Re: [Idr] WGLC on draft-ietf-idr-as-private-reservation-00

[Nick Hilliard <[e-mail]>]

On 11/12/2012 19:20, Robert Raszuk wrote:
> Based on number of use cases it seems to me like perhaps we should
> just declare the top most bit in 4 octet AS space as indicator of
> private vs public.

Excuse me if this sounds atrociously naive, but putting my operator hat on:

- i'm human and regularly need to parse asns in as-paths

- i don't naturally count in binary

- i would really like a set of numbers which I can easily see are private
or public.  Decimal alignment would work nicely, both for me and the NOC
staff that are going to be dealing with these numbers on a daily basis.

- looking at 4278190079 and 4278190080, I simply cannot tell which might be
public and which might be private.  This will cause leakage because it will
not be obvious on sight to an operator looking at a bgp aspath whether the
range should be propagated or not #operatorfail

- unless something drastic has changed recently, I'm not aware of any plans
to assign bitwise masks to ASNs.  Consequently bitwise alignment for a
private ASN range is a purely arbitrary construction

- please change to using a decimal-aligned range.  it would make my life
and the lives of many operators a little easier.  I suggest 4000000000 or
maybe 4200000000, or something else.  Not 4278190080, because that number
means nothing to me

- otherwise, I love the idea of this draft.  Please let us have private
ASNs and lots of them.  There are plenty of them to spare in the can.

thanks,
Nick