Re: draft-ietf-l2tpext-circuit-status-extensions-02.txt

Hi Ignacio,

Many thanks for the comments, please see inline.

On 3/13/2009 1:16 PM, Ignacio Goyret wrote:
> Hi,
> A few comments on the circuit-extensions draft:
> 
> 1) Since this draft adds some new bits, including an
>    "Standby" bit that affects the data plane, would it
>    make sense to add a bit to request a remote loop?
>    (ie, ask the remote LCCE to send all packets received
>    over the L2TP tunnel back to the sender LCCE)
> 
>    Mmm, come to think of it, this may be the wrong AVP
>    for that kind of function.

This is a very interesting suggestion; I tend yo agree with your
conclusion though.

I think that such loopback capability likely requires more fields than a
single bit (e.g., type of loop, direction, all dataplane packets versus
only some packets, some locking mechanism, etc.). It would likely need
its own AVP and separate definition.

> 
> 2) Besides updating RFC3931, this draft is also implicitly
>    updating 4591, as well as 4349, 4454 and 4719.

This is a great point. I checked and I don't think there's other RFC
specifying Circuit Status AVP and N-bit usage, in addition to those
four. I'm not sure how implicit or explicit the update is (or needs to
be), see below;

> 
>    The new bits and the effect on the N-bit are applicable
>    to all protocols, right?

Yes.

> 
>    If so, RFC4591 should probably be moved to the normative
>    section of the references, and the others should be added.

OK. RFC4591 is currently Informative because it was only included to
show that, while RFC3931 says that the N bit is for FRoL2TPv3, RFC4591
uses it exactly the saw way as the other protocols.

I agree, we'll add Normative references to RFC4349, RFC4454 and RFC4719,
and move RFC4591 as Normative. We'll add a paragraph with citations to
these 4 explaining why they are there.

Q: How do you suggest we handle the "Updates:" header of the document?
Right now it reads:
	Updates: 3931 (if approved)
Should the other 4 RFCs be added here?

> 
> That's all I have.
> 
> 
> I'd like to start a WGLC on this document but,
> with IETF74 so close, we should wait until after
> the meeting.

Maybe it's a feature that it's so close to IETF74, to maximize the
review and comment. You could start a WGLC extending it by 1 or 2
additional weeks? Or otherwise, we can publish a new revision with these
changes when submission re-opens and WGLC it at that time; what do you
prefer?

> 
> In the meantime, please, take the time to read this
> document: it is an easy read.

Thanks !

-- Carlos.

> 
> Cheers,
> -Ignacio
> 
> _______________________________________________
> L2tpext mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/l2tpext
>