropriate range of bits in the prefix), and
> then hand out the appropriate /48 implicitly with any AS received from
> an RIR (including legacy AS holders, obviously).

I'm also in favor of this.

In fact my view is that we really don't need ip allocations if we 
engineered ipv6 properly. We could have BGP table where it is entirely
one asn announcing routes to other ASNs and and ip addresses are composed 
of global network part that is ASN, local network part that is entirely 
dependent on how each ASN wants to set it up (and it would be cool if 
local network part was 32-bit in size so you could just use existing
ipv4 numbers for it without any changes) and then additional bits for 
local device specific address (and in my view 64bit is way too much).

We could probably try it as a global experiment with existing ipv6
with setup like:
  8 bits - IPv6 ASN experiment global prefix
  32 bits - ASN number
  24 bits - local network address if possible last 24 bits of ipv4

Last one (24 bit local) would not be enough to fully map ipv4 address
for those who have ip addresses from different /8 ipv4 blocks currently.
(unless we abandon 64 bit boundary for local device id...)

-- 
William Leibzon
Elan Networks
william at elan.net