Re: WGLC for BFD Multipoint documents (last round)

[Greg Mirsky <[e-mail]>]

Hi Carlos,
thank you for your thorough review and detailed comments. Please find my
responses in-lined tagged GIM>>.

Regards,
Greg

On Thu, Feb 8, 2018 at 7:19 PM, Carlos Pignataro (cpignata) <
[e-mail]> wrote:

> Hi, Greg,
>
> When looking for this specific sentence, I got a chance to scan through
> the document a bit.
>
> It seems to me there are still a number of editorials and potentially
> non-editorials to be fixed.
>
> Looking at S4.8 only:
>
>    BFD packets received on tails for an IP multicast group MUST be
>    consumed by tails and MUST NOT be forwarded to receivers.  Session of
>    type MultipointTail MUST identify the packet as BFD with the help of
>    destination UDP port number "3784" on IP multipoint path.
>
> CMP: There are a number of “the” or “a[n]” articles missing.
>
> CMP: What is “with the help of”? What is "port number "3784" on IP
> multipoint path”?
>
GIM>> Agree, will remove the quotation marks around 3784. Proposed update:
OLD TEXT:
 Session of
   type MultipointTail MUST identify the packet as BFD with the help of
   destination UDP port number "3784" on IP multipoint path.
NEW TEXT:
Session of type MultipointTail MUST identify packet received on an IP
multipoint path
as BFD control packet if the destination UDP port value equals 3784.


>    For multipoint LSPs, when IP/UDP encapsulation of BFD control packets
>    is used, MultipointTail MUST use destination UDP port "3784" and
>    "127.0.0.0/8" range for IPv4 or "0:0:0:0:0:FFFF:7F00:0/104" range for
>    IPv6 ([RFC8029]).  Packets identified as BFD packets MUST be consumed
>    by MultipointTail and demultiplex as described in Section 4.13.2
> .
> CMP: The first sentence is very confusing (besides the “MUST use” that you
> already called out and that I agree).
> CMP: For example, is this the source or destination endpoint? Which
> address is this (i.e., Destination)?
>
GIM>> Since the sentence refers to MultipointTail, I conclude this refers
to the destination endpoint. As Reshad agreed to s/use/expect/
here's proposed update to the sentence:
OLD TEXT:
   For multipoint LSPs, when IP/UDP encapsulation of BFD control packets
   is used, MultipointTail MUST use destination UDP port "3784" and
   "127.0.0.0/8" range for IPv4 or "0:0:0:0:0:FFFF:7F00:0/104" range for
   IPv6 ([RFC8029]).
NEW TEXT:
   For multipoint LSPs, when IP/UDP encapsulation of BFD control packets
   is used, MultipointTail MUST expect destination UDP port 3784 and
   destination IP address in 127.0.0.0/8 range for IPv4 or
0:0:0:0:0:FFFF:7F00:0/104 range for
   IPv6 ([RFC8029]).


>    Use of other types of encapsulation for multipoint LSP is outside the
>    scope of this document.
>
> CMP: For BFD Control?
>
GIM>> Will add 'of the BFD control message over' and remove 'for' to result:
Use of other types of encapsulation of the BFD control message over
multipoint LSP  is outside the scope of this document.


> Also, checking the following section:
>
> 4.13.1.  Reception of BFD Control Packets
>
>    The following procedure replaces section 6.8.6 of [RFC5880].
> …
>       If bfd.SessionType is PointToPoint, update the Detection Time as
>       described in [RFC5880] section 6.8.4.  Otherwise, update the
>       Detection Time as described in Section 4.11 above.
>
> CMP: The actual set of citations is not clear. If this is a replacement to
> section 6.8.6 of [RFC5880], then why a citation like “[RFC5880] section
> 6.8.4"? Isn’t it implied that it is RFC 5880? Or conversely, if it is a
> replacement updating RFC 5880, how does “in Section 4.11 above” work when
> inserted in RFC 5880? Its ask relative :-)
>
> Thanks!
>
> —
> Carlos Pignataro, [e-mail]
>
> *“Sometimes I use big words that I do not fully understand, to make myself
> sound more photosynthesis."*
>
> On Feb 8, 2018, at 12:14 AM, [e-mail] wrote:
>
> Hi Reshad,
>
> thank you for your consideration. I've came across what looks as simple
> editorial change. Appreciate your comment.
>
> In the second paragraph of section 4.8 Packet consumption on tails the
> following
>
>   For multipoint LSPs, when IP/UDP encapsulation of BFD control packets
>
>   is used, MultipointTail MUST use destination UDP port "3784" ...
>
> reads akwardly because, I think, of 'MUST use'. I propose simple
> s/use/look for/ or s/use/expect/. What do you think? Is the text god as-is
> or minor editing may help?
>
>
> Regards,
>
> Greg Mirsky
>
>
> Sr. Standardization Expert
> 预研标准部/有线研究院/有线产品经营部 Standard Preresearch Dept./Wireline Product R&D
> Institute/Wireline Product Operation Division
>
>
>
> <9ae3e214c17d49ed935d87c674ba3ee2.jpg> <24242e5637af428891c4db731e7765
> ad.jpg>
> E: [e-mail]
> www.zte.com.cn
> Original Mail
> *Sender: *ReshadRahman(rrahman) <[e-mail]>
> *To: *gregory mirsky10211915;[e-mail] <[e-mail]>
> *Date: *2018/02/08 11:03
> *Subject: **Re: WGLC for BFD Multipoint documents (last round)*
>
> <Sorry for the delay>
>
> Hi Greg,
>
>
>
> I would go with normative SHOULD. What you proposed below is fine.
>
>
>
> Regards,
>
> Reshad.
>
>
>
>
>
> *From: *"[e-mail]" <[e-mail]>
> *Date: *Sunday, February 4, 2018 at 8:33 PM
> *To: *"Reshad Rahman (rrahman)" <[e-mail]>, "[e-mail]" <
> [e-mail]>
> *Subject: *Re: WGLC for BFD Multipoint documents (last round)
>
>
>
> Hi Reshad,
>
> you've said:
>
> Hi Greg,
>
>
>
> While OOB mechanism would improve security, my personal opinion is that we should try to improve security without requiring an OOB mechanism. I think we can add text to the security considerations to address the concerns below, e.g. A tail SHOULD prevent the number of MultipointTail sessions from exceeding the number of expected streams.  The concern expressed in b) cannot be fixed by what I proposed because of multiple streams.  So just preventing the number of MultipointTail sessions from exceeding the number of expected streams should be good enough.
>
>
>
> Regards,
>
> Reshad.
>
>
>
> If we make it normative SHOULD, i.e. s/should/SHOULD/ in the third
> recommendation to the implementers:
>
>      The implementation SHOULD have a reasonable upper bound on the
>
>       number of MultipointTail sessions that can be created, with the
>
>       upper bound potentially being computed based on the number of
>
>       multicast streams that the system is expecting.
>
>
>
> Or keep the lower case as it is consistent with the rest of the section,
> e.g. 'a MultipointTail session should not be created'?
>
>
>
> Kind regards,
>
> Greg Mirsky
>
>
>
> Sr. Standardization Expert
> 预研标准部/有线研究院/有线产品经营部 Standard Preresearch Dept./Wireline Product R&D
> Institute/Wireline Product Operation Division
>
>
>
> <image001.gif>
>
> <image002.gif>
> E: [e-mail]
> www.zte.com.cn
>
>
>
>
>
>