Re: [OSPF] OSPF WG Last Call for Traffic Engineering Extensions to OSPFversion 3 - draft-ietf-ospf-ospfv3-traffic-10.txt

[Acee Lindem <acee@redback.com>]

Adrian,

How's this:

***************
*** 342,348 ****
     The Link TLV describes a single link and consists of a set of sub-
     TLVs [TE].  All of the sub-TLVs in [TE] other than the Link ID  
sub-
     TLV are applicable to OSPFv3.  The Link ID sub-TLV can't be  
used in
!    OSPFv3 due to the protocol differences between OSPFv2 and OSPFv3.

     Three new sub-TLVs for the Link TLV are defined:

--- 342,351 ----
     The Link TLV describes a single link and consists of a set of sub-
     TLVs [TE].  All of the sub-TLVs in [TE] other than the Link ID  
sub-
     TLV are applicable to OSPFv3.  The Link ID sub-TLV can't be  
used in
!    OSPFv3 since it defined to contain the IPv4 address of the  
Designated
!    Router (DR) for multi-access interfaces.  In contrast to OSPFv2,
!    OSPFv3 always identifies a neighboring router by the Router ID  
(Refer
!    to section 2.11 in [OSPFV3]).

     Three new sub-TLVs for the Link TLV are defined:

***************

I plan to wait until the WG last call has completed to submit the  
update.
Thanks,
Acee

On Apr 5, 2008, at 11:50 AM, Acee Lindem wrote:

> Hi Adrian,
>
> Thanks for the review.
>
> On Apr 4, 2008, at 3:56 PM, Adrian Farrel wrote:
>
> > Hi,
> >
> > Just a couple of comments...
> >
> > ===
> > Section 1
> > s/proposes the addition of/defines/
>
> Changed.
>
>
> > ===
> > Section 4
> > Forgive me for not remembering this discussion...
> > The draft says that we cannot use the Link ID sub-TLV "due to the
> > protocol
> > differences."
>
> The link-ID is cannot be used since, in the case of multi-access
> network, it contains the IPv4 address of the Designated Router (DR).
> OSPFv3 doesn't have this information.
>
>
> > It then says that the Link ID sub-TLV SHOULD NOT be included
> > (implying that
> > it MAY be included under certain circumstances) but MUST be ignored.
>
> This is the spirit of being conservative in what one sends and
> liberal in what one excepts.
>
>
>
> > 1. Does ignored mean "continue to be flooded" or "stripped from the
> > LSA"?
>
> In OSPF, only the originator should modify an LSA. So, it means  
> neither.
>
>
>
> > 2. Is it not possible to consider operating a GMPLS control plane
> > in an IPv6
> > network where the routers use IPv6 addresses to communicate (so all
> > control
> > plane messages will be addressed using IPv6, and the router address
> > will be
> > IPv6 as described in Section 3) but where the data channel
> > identifiers are
> > assigned from an IPv4 address space? Recall that in GMPLS the
> > interfaces
> > used for OSPF exchange are not those used for data exchange.
>
> I believe it is probable that IPv4 and IPv6 will coexist. However,
> OSPFv3 doesn't know the IPv4 address of the DR (at least it is not
> standardized). Hence, this isn't the right sub-TLV to reflect this
> topology.
>
>
>
> > Whatever the answers, I think it would help if the reasons were
> > clarified
> > beyond "protocol differences."
>
> I'll expand this to describe the multi-access network case. Sound  
> good?
>
> Thanks,
> Acee
>
>
>
> > ===
> >
> > Cheers,
> > Adrian
> >
> > PS I wouldn't mind if you spelled my name right in the acks
> > section :-)
> >
> >
> > _______________________________________________
> > OSPF mailing list
> > OSPF@ietf.org
> > https://www.ietf.org/mailman/listinfo/ospf
>
> _______________________________________________
> OSPF mailing list
> OSPF@ietf.org
> https://www.ietf.org/mailman/listinfo/ospf