Greg,
thanks. please see in-line.
once I see the update published I will request IETF LC.
-m
Le 2018-04-17 à 18:34, Greg Mirsky a écrit :
Hi Martin,
thank you for your thorough review, thoughtful comments and kind words.
Please find my answers to your questions in-line and tagged GIM>>.
Regards,
Greg
On Tue, Apr 17, 2018 at 8:06 AM, Martin Vigoureux
<[email protected] <mailto:[email protected]>> wrote:
[resend, wrong bfd wg address in first attempt ...]
Authors, WG,
thank you for this document. It is clear and well written.
I didn't find any technical comment to make so I've been nit picking :-)
Please find those comments below.
regards,
martin
---
Please check and address the nits:
https://tools.ietf.org/idnits?url=https://tools.ietf.org/id/draft-ietf-bfd-multipoint-14.txt
<https://tools.ietf.org/idnits?url=https://tools.ietf.org/id/draft-ietf-bfd-multipoint-14.txt>
On that aspect, does this document really update 7880 as the header
says? The Introduction only refers to 5880 and it is not clear in
the body of the Document what effectively impacts 7880. The only
thing I saw is the addition of new session types but that does not
require an update in my opinion. Could you clarify?
GIM>> Yes, you'correct, the only connection to RFC 7880 are the new
values of bfd.sessionType. The proposal to list RFC 7880 as updated
by this specification was inteded to address Errata to RFC 7880
<https://www.rfc-editor.org/errata_search.php?rfc=7880>.
I am not sure how this document relates to or addresses the errata. So I
still think it does not update 7880.
i.e. existence of a path between the sender and the receiver.
do you mean "forwarding path"?
GIM>> Yes. Updated to
i.e. existence of a forwarding path between the sender and the receiver
thx
Section 2. and Section 3. seem a bit redundant. They both state the
same thing but from a different angle. Not critical.
Although this document describes a single head and a set of tails
spanned by a single multipoint path, the protocol is capable of
supporting (and discriminating between) more than one multipoint
path
at both heads and tails.
There is no text to describe how one could achieve that. Wouldn't it
be worth adding some?
GIM>> The question of applicability of this specification to mp2mp
scenario came up at BIER WG meeting in London. Perhaps we can say the
these questions are ouside the scope of this document and discuss
whether there are interested to work on mp2mp case as a separete document?
I don't read this part of the document as talking about mp2mp but rather
as talking about multiple p2mp trees.
Point-to-point sessions, as described in [RFC5880], are of type
PointToPoint.
Does this really fit in Section 4.2 which looks to be about the
mpBFD session model.
GIM>> I think that this short reminder is helpful to explain why this
document adds value PointToPoint, section 4.4.1, to the defined in RFC
7880 bfd.sessionType variable.
Well, I would move the text to 4.4.1 then, but not critical.
Sessions of type MultipointHead MUST NOT send BFD control packets
with the State field being set to INIT, and MUST be ignored on
receipt.
English is not my native language but I wonder if this really says
what you want. It seems that "Sessions" is the subject of "MUST be
ignored" while I think it's the packets which are the intended
subject. So I'd write:
and those packets MUST be ignored on receipt.
You chose to ignore that one or simply missed it?
Because there is no three-way handshake in Multipoint BFD, a newly
started head (that does not have any previous state information
available) SHOULD start with bfd.SessionState set to Down and with
bfd.RequiredMinRxInterval set to zero in the MultipointHead session.
To shut down a multipoint session a head MUST administratively set
bfd.SessionState in the MultipointHead session to either Down or
AdminDown and SHOULD set bfd.RequiredMinRxInterval to zero. The
In both these paragraphs one could read that the head "SHOULD set
bfd.RequiredMinRxInterval to zero" while 4.4.2 says MUST.
Clarification needed?
GIM>> Section 4.4.2 mandates initialization of
bfd.RequiredMinRxInterval and, I think, applies to the first paragraph
you've pointed. Would the following be clear and consistent:
Because there is no three-way handshake in Multipoint BFD, a newly
started head (that does not have any previous state information
available) SHOULD start with bfd.SessionState set to Down and
bfd.RequiredMinRxInterval _MUST be_ set to zero in the
MultipointHead session.
The second paragraph describes manipulation with the value of
bfd.RequiredMinRxInterval which process, as noted in section 4.10, "is
outside the scope of this document". That may be reason to use SHOULD
and not MUST.
Yes, i'd live with that. But then you might also want to clarify in 4.4.2.:
OLD:
This variable MUST be set to 0 for session type MultipointHead.
NEW:
This variable MUST be initialized to 0 for session type
MultipointHead.
s/M, P bit/M and P bits/
GIM>> Thanks, done.
---
