Re: [Idr] Early MIB Dr. Review of draft-ietf-idr-bgp4-mibv2-06.txt

----- Original Message -----
From: "Jeffrey Haas" <jhaas@pfrc.org>
To: "tom.petch" <cfinss@dial.pipex.com>
Sent: Wednesday, May 14, 2008 4:15 PM

> Tom,
>
> On Wed, May 14, 2008 at 11:41:15AM +0200, tom.petch wrote:
> > Jeff
> >
> > As far as I can see, there is but one place where working group feedback is
> > requested and that is for
> > > 5) InetAddress/InetAddressType
> > but I am unclear what point you are asking about:-(
>
> Specifically:
>
> > OBJECT bgpPeerAfLocalAddr
> > SYNTAX InetAddress (SIZE(4|16|20))
> > DESCRIPTION
> >     "An implementation is required to support IPv4 peering
> >      sessions.  An implementation MAY support IPv6 peering
> >      sessions.  IPv6 link-local peering sessions MAY
> >      supported by this MIB."
> >
> > The reason for this wording is that we've never come to proper
> > consensus
> > about ipv6-only routers.
>
> I don't have any issues with moving the sizes into the conformance
> statements.  My question specifically is about what we'd require in the MIB
> for peering session support.
>
> The BGP-4 spec pretty much says you're going to have your peering
> sessions via TCP on IPv4.
>
> We know you can do IPv6-only peering sessions.  Further, there has been
> some experimentation (largely related to IXes I think) that were IPv6
> link local only peering.
>
> The above language is intended to say:
> 1) You must support IPv4 peering sessions.
>
> I think this is implied by the BGP-4 spec.  However, that might be
> inconsistent at a later date by people who have IPv6 only routers.  This
> is, admittedly, looking far into the future.
>
> 2) You can support IPv6 peering sessions.  That's not really called out
> in any specific spec but is bourne out by operational practice.  I don't
> believe this is controversial.  However, the working group may not want
> this in the MIB.
>
> 3) You can support IPv6 link local peering sessions.  The only draft
> that I remember on the topic is long expired.  There were some
> operational details that weren't controversial but do have impact on
> RFC 2545 behaviors.  The working group may choose to say "no, we don't
> want this in the MIB at all".
>

Understand; I do not have a view, seeing very little IPv6 and then only
experimental.  Since this is a functional question about BGP4, and not one about
the MIB, you might want to resend this message with a different title, in order
to circumvent those whose blacklists include such words as 'MIB' and 'SNMP' :-)

I think that the current wording is a fudge - 'is required to' for Ipv4, MAY for
IPv6.  I think that the former needs turning into a RFC2119 word, but do not
know which it should be.

Tom Petch

> > Tom Petch
>
> -- Jeff

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