Re: [Bier] WGLC: draft-ietf-bier-php

Working Group, authors,

I don't think I saw a response on this.

/Loa

On 22/08/2020 13:57, Loa Andersson wrote:
> Working Group,
> 
> I've had time to read the draft a bit more in detail.
> 
> I'm convinced that this will work, and support progressing the draft.
> 
> I still think there are some minor description details that could be
> improved.
> 
> The draft says:
> 
>     One way to solve this problem is to use Penultimate Hop Popping (PHP)
>     so that the upstream BFR can pop the BIER header and send the payload
>     "natively" (note that the upstream BFR can be connected directly or
>     indirectly via any type of tunnel to the PE).  This is similar to
>     Multi-Protocol Label Switching (MPLS) PHP though it is the BIER
>     header that is popped.
> 
> This indicate that this is a "per node" decision, while in MPLS it is a
> per LSP decision.
> 
> I would think that it is possible for one node to be the penultimate
> node for some traffic but not for other traffic, even though the traffic 
> is sent to the same node, which can be both ultimate node, pen-ultimate
> node or transit node.
> 
> 
>       +-------+              +-------+          +-------+
>       |       |              |       |          |       |
> -a,b-+   A   +-----a',b-----+   B   +----a'----+  CPE1 |
>       |       |              |       |          |       |
>       +-------+              +---+---+          +-------+
>                                  |
>                                  b'
>                                  |
>                              +---+---+          +-------+
>                              |       |          |       |
>                              +   C   +----b'----+  CPE2 |
>                              |       |          |       |
>                              +---+---+          +-------+
> 
> 
> Two streams comes in to A from the left (a and b).
> 
> A is the penultimate node for traffic to to CPE1 (a) so the BIER header
> is popped.
> 
> A is not the  penultimate node for traffic to to CPE2 (b) so the BIER
> header is popped.
> 
> Both a' and b is sent to B.
> 
> B forward a' natively to CPE1.
> 
> B is the penultimate node for traffic to to CPE2 (b) so the BIER header
> is popped.
> 
> b' is forwarded to C and onto CPE2
> 
> Works just fine.
> 
> Question: If the PHP is a "per node decision" how is A stopped from 
> doing PHP for b?
> 
> /Loa
> 
> 
> 
> On 22/08/2020 05:20, Gyan Mishra wrote:
>>
>> I support advancing this critical draft forward.  I agree that the PHP 
>> feature traditionally used with MPLS and has been adopted by SR and 
>> now by BIER for similar reasons of being able to pop the BIER header 
>> similar to what is done with SR PHP concept and now extensible with 
>> BIER for Non-BRF capable PE's.
>>
>> BIER IPv6 is still being developed, however I think maybe BIERin6 
>> could take advantage of  PHP by popping the encapsulated L2 1/2 BIER 
>> shim for "Non-BFR" capable PE.  I think for BIERv6 as the BIER header 
>> is encoded in DOH it may be difficult to support PHP.
>>
>> Thanks
>>
>> Gyan
>>
>> On Wed, Aug 19, 2020 at 2:01 PM Mankamana Mishra (mankamis) 
>> <mankamis=40cisco.com@dmarc.ietf.org 
>> <mailto:40cisco.com@dmarc.ietf.org>> wrote:
>>
>>     Support moving this document forward. ____
>>
>>     __ __
>>
>>     *From: *BIER <bier-bounces@ietf.org <mailto:bier-bounces@ietf.org>>
>>     on behalf of "IJsbrand Wijnands (iwijnand)"
>>     <iwijnand=40cisco.com@dmarc.ietf.org
>>     <mailto:40cisco.com@dmarc.ietf.org>>
>>     *Date: *Monday, August 17, 2020 at 11:56 AM
>>     *To: *Greg Shepherd <gjshep@gmail.com <mailto:gjshep@gmail.com>>
>>     *Cc: *"bier@ietf.org <mailto:bier@ietf.org>" <bier@ietf.org
>>     <mailto:bier@ietf.org>>
>>     *Subject: *Re: [Bier] WGLC: draft-ietf-bier-php____
>>
>>     __ __
>>
>>     In support moving this draft forward, ____
>>
>>     __ __
>>
>>     Thx,____
>>
>>     __ __
>>
>>     Ice.____
>>
>>     Sent from my iPhone____
>>
>>
>>
>>     ____
>>
>>         On 13 Aug 2020, at 05:31, Greg Shepherd <gjshep@gmail.com
>>         <mailto:gjshep@gmail.com>> wrote:____
>>
>>
>>         Please read and respond to this thread. We had a WGLC for this
>>         draft previously, but had no response. Let’s do better. ____
>>
>>         __ __
>>
>>         https://datatracker.ietf.org/doc/draft-ietf-bier-php/____
>>
>>         __ __
>>
>>         Start of short LC timer. Respond by Aug 19..____
>>
>>         __ __
>>
>>         - Shep____
>>
>>         (Chairs)____
>>
>>         _______________________________________________
>>         BIER mailing list
>>         BIER@ietf.org <mailto:BIER@ietf.org>
>>         https://www.ietf.org/mailman/listinfo/bier____
>>
>>     _______________________________________________
>>     BIER mailing list
>>     BIER@ietf.org <mailto:BIER@ietf.org>
>>     https://www.ietf.org/mailman/listinfo/bier
>>
>>
>>
>> -- 
>>
>> <http://www.verizon.com/>
>>
>> *Gyan Mishra*
>>
>> /Network Solutions A//rchitect /
>>
>> /M 301 502-1347
>> 13101 Columbia Pike
>> /Silver Spring, MD
>>
>>
>>
>> _______________________________________________
>> BIER mailing list
>> BIER@ietf.org
>> https://www.ietf.org/mailman/listinfo/bier
>>
> 

-- 

Loa Andersson                        email: loa@pi.nu
Senior MPLS Expert                          loa.pi.nu@gmail.com
Bronze Dragon Consulting             phone: +46 739 81 21 64