Re: [mpls] thought about the ADI name

Loa Andersson <loa@pi.nu> Sun, 26 December 2021 07:46 UTC

Return-Path: <loa@pi.nu>
X-Original-To: mpls@ietfa.amsl.com
Delivered-To: mpls@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 068C33A0920 for <mpls@ietfa.amsl.com>; Sat, 25 Dec 2021 23:46:53 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.749
X-Spam-Level:
X-Spam-Status: No, score=-3.749 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, NICE_REPLY_A=-1.852, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id QhQVuFg288iC for <mpls@ietfa.amsl.com>; Sat, 25 Dec 2021 23:46:48 -0800 (PST)
Received: from pipi.pi.nu (pipi.pi.nu [83.168.239.141]) (using TLSv1.1 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 319233A091D for <mpls@ietf.org>; Sat, 25 Dec 2021 23:46:46 -0800 (PST)
Received: from [192.168.1.98] (unknown [122.2.110.30]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: loa@pi.nu) by pipi.pi.nu (Postfix) with ESMTPSA id 28ADF365946; Sun, 26 Dec 2021 08:46:40 +0100 (CET)
Message-ID: <67d79e03-8573-87d4-de06-fcc85ffb7062@pi.nu>
Date: Sun, 26 Dec 2021 15:46:34 +0800
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Thunderbird/91.4.1
Content-Language: en-CA
To: Greg Mirsky <gregimirsky@gmail.com>
Cc: Haoyu Song <haoyu.song@futurewei.com>, "mpls@ietf.org" <mpls@ietf.org>
References: <9e553dc9-34d1-44d5-1e33-41e4a3372597@pi.nu> <BY3PR13MB47870B6D4018A330A4EDEA279A7D9@BY3PR13MB4787.namprd13.prod.outlook.com> <dc108dd7-dad6-1d94-268e-d54bdda48719@pi.nu> <BY3PR13MB4787A33AE3806444F48086449A7E9@BY3PR13MB4787.namprd13.prod.outlook.com> <CA+RyBmWc-YGvYBFZeyQ0=oaubNx8VwvaQLJTtsf9YqUNqp_VBA@mail.gmail.com> <c673797e-19e7-b636-acdf-2a0667db9919@pi.nu> <CA+RyBmVuG7sYVt6D_xcN8x7pcGacA+pvT4BYOsX3sW1ToToXhw@mail.gmail.com>
From: Loa Andersson <loa@pi.nu>
In-Reply-To: <CA+RyBmVuG7sYVt6D_xcN8x7pcGacA+pvT4BYOsX3sW1ToToXhw@mail.gmail.com>
Content-Type: text/plain; charset="UTF-8"; format="flowed"
Content-Transfer-Encoding: 8bit
Archived-At: <https://mailarchive.ietf.org/arch/msg/mpls/ZzqdnH4cWgauUbDtuDczqt81m4k>
Subject: Re: [mpls] thought about the ADI name
X-BeenThere: mpls@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Multi-Protocol Label Switching WG <mpls.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/mpls>, <mailto:mpls-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/mpls/>
List-Post: <mailto:mpls@ietf.org>
List-Help: <mailto:mpls-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/mpls>, <mailto:mpls-request@ietf.org?subject=subscribe>
X-List-Received-Date: Sun, 26 Dec 2021 07:46:53 -0000

Greg, et.al.,

Happy Holidays!

Online please!

On 26/12/2021 03:50, Greg Mirsky wrote:
> Hi Loa,
> Happy Holidays to All!
> Please find my answers in-line below under the GIM>> tag.
> 
> Regards,
> Greg
> 
> On Fri, Dec 24, 2021 at 9:33 AM Loa Andersson <loa@pi.nu 
> <mailto:loa@pi.nu>> wrote:
> 
>     Greg,
> 
>     Questions for clarification:
> 
>     - the abbreviation ADAI is that Ancillary Data and Action Indicator?
> 
> GIM>> Yes, that was my intention.

OK - we'll give Matthew and Stewart a chance to think about this.
> 
> 
>     - when you say "e2e is always at the BoS" do you intend to say that
>         this is PSD?
> 
> GIM>> Thank you for the question. I don't have a case for e2e using ISD. 
> Thus, I think that e2e is only PSD (but I am open to the discussion).
> 

hmmm - I have feeling that we are in a kind of agreement, but also that 
we get kind of circular.

For the sake of the argument let us say that e-2-e is always using PSD.
What would the value be to say it is at the BoS? In my world "at" could 
be interpreted as "either side of" and thus be incorrect. If we say that 
data that e-2-e relies on are "after" the BoS, that would be correct, 
but just repeting the definitin of PSD.
> 
>     Open DT,
> 
>     We have a lot of terms/abbreviations that has more or less the same
>     thing.
> 
>     Should we start an Abbreviation/Terminology page for normative
>     information on this?
> 
> GIM>> Wholeheartedly agree. Let's record all acronyms and abbreviations, 
> establish equivalency groups and try to converge on a single term for 
> each construct.

I came up with this as a starting point:

https://trac.ietf.org/trac/mpls/wiki/abbreviations

It is not linked to the Open DT wiki pafa, but I can do that.

/Loa


> 
> 
>     /Loa
> 
>     On 24/12/2021 02:24, Greg Mirsky wrote:
>      > Hi Haoyu, et al,
>      > I might be missing something important in the discussion of the
>      > processing ADI/ADAI. My impression is that e2e is always at the
>     BoS. Is
>      > that correct? Now, what do we do for HbH? As I understand it, you're
>      > considering ELI-style when ADAI is present in the stack and the node
>      > needs to locate it first. Do I understand the scenario you're
>     talking
>      > about correctly? But wouldn't processing be simpler if we mandate
>     that a
>      > HbH ADAI is at the top of the stack? Also, to optimize the length
>     of the
>      > stack, a transit node may be allowed to place the acted-upon HbH
>     ADAI
>      > below the transport label.
>      > What do you think? Is that completely off the rails?
>      >
>      > Happy Holidays to All!
>      >
>      > Regards,
>      > Greg
>      >
>      > On Thu, Dec 23, 2021 at 9:58 AM Haoyu Song
>     <haoyu.song@futurewei.com <mailto:haoyu.song@futurewei.com>
>      > <mailto:haoyu.song@futurewei.com
>     <mailto:haoyu.song@futurewei.com>>> wrote:
>      >
>      >     Hi Loa,
>      >
>      >     Let me try to explain a bit more. I have examples to back the
>      >     following claims which I can present in a future meeting.
>      >
>      >     So far all the existing header design follows a chain
>     structure. The
>      >     header are parsed linearly one by one from the start of the
>     packet,
>      >     until you reach a header that is considered the last one
>     concerned
>      >     by the forwarding plane. You know the presence of a specific
>     header
>      >     only when you reach it. In this design, the number of parser
>     states
>      >     as well as the time for parsing is both linear to the number of
>      >     headers scanned.
>      >
>      >     Now people may think using some extra indicators (i.e., a bitmap
>      >     with each bit indicating the presence of a header later in the
>      >     packet) may improve the parsing performance. To this we must
>     ask "in
>      >     what sense"?
>      >
>      >     We can consider two possible types of improvements. First is the
>      >     reduction of parsing states which can help to save the parser
>      >     resource (i.e., fewer nodes in the parser FSM); second is the
>      >     reduction of parsing cycles which can help to parse a packet
>     faster
>      >     (we have a fixed cost for parsing each header, no matter the
>     size of
>      >     it. E.g, each MPLS label is considered a header, the entire IPv6
>      >     header, excluding EHs, is also considered a header).
>      >
>      >     For the first one, if you start to actually draft the parsing
>     graph,
>      >     you will find the opposite results. In two different parsing
>     styles,
>      >     both requires more parser states than a simple header chain.
>      >
>      >     For the second one, we need to understand that the headers
>     concerned
>      >     by a forwarding plane need to all be parsed anyway. You can't
>     ignore
>      >     some headers in the middle because you will need to
>     reconstruct the
>      >     packet headers at the egress (a process also known as
>     deparsing). So
>      >     the extra indicator encoding doesn't help to improve the parsing
>      >     speed either.
>      >
>      >     There's a reason why so far all the headers are simply
>     organized as
>      >     a chain. It's the most efficient and straightforward way. My
>     study
>      >     is based on current switch ASICs and some NPs.  If people don't
>      >     believe me, then evidence (e.g., pseudo code or parsing
>     graph) needs
>      >     to be provided. Perhaps there are some different forwarding plane
>      >     designs which can play magic. We need to learn that before
>      >     introducing any new mechanism.
>      >
>      >     A caveat is that, an extra indicator for the presence of HBH
>     headers
>      >     might be useful in some cases. For example, on an LSP path
>     nodes, if
>      >     there's no HBH headers later in the packet, the parser can stop
>      >     further parsing immediately which can save some parsing cycles.
>      >     Even in this case, if the forwarding plane still requires to
>      >     continue parsing, this mechanism doesn’t help.
>      >
>      >     In general, we really just need to concern the packet header
>     buffer
>      >     (aka packet window) size. As long as all the headers
>     concerned by a
>      >     forwarding plane is within the buffer limit, the parsing cost
>     is a
>      >     negligible concern for a simple header chain. Other
>     mechanisms are
>      >     of no help at best and could be harmful at worst. Of course, I'd
>      >     like to see evidence if people think the other way.
>      >
>      >     Happy Holidays!
>      >
>      >     Best regards,
>      >     Haoyu
>      >
>      >     -----Original Message-----
>      >     From: Loa Andersson <loa@pi.nu <mailto:loa@pi.nu>
>     <mailto:loa@pi.nu <mailto:loa@pi.nu>>>
>      >     Sent: Wednesday, December 22, 2021 2:54 PM
>      >     To: Haoyu Song <haoyu.song@futurewei.com
>     <mailto:haoyu.song@futurewei.com>
>      >     <mailto:haoyu.song@futurewei.com
>     <mailto:haoyu.song@futurewei.com>>>; mpls@ietf.org
>     <mailto:mpls@ietf.org> <mailto:mpls@ietf.org <mailto:mpls@ietf.org>>
>      >     Subject: Re: [mpls] thought about the ADI name
>      >
>      >     Haoyu,
>      >
>      >     OK, I simply don't understand.
>      >
>      >     If you don't know what action you'll take, what good is it to
>     know
>      >     where to find the data?
>      >
>      >     It might be that this is not what you say, but that is what
>     if get
>      >     from your text below. Sorry if I'm misunderstanding.
>      >
>      >     /Loa
>      >
>      >     On 23/12/2021 03:25, Haoyu Song wrote:
>      >      > Hi Loa,
>      >      >
>      >      > In my opinion the ADI should only be used to indicate the
>      >     presence of AD.  E2E or HBH AD could be differentiated because in
>      >     some case it can help stop further parsing beyond ADI.  Other
>      >     information encoded in it won't help but complicate the parsing
>      >     process. I strongly suggest any such proposal should give a clear
>      >     presentation on why it's necessary and how it can help from
>     the view
>      >     of implementors, otherwise, we may end up with an over
>     complicated
>      >     design without tangible benefits.
>      >      >
>      >      > Best regards,
>      >      > Haoyu
>      >      >
>      >      > -----Original Message-----
>      >      > From: mpls <mpls-bounces@ietf.org
>     <mailto:mpls-bounces@ietf.org> <mailto:mpls-bounces@ietf.org
>     <mailto:mpls-bounces@ietf.org>>>
>      >     On Behalf Of Loa Andersson
>      >      > Sent: Sunday, December 19, 2021 8:32 AM
>      >      > To: mpls@ietf.org <mailto:mpls@ietf.org>
>     <mailto:mpls@ietf.org <mailto:mpls@ietf.org>>
>      >      > Subject: [mpls] thought about the ADI name
>      >      >
>      >      > Working Group,
>      >      >
>      >      > The MIAD Requirement Specification use the abbreviation
>     ADI, it
>      >     stands for Ancillary Data Indicator. Which is all nice and dandy.
>      >      >
>      >      > But isn't it he case  that the indicator gives us two
>     things, the
>      >     action to be performed and where to find the data needed,
>     i.e., an
>      >     Ancillary Data and Action indication (ADAI?).
>      >      >
>      >      > No I'm not suggesting that we change, but we should be
>     aware, and
>      >     it would be nice to have it mentioned somewhere.
>      >      >
>      >      > /Loa
>      >
>      >     --
>      >     Loa Andersson                        email: loa@pi.nu
>     <mailto:loa@pi.nu> <mailto:loa@pi.nu <mailto:loa@pi.nu>>
>      >     Senior MPLS Expert loa.pi.nu@gmail.com
>     <mailto:loa.pi.nu@gmail.com> <mailto:loa.pi.nu@gmail.com
>     <mailto:loa.pi.nu@gmail.com>>
>      >     Bronze Dragon Consulting             phone: +46 739 81 21 64
>      >     _______________________________________________
>      >     mpls mailing list
>      > mpls@ietf.org <mailto:mpls@ietf.org> <mailto:mpls@ietf.org
>     <mailto:mpls@ietf.org>>
>      > https://www.ietf.org/mailman/listinfo/mpls
>     <https://www.ietf.org/mailman/listinfo/mpls>
>      >     <https://www.ietf.org/mailman/listinfo/mpls
>     <https://www.ietf.org/mailman/listinfo/mpls>>
>      >
> 
>     -- 
>     Loa Andersson                        email: loa@pi.nu <mailto:loa@pi.nu>
>     Senior MPLS Expert loa.pi.nu@gmail.com <mailto:loa.pi.nu@gmail.com>
>     Bronze Dragon Consulting             phone: +46 739 81 21 64
> 

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