Re: [mpls] draft-fang-mpls-label-forwarding-no-swap - how much does it really save?

Andrew Qu <> Mon, 27 July 2015 18:56 UTC

Return-Path: <>
Received: from localhost ( []) by (Postfix) with ESMTP id B8E7F1B3212 for <>; Mon, 27 Jul 2015 11:56:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: 0.667
X-Spam-Status: No, score=0.667 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, IP_NOT_FRIENDLY=0.334, MIME_BAD_LINEBREAK=0.5, MIME_BASE64_TEXT=1.741, T_RP_MATCHES_RCVD=-0.01, UNPARSEABLE_RELAY=0.001] autolearn=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id XSZdQxfOGnO2 for <>; Mon, 27 Jul 2015 11:56:36 -0700 (PDT)
Received: from ( []) by (Postfix) with ESMTP id E9A331ACCEB for <>; Mon, 27 Jul 2015 11:56:22 -0700 (PDT)
X-Listener-Flag: 11101
Received: from [(] by (envelope-from <>) (Cellopoint E-mail Firewall v3.9.12 Build 0312 with TLS) with ESMTP id 1758562138; Mon, 27 Jul 2015 14:54:54 -0500
Received: from ( by ( with Microsoft SMTP Server (TLS) id; Mon, 27 Jul 2015 14:56:09 -0400
Received: from ([fe80::2898:86df:e627:42ee]) by ([fe80::69ba:ee95:46f7:711d%10]) with mapi id 14.03.0181.006; Mon, 27 Jul 2015 14:56:08 -0400
From: Andrew Qu <>
To: "" <>, Alexander Vainshtein <>, Robert Raszuk <>, Shahram Davari <>
Thread-Topic: [mpls] draft-fang-mpls-label-forwarding-no-swap - how much does it really save?
Thread-Index: AQHQxdE8jUti3orAQEqCOvm3DZDbXZ3qQZhggAVnVgCAAAFBUA==
Date: Mon, 27 Jul 2015 18:56:07 +0000
Message-ID: <>
References: <> <> <>
In-Reply-To: <>
Accept-Language: en-US
Content-Language: en-US
x-originating-ip: []
Content-Type: multipart/alternative; boundary="_000_EA360A7AB9D90D4B9E9173B6D27C371EE3F624BEMTKMBS61N1media_"
MIME-Version: 1.0
X-TM-AS-Product-Ver: SMEX-
X-TM-AS-Result: No--29.665400-8.000000-31
X-TM-AS-User-Approved-Sender: No
X-TM-AS-User-Blocked-Sender: No
Archived-At: <>
Cc: "" <>
Subject: Re: [mpls] draft-fang-mpls-label-forwarding-no-swap - how much does it really save?
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Multi-Protocol Label Switching WG <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Mon, 27 Jul 2015 18:56:39 -0000

Hi Stewart,

From: Stewart Bryant []
Sent: Monday, July 27, 2015 4:30 PM
To: Andrew Qu; Alexander Vainshtein; Robert Raszuk; Shahram Davari
Subject: Re: [mpls] draft-fang-mpls-label-forwarding-no-swap - how much does it really save?

On 24/07/2015 09:06, Andrew Qu wrote:
Hi Sasha,

My comments in lines...

From: mpls [] On Behalf Of Alexander Vainshtein
Sent: Friday, July 24, 2015 7:26 AM
To: Robert Raszuk; Shahram Davari
Subject: Re: [mpls] draft-fang-mpls-label-forwarding-no-swap - how much does it really save?

Hi all,

Please note that CONTINUE in SPRING is:

1. A DP-agnostic primitive

2. *Implemented* as SWAP instruction with MPLS DP.

[Andrew]  You are contradicting with 1) and 2).
What Sasha says is correct
[Andrew]  I don't think so,  points are listed below.

With "CONTINUE" defined in SR draft,  for NEW ASIC to be designed,

Why should I still use "SWAP" to achieve it?
.. because you need to be backwards compatible with
the deployed equipment and deployed networks.

[Andrew] "backwards compatible" is right consideration, hence I don't object
                to mention 2) as an option, so people read such RFC will be clear about
how to support that feature using existing device.  but "NO_SWAP" is another option
to support such primitive called "CONTINUE", to be REALLY "DP-agnostic", they you must
exhaust all options to also mention "NO_SWAP" simply from being logically consistent
"agnostic", let alone there are technical benefits to mention.

Adding "NO_SWAP" for new/future generation DEVICE does NOT break backward capability.

Rule number 1 says "DP-agnostic", but at the same time you purposely leave one option out but
Insists one more two options, then it is contradicting, simply from logic point of view.

If that is not a requirement in your world, then I suggest
that you get a new ethertype and call it MPLS2

[Andrew] You mis-understood my points,  I am fully aware of backward compatibility requirement by asking
adding NO_SWAP primitive.
                Please think again if I need MPLS2 or not,  I don't think so.


- Stewart