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

Marshall Eubanks tme at multicasttech.com
Wed Dec 21 07:42:10 EST 2005


Here is my take on things

1.) I support 2005-9 with the exception of the "Nomenclature"  
section, and I have  a suggestion for the text.

2.) I think that much of the nomenclature discussion is pretty silly,  
including some of mine. Unless someone comes up with a scheme to  
really burn through them, I would  predict that we would not need to  
use more than 6 digits for ASN's a few decades to come (I estimate  
about 20 years).

- We can live with expressing 5 digit ASN's as simple decimal  
numbers. The existence proof for that is that we do.
I see no reason to worry about adding a digit to that, and I can  
certainly live with 6 digit decimal numbers.

- If someone does come up with a scheme to burn through lots more  
ASN, I predict that will involve applying
some sort of structure or aggregation or overlay to ASN numbers, and  
that will likely impose its own numbering convention. In that case,  
anything we do now will likely be superseded.

- If people years from now feel that they need more order to ASN  
numbers, there is nothing stopping them from doing it.

- So, my vote is for simple decimal numbers to describe ASN. This is  
a preference, not a MUST. If this
comes up at the meeting  with the current nomenclature, I would  
support it.

3.) 2005-9 says nothing about whether or not people requesting 4 byte  
ASN can request specific 4 byte
ASN. I think that the current sequential assignment should continue.  
The text

On 1 January 2007 the registry will process applications that  
specifically request 4-byte only AS Numbers and allocate such AS  
Numbers as requested by the applicant. In the absence of any specific  
request for a 4-byte only AS Number, a 2-byte only AS Number will be  
allocated by the registry.

is ambiguous. I predict that if specific numbers CAN be requested,  
they WILL be and a land rush will ensue.
(I want ASN 1,000,000 or ASN  1.11111, depending on  the nomenclature  
adopted.) So, I would add

In accordance with current policy, applications will only be able to  
request AS Numbers that require 4 bytes to express, not specific AS  

Marshall Eubanks

On Dec 21, 2005, at 12:05 AM, Bill Darte wrote:

> William,
> Your post below is excellent....thanks.
> Using "." or ":" makes sense because these are already in use for ip
> and other addresses. It has also become a kind-of tradition that ":"
> separates parts of the numbers when they are written in hexadecimal
> where as "." separates when parts are written as decimal.
> Also that we separate 32-bit number into several parts in writing does
> not mean somebody else can not write and use the same number as full
> 32-bit [unsigned] integer. In fact as many of you probably know you  
> can
> use use full 32-bit number as IPv4 address, for example just for fun
> try "ping 3231054855" from your dos/unix prompt and then also try
> "http://3231054855" at the browser and see where it gets you :)
> ---
> William Leibzon
> Elan Networks
> william at elan.net
> _______________________________________________
> PPML mailing list
> PPML at arin.net
> http://lists.arin.net/mailman/listinfo/ppml
> _______________________________________________
> PPML mailing list
> PPML at arin.net
> http://lists.arin.net/mailman/listinfo/ppml

More information about the ARIN-PPML mailing list