[ppml] Proposed Policy: 4-Byte AS Number Policy Proposal

[Geoff Huston]

At 01:46 AM 14/12/2005, Rich Emmings wrote:
>On Tue, 13 Dec 2005, Geoff Huston wrote:
>
>[regarding separation of policy proposals on documentation/private space]
>
>No real objection to that.
>
> > While I have heard of folk indicating that the private IPv4 address 
> space is
> > too small, I have not heard of anyone saying that the 1023 AS numbers
> > reserved for private use is too small. [ balance deleted]
> >
>
>More thinking that 1023 could perceived as some as being too large a
>reservation of the top two bytes.
>
> >
> >>         b) administrative ease is addressed here by using the same range,
> >> at the possible expense of "c"
> >>         c) efficiency of allocation is not -- could be with a small range
> >> at the top, at the expense of "b"
> >
> >
> > it is unclear to me what your point is here.
> >
>
>Reservation of the top 1023 2 bytes in 4 byte space is a larger chunk of
>space, than is probably needed, but using the same numbers as in 2 byte
>space is easier to remember -- administrative ease.  The other side is to
>only strip off a few addresses at the top, say: 65520.x - 65534.x, instead
>of 1023 (in the first 2 bytes of the 4.)


This policy proposal says absolutely nothing about expansion of the private 
AS number space. Currently the AS numbers 0 (or 0.0 in a 4 byte notation), 
23456 (0.23456) and 65535 (0.65535) are reserved by IANA, and the range 
64512 - 65534 (0.64512 - 0.65534) are designated for private use 
(http://www.iana.org/assignments/as-numbers) Neither the IETF 4 Byte draft 
nor this proposal say anything about extending this set of reserved or 
designated numbers in any way. i.e. in a 4byte AS world private AS numbers 
will still be the range of AS numbers from 0.64512 to 0.65534.



>AS2 -> AS4 transition will mirror what will happen in IPv6 implementation.
>Won't happen until/if forced, workarounds allowing the continued use of AS2
>space will probably be developed ala some of the tools which extended the
>IPv4 lifetime (both good and ugly), they'll be some things that won't leave
>IPv4 space for 10+ years, if ever, and they'll be grumbling all over.  In
>other words, same as always.  (but no reason not to get started now)


I have to disagree with this assertion. As noted in the 4byte AS draft, and 
as noted in http://www.potaroo.net/ispcol/2005-08/as.html, the transition 
is entirely different to that associated with IPv6. In this case existing 
players in the inter-domain routing space have no requirement to do 
anything now, or in the future. Let me say it again: Folk who have 2 byte 
AS numbers need do _nothing_ , i.e. the transition allows for "the 
continued use of AS2  space" indefinitely. Please read these references, as 
they are intended to describe the transition situation in as comprehensive 
manner as possible.


thanks,

   Geoff