ietf
[Top] [All Lists]

Re: Last Call: <draft-housley-number-registries-02.txt> (Internet Numbers Registries) to Informational RFC

2014-02-03 19:51:41
I believe if RFC 6890 were used as a template, a Draft could be put together quickly, I'm personally willing to help. I don't see why it should be controversial within the community, but unfortunately those can be famous last words. :)

On 2/3/14, 14:20 , Russ Housley wrote:
David:

With respect to the AS Special purpose Registry, I might sit more comfortably 
in its own document, and then it could be referenced from this document. At 
this point, I'm concerned that he delay to this document would be quite 
unfortunate.  Do you believe that such a document can be produced and approved 
rapidly?  If not, then I'd like to see this document move forward and a 
subsequent document could make an update if there is community consensus to 
evolve the form of the special-purpose AS number registry.

Russ



On 1/25/14, 14:14 , Adrian Farrel wrote:
Hi Russ,

Thanks for this. The -03 revision just posted addresses my comments.

Adrian

Russ,

I also like the -03 revision, its has generally moved everything in the right 
direction.  But I have a few comments on it.

1. The part about RFC2860 in Section 1 seems like an incomplete thought, at the 
very least it seems awkward to me.  I'm really not sure what you are intending 
to say.  But, I agree RFC2860 is relevant to the discussion.  I'm just not sure 
you have nailed what to say about it.

2. I really like the idea of creating a "Special-Purpose AS Number Registry".  However, it may be a 
better idea to spin-off the creation of a "Special-Purpose AS Number Registry" into a separate 
draft.  I'm concerned that trying to do two important things in the same document will fail to achieve one or 
both of the important things.  For instance the simple section 3 you have currently worries me.  I'd really 
think RFC6890 is the template to use for creating a "Special-Purpose AS Number Registry", more 
below.

If you spin-off the creation of a "Special-Purpose AS Number Registry", the 
following items are for that draft.

3. There is a draft in WG Last Call in the IDR that documents the reservations of ASNs 
65535 and 4294967295, the "Last ASNs".
http://tools.ietf.org/html/draft-ietf-idr-last-as-reservation-01

4. As I said, I like the idea of creating a "Special-Purpose AS Number Registry", but it 
should have certain information recoded about each AS number range.  This information should be 
similar to the format described in Section 2.2.1 of RFC6890 for the "Special-Purpose IP 
Address Registries" for IPv4 and IPv6, that you reference elsewhere in the draft.

First, there should be general information describing each of the 
special-purpose ASN ranges, essentially the first half of the columns specified 
in Section 2.2.1 of RFC6890;

o AS Number Range - An AS number or range of AS numbers that has been
   registered for a special purpose.

o Name - A descriptive name for the special-purpose AS number range.

o RFC - The RFC through which the special-purpose AS number range was
   requested.

o Allocation Date - The date upon which the special-purpose AS number
   range was allocated.

o Termination Date - The date upon which the allocation is to be
   terminated.  This field is applicable for limited-use allocations
   only.

Then, information specific to how the special-purpose ASNs are to be used.  
There may be others as well, I but I think these are the minimum.

o Local - A boolean value indicating whether an AS number from the
   allocated special-purpose AS Number range is valid when used as the
   value of the "My Autonomous System" field of a BGP OPEN message, or
   equivalently in the Capability Value of the "support for four-octet
   AS number capability".

o Global - A boolean value indicating whether an ASN from the allocated
   special-purpose ASN range is valid in the AS_PATH or AS4_PATH
   attributes when advertised to the global Internet.

o Reserved-by-Protocol - A boolean value indicating whether the
   special-purpose address block is reserved by BGP, itself.  This
   value is "TRUE" if the RFC that created the special-purpose ASN
   range requires all compliant implementations to behave in a special
   way when this ASN is used.

I believe this is the table for these last three items

   ASN Range              Local    Global   Reserved-by-Protocol
   ---------------------  -------  -------  --------------------
   0                      FALSE    FALSE    TRUE
   23456                  TRUE[1]  TRUE[1]  TRUE
   64496-64511            FALSE    FALSE    FALSE
   64512-65534            TRUE     FALSE    FALSE
   65535                  FALSE    FALSE    FALSE
   65536-65551            FALSE    FALSE    FALSE
   4200000000-4294967294  TRUE     FALSE    FALSE
   4294967295             FALSE    FALSE    FALSE

   [1] Only valid in a BGP OPEN message or AS_PATH, not valid within a
   four-octet AS number capability negotiation or AS4_PATH.


Thanks.

--
================================================
David Farmer               Email: farmer(_at_)umn(_dot_)edu
Office of Information Technology
University of Minnesota
2218 University Ave SE     Phone: 1-612-626-0815
Minneapolis, MN 55414-3029  Cell: 1-612-812-9952
================================================



--
================================================
David Farmer               Email: farmer(_at_)umn(_dot_)edu
Office of Information Technology
University of Minnesota
2218 University Ave SE     Phone: 1-612-626-0815
Minneapolis, MN 55414-3029  Cell: 1-612-812-9952
================================================