[ppml] Policy Proposal 2005-9: 4-Byte AS Number

Owen DeLong owen at delong.com
Tue Dec 20 08:14:25 EST 2005



--On December 20, 2005 11:59:24 AM +0000 Michael.Dillon at btradianz.com wrote:

>> To reiterate for the record now that this is a formal policy proposal, I
>> would agree with the apparent ppml consensus
>
> Consensus?
>
>> that the colon separator
>> should be changed to a period to avoid confusion with the nomenclature
>> used by BGP communities.
>
> Several people pointed out that there is no need
> for any special separator since the decimal number system
> in common use already allows for representing numbers
> greater than 65535. In other words, there was not any
> CONSENSUS about sticking spurious punctuation marks
> into AS numbers.
>

While it is true that there were a few comments of this nature, the vast
majority of respondents who commented on this did seem to favor the use
of the decimal and segmenting the AS number into 16 bit chunks for
human readability.  I know you don't agree, but, I think there was a
strong enough majority to call it consensus.

> Is the number 0.63535 a valid AS number?

yes

> What about 0.65553?

I would say that it is equally valid to 0.0.0.384 as an IP address.

Technically, I suppose, an argument could be made that it is valid, but,
I doubt most parsers would treat it the way you would expect.

> When the dot notation was introduced for IP addresses,
> they marked an important bit boundary that was a fundamental
> part of the IP address. The 32 bit identifier was divided
> into a NETWORK portion and a HOST portion. This division

However, there's three dots and only one boundary for any given
address.  I don't know whether you'll accept this or not, but,
I do believe that a portion of the thinking of using dotted quads
was to make IP addresses more human-readable and reduce human
audio transmission/reception errors.

> was done on one of three 8-bit boundaries depending on
> address class and therefore there are 3 dots in an IP
> address marking the boundaries.
>
I'll also point out that dotted quad did not go away with the
advent of CIDR and that there was no dot to mark the difference
between unicast and multicast space or multicast and experimental
E/F/G/... space.  The dots remain today because they are useful
for human readability.

> AS numbers are simple integers with no such internal
> structure.
>
IP addresses today are simple integers as well.  You can either say they 
have
no such internal structure, or, you can say that the internal structure is 
no
longer related to octet boundaries.  Either way, the position of the dots
in IP addresses is no longer related to that structure, but, the dots have
not moved or been abandoned.

When was the last time you saw someone write an address within a /16 as
141.32892?  (sure, it's a legal representation, but, nobody writes IP
addresses that way).

There are a lot of parallels between ASNs and Postal Codes.  I would expect
that when ICANN/IANA starts issuing 32 bit ASNs, they will likely issue
16bit blocks to RIRs who will then allocate individual ASNs from them.
As such, it might be useful to have A.B format notation where A can be
related back to:
	A=0	=	Original ASN swamp, non RIR specific
	A>0	=	Specific RIR to which that chunk was assigned

Of course, if that happens, then, while not technically structure within the
ASN, it does mean the bits have somewhat separate meaning.

Owen

-- 
If this message was not signed with gpg key 0FE2AA3D, it's probably
a forgery.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20051220/b26e8b71/attachment.sig>


More information about the ARIN-PPML mailing list