IETF Logo Mail Archive

Re: [babel] ECMP routing

"Acee Lindem (acee)" <acee@cisco.com> Mon, 15 October 2018 18:15 UTC

Return-Path: <acee@cisco.com>
X-Original-To: babel@ietfa.amsl.com
Delivered-To: babel@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 685B3130DC5 for <babel@ietfa.amsl.com>; Mon, 15 Oct 2018 11:15:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -14.502
X-Spam-Level:
X-Spam-Status: No, score=-14.502 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIMWL_WL_MED=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, USER_IN_DEF_DKIM_WL=-7.5] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (1024-bit key) header.d=cisco.com
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 odws4u8yXr0W for <babel@ietfa.amsl.com>; Mon, 15 Oct 2018 11:15:05 -0700 (PDT)
Received: from alln-iport-5.cisco.com (alln-iport-5.cisco.com [173.37.142.92]) (using TLSv1.2 with cipher DHE-RSA-SEED-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 52019127133 for <babel@ietf.org>; Mon, 15 Oct 2018 11:15:05 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=2366; q=dns/txt; s=iport; t=1539627305; x=1540836905; h=from:to:cc:subject:date:message-id:references: in-reply-to:content-id:content-transfer-encoding: mime-version; bh=PjKhE/dZvUT8BjC+oWx7BkfCBuzb3hyoVeYxUKVrOBo=; b=K0cxUGZ4a7OTVjm5JZP9jJzP8cvUZxtr8fAFjXkEu25gX6TJF6hJdVhj chR5AUjq8rl/BOd4AfAVTI+glAX14wDZlhN+Q05LYY1lAr1qV3YuOAuuY r2/y63lf4bp/w6l/gKrHjfDsPExsjBb09IdZrKrRw4wVjCAp80DTXQZLP g=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0AHAADV2MRb/5FdJa1jGgEBAQEBAgEBAQEHAgEBAQGBUgQBAQEBCwGCA4FlKAqDa5RBgWgliHSOE4F6CwEBhGwCF4RGITUMDQEDAQECAQECbSiFOgEEASMRRRACAQYCGgImAgICHxEVEAIEAQ0FgyCBagMNCItJm02BLoR3gjkNghiBC4pBF4IAgRInDBOCTIJWggEnF4JsMYImAohklRguCQKNNoMjF5AejTmITQIRFIEmHgE2gVVwFWUBgkGQVm+JG4EugR8BAQ
X-IronPort-AV: E=Sophos;i="5.54,385,1534809600"; d="scan'208";a="185682720"
Received: from rcdn-core-9.cisco.com ([173.37.93.145]) by alln-iport-5.cisco.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 15 Oct 2018 18:15:00 +0000
Received: from XCH-RTP-012.cisco.com (xch-rtp-012.cisco.com [64.101.220.152]) by rcdn-core-9.cisco.com (8.15.2/8.15.2) with ESMTPS id w9FIExta028478 (version=TLSv1.2 cipher=AES256-SHA bits=256 verify=FAIL); Mon, 15 Oct 2018 18:15:00 GMT
Received: from xch-rtp-015.cisco.com (64.101.220.155) by XCH-RTP-012.cisco.com (64.101.220.152) with Microsoft SMTP Server (TLS) id 15.0.1395.4; Mon, 15 Oct 2018 14:14:22 -0400
Received: from xch-rtp-015.cisco.com ([64.101.220.155]) by XCH-RTP-015.cisco.com ([64.101.220.155]) with mapi id 15.00.1395.000; Mon, 15 Oct 2018 14:14:22 -0400
From: "Acee Lindem (acee)" <acee@cisco.com>
To: "7riw77@gmail.com" <7riw77@gmail.com>, 'Juliusz Chroboczek' <jch@irif.fr>
CC: "'STARK, BARBARA H'" <bs7652@att.com>, "babel@ietf.org" <babel@ietf.org>
Thread-Topic: [babel] ECMP routing
Thread-Index: AQHUYuor2J+RO1QbtUCKlIOnaApJpaUfhH6AgAAM5oCAAEhLAIAAxkaA
Date: Mon, 15 Oct 2018 18:14:22 +0000
Message-ID: <6621201C-1926-4CEF-8688-4F886844F640@cisco.com>
References: <2D09D61DDFA73D4C884805CC7865E6114DEEA160@GAALPA1MSGUSRBF.ITServices.sbc.com> <87a7nink9j.wl-jch@irif.fr> <088901d46403$a2e4a540$e8adefc0$@gmail.com> <87r2gsi3p6.wl-jch@irif.fr> <006a01d4642e$3b8b86e0$b2a294a0$@gmail.com>
In-Reply-To: <006a01d4642e$3b8b86e0$b2a294a0$@gmail.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach:
X-MS-TNEF-Correlator:
x-ms-exchange-messagesentrepresentingtype: 1
x-ms-exchange-transport-fromentityheader: Hosted
x-originating-ip: [10.116.152.197]
Content-Type: text/plain; charset="utf-8"
Content-ID: <2DADF3E538E60043B5DDCBCD21FB90F5@emea.cisco.com>
Content-Transfer-Encoding: base64
MIME-Version: 1.0
X-Outbound-SMTP-Client: 64.101.220.152, xch-rtp-012.cisco.com
X-Outbound-Node: rcdn-core-9.cisco.com
Archived-At: <https://mailarchive.ietf.org/arch/msg/babel/oCLDKKwTEaB14F_qzblXgSivlaA>
Subject: Re: [babel] ECMP routing
X-BeenThere: babel@ietf.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: "A list for discussion of the Babel Routing Protocol." <babel.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/babel>, <mailto:babel-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/babel/>
List-Post: <mailto:babel@ietf.org>
List-Help: <mailto:babel-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/babel>, <mailto:babel-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 15 Oct 2018 18:15:08 -0000

Hi Russ, Juliusz, 

I agree totally with Russ - I really don't see how ECMP could be more of an issue in Babel. In fact, it could lessen the impact of dynamic metric oscillations if you were to support CECMP (Close Enough Cost Multi-Path) with some % of variance allow between multipath routes. 
Thanks,
Acee

On 10/14/18, 10:24 PM, "7riw77@gmail.com" <7riw77@gmail.com> wrote:

    
    > ECMP only improves throughput, and in many networks you're not limited by
    > throughput -- you want to minimise latency, jitter, or packet loss.  And I'm not
    > sufficiently familiar with the literature to know whether the effects of per-flow
    > ECMP on metrics other than throughput have been studied in detail.
    
    This is a discussion we need to undertake over something to drink. 😊 I'll just say that ECMP can, in fact, be used to reduce jitter and latency, as well as throughput -- in fact, the three are sometimes intimately related. If the "cost" in ECMP included delay and/or jitter, ECMP would explicitly adjust for links equal in this regard -- or at least attempt to. Reality may be different than theory in this neighborhood. 
    
    > I'm not saying ECMP cannot be done right, what I'm saying is that doing ECMP
    > right in the wide variety of networks that Babel has been designed to handle
    > with no human intervention is an open research problem, and therefore one
    > that it would be premature to standardise.
    
    My larger concern would be positive feedback loops, rather than specific issues with optimizing for a single metric. 
    
    I'm okay with leaving this as a research project, though, for the moment, in BABEL.
    
    😊 /r

v2.46.0 | Report a Bug | By Email | System Status