> On Nov 19, 2016, at 10:16 AM, Robert Raszuk <[email protected]> wrote:
> 
> Jeff,
> 
> > Part of the motivation I had for suggesting the use of an existing sub-code 
> > is we have *not* changed the semantics in a machine-readable format.  
> > It's still "administrative shutdown".  Today, you'd have zero extra 
> > information 
> > what it's about.
> 
> ​My interpretation of 4271 does not prohibit one to send more then one cease 
> notification message. 
> 
> One would be subcode 2 as today .. the other would be subcode 9 INFO. 

What would the distinction between admin shutdown (clear) vs. info-read the 
friendly message be?  

> 
> > The shutdown draft requires no changes for any useful backward 
> > compatibility.  
> 
> If you add length as proposed by Job it does. 

4271, section 6.4.  I'm being a bit of a pain in my interpretation here but 
simply:
- You still have a valid code and sub-code.  (The admin shutdown subcode is 
already not in 4271 and thus not in compliance with 4271-only implementation.)
- If you don't understand the data, you can't report it back.
- You're instructed to "bring this to the attention of the administration of 
the peer".  This is one of the motivators already for the hexdump many 
implementations use.

I derive my justification from the same used when RFC 4486 was discussed. :-)

> 
> > But in order to signal such other information, the routers in question 
> > would need to support Advisory.
> 
> Advisory is also not structured.

Partially my point.  "SMS over BGP" is how someone commented on this previously.

-- Jeff


_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow

Reply via email to