Re: [Idr] Question regarding bgpM2AsPath4byteTable in BGP4 MIBv2

Tom,

On Tue, Dec 18, 2007 at 04:39:16PM +0100, tom.petch wrote:
> Geoff Huston has picked up and developped the
> draft-michaelson-4byte-as-representation
> I-D that was last called here a year ago, and this now says that the
> representation MUST be asdot ie a decimal number below 65536, or x.y when it
> would be above 65535.  This is gaining some traction in RIPE.  If and when a
> standard is adopted, then I think that there should be a TC in this MIB with an
> appropriate DISPLAY-HINT (I cannot recall seeing anything like it anywhere).

This MIB makes use of InetAutonomousSystemNumber from INET-ADDRESS-MIB.

I'd like to say "when that MIB is updated, all will be well".  However,
the mapping of a DISPLAY-HINT for INTEGER dervied object types won't do
"the Right Thing".  DISPLAY-HINTs for INTEGER, by their nature, can't do
math.

There is a strong push to make use of the textual conventions defined in
the INET-ADDRESS-MIB when possible.  If the goal is to make use of the
display representation in that draft, the wider SNMP community has a
problem, not just the BGP MIBv2.

I would strongly suggest that this issue be raised to the OPS group
directly.  I believe any changes to the BGP MIBv2 will fall out from
there. 

A weak hack would be to partition an AS Number into a "top half" and a
"bottom half" and leave the DISPLAY-HINT to be "d".  However, that won't
give you your dotted notation, it'll simply give you two objects.  From
a MIB perspective, that's not a problem since the AS numbers are not
part of an INDEX.  I think it's a poor solution.  I'd rather punt this
issue to the OPS group. :-)

-- Jeff

_______________________________________________
Idr mailing list
Idr@ietf.org
https://www1.ietf.org/mailman/listinfo/idr