Dear Jeff,

The sentences are hard to parse. Let me summarize my current understanding:

 RFC 9515 updated the following BMP registries at IANA from "Specification 
Required" to "First Come First Served":

BMP Statistics Type 32768-65530
BMP Initiation Information TLVs 32768-65530
BMP Termination Message TLVs 32768-65530
BMP Termination Message Reason Codes 32768-65530
BMP Route Mirroring TLVs 32768-65530
BMP Route Mirroring Information Codes 32768-65530
BMP Peer Up Message TLVs 32768-65530

That lowers the bar for adding additional information's into the existing IANA 
managed BMP message types. See 
https://datatracker.ietf.org/doc/html/rfc8126#section-4.4 vs. 
https://datatracker.ietf.org/doc/html/rfc8126#section-4.6.

While draft-ietf-grow-bmp-tlv-ebit introduces enterprise managed BMP message 
types by suppressing the  experimental 
(https://datatracker.ietf.org/doc/html/rfc8126#section-4.2) code points of IANA 
managed BMP message types.

Taking https://datatracker.ietf.org/doc/html/rfc7120#section-2 into 
consideration, that would mean the following:

- RFC 9515 enables without an IETF document to add additional information's to 
existing BMP message types
- draft-ietf-grow-bmp-tlv-ebit enables enterprises to add additional 
information's to existing and future BMP message types
- https://datatracker.ietf.org/doc/html/rfc7120#section-2 governs at which 
stage code points from the "Standards Action" range can be allocated
- experimental code point allocation is not supported anymore

Quoting from: https://datatracker.ietf.org/doc/html/rfc3692#section-1

  By definition, experimental numbers are not guaranteed to be
   unique in any environment other than one where the local system
   administrator has chosen to use a particular number for a particular
   purpose and can ensure that a particular value is not already in use
   for some other purpose.

We understand that enterprise governed and IANA experimental code points are 
very similar. The only difference is that the company who issued the code point 
is now identifiable.

To summarize, both documents together lowers the bar for IANA code point 
allocation in above mentioned ranges and enables enterprises to register their 
own code points in favor of supporting experimental code points.

This mechanism has been proven well for IPFIX and experimental code points in 
BMP appears to be not being used since introduction of the protocol.

If I understood correctly you are raising the point wherever the working group 
understood that experimental code points are no longer being supported with 
draft-ietf-grow-bmp-tlv-ebit anymore and the role of RFC 9515 correct?

My response: draft-ietf-grow-bmp-tlv-ebit anymore and RFC 9515 are 
complementary and together setting with RFC7120 the right setting for managing 
codepoints in BMP.  

Quoting from: 
https://datatracker.ietf.org/doc/html/draft-ietf-grow-bmp-tlv-ebit-06#section-3.3
Of these, the Experimental values are being suppressed in favor of using the 
E-bit mechanism described in this document;

Should this be mentioned Section 1 in the document to make it more evident?

Best wishes
Thomas

-----Original Message-----
From: Jeffrey Haas <[email protected]> 
Sent: Saturday, March 22, 2025 6:36 AM
To: [email protected]; [email protected]
Subject: [GROW] bmp ebit and rfc 9515


Be aware: This is an external email.



As I'm glancing through the ebit draft, it occurs to me that the changes to RFC 
9515 have a potentially perverse interaction with it: All new FCFS fields are 
now required to use the ebit.

This would include the experimental ranges.

I had noted a prior set of emails last year from Ahmed when this bit of 
interaction occurred to me.

This isn't necessarily a bad thing.  But it does have the interesting property 
that if grow doesn't want some vendor's PEN stuck in something forever for a 
feature that might be popular, it means that "early adoption"
from the standards point becomes almost a required thing.

The WG should ponder if this is a set of properties it's really wanting and, if 
so, the ebit document should call it out explicitly as an interaciton with RFC 
9515.

-- Jeff

_______________________________________________
GROW mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

_______________________________________________
GROW mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to