Re: [Dime] DOIC endpoint behavior - Agent Overload Report handling

Jouni Korhonen <jouni.nospam@gmail.com> Fri, 07 February 2014 09:45 UTC

Return-Path: <jouni.nospam@gmail.com>
X-Original-To: dime@ietfa.amsl.com
Delivered-To: dime@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 88A781A05E5 for <dime@ietfa.amsl.com>; Fri, 7 Feb 2014 01:45:13 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2
X-Spam-Level:
X-Spam-Status: No, score=-2 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, SPF_PASS=-0.001] autolearn=ham
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 31hmE4BjMUwW for <dime@ietfa.amsl.com>; Fri, 7 Feb 2014 01:45:11 -0800 (PST)
Received: from mail-lb0-x22c.google.com (mail-lb0-x22c.google.com [IPv6:2a00:1450:4010:c04::22c]) by ietfa.amsl.com (Postfix) with ESMTP id 282AB1A05E6 for <dime@ietf.org>; Fri, 7 Feb 2014 01:45:10 -0800 (PST)
Received: by mail-lb0-f172.google.com with SMTP id c11so2439826lbj.17 for <dime@ietf.org>; Fri, 07 Feb 2014 01:45:09 -0800 (PST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=content-type:mime-version:subject:from:in-reply-to:date:cc :content-transfer-encoding:message-id:references:to; bh=78KdxZ+/vNKH2Tr81vNr38ctxvvOXGevofIWhNKqfQE=; b=AfG+ACTcFHXO7em60htF53nhfOPb4qeypsbtuBl0FOSoNsrup2cIi+C36YYV56BmP1 eGwkcyjeINmwdDyUdwBp/WbCEfyYTGizrsxm9drbWcy/thDpCfp4OZxR6JjKp4aKhBfc x3bXAh3iitZ9oJoLKW9NelOTZyT0Pw23ky/9iFrsVpg7xsxxTb5qlPvIj5xZIo+44ysR kdHJrnrgIUypMrsKSxa0/XXBBv/sEPdl94BQpbG1MDh4mKW8F3MuYBnOfchqD4oFbxdv TfwsqZyDurEYTfsLvu1RDnmUk2O5zODIeoO+3dQhM4G5jFNosR5j1xeupH4gGd6B/MaB EYKQ==
X-Received: by 10.112.204.104 with SMTP id kx8mr8852483lbc.12.1391766309074; Fri, 07 Feb 2014 01:45:09 -0800 (PST)
Received: from ?IPv6:2001:6e8:480:60:5cf:580e:21dd:449d? ([2001:6e8:480:60:5cf:580e:21dd:449d]) by mx.google.com with ESMTPSA id yq2sm5946300lab.3.2014.02.07.01.45.07 for <multiple recipients> (version=TLSv1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Fri, 07 Feb 2014 01:45:07 -0800 (PST)
Content-Type: text/plain; charset="iso-8859-1"
Mime-Version: 1.0 (Mac OS X Mail 6.6 \(1510\))
From: Jouni Korhonen <jouni.nospam@gmail.com>
In-Reply-To: <52F0EB22.8090804@usdonovans.com>
Date: Fri, 07 Feb 2014 11:45:08 +0200
Content-Transfer-Encoding: quoted-printable
Message-Id: <80E5E47E-9AD1-432F-9CA5-BA7D3D73D894@gmail.com>
References: <52F0EB22.8090804@usdonovans.com>
To: Steve Donovan <srdonovan@usdonovans.com>
X-Mailer: Apple Mail (2.1510)
Cc: "dime@ietf.org" <dime@ietf.org>
Subject: Re: [Dime] DOIC endpoint behavior - Agent Overload Report handling
X-BeenThere: dime@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: Diameter Maintanence and Extentions Working Group <dime.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/dime>, <mailto:dime-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/dime/>
List-Post: <mailto:dime@ietf.org>
List-Help: <mailto:dime-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/dime>, <mailto:dime-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 07 Feb 2014 09:45:13 -0000

Comments inline:

On Feb 4, 2014, at 3:29 PM, Steve Donovan <srdonovan@usdonovans.com> wrote:

> The following wording is proposed to be added to section 5.5 on Overload Report Processing.
> 
> This goes along with the proposed wording for agent involvement in capability exchange.
> 
> The most important piece of behavior proposed is for the case when the agent is a back-to-back DOIC association agent as illustrated here:
> 
>   DOC node <--downstream DOIC association--> DOC agent <--upstream DOIC association--> DOC  node
> 
> In this case, it is proposed that when the DOC agent receives a host or realm overload report the DOC agent simply passes it on without taking further action.  The goal being to do the overload abatement as close to the source of the request as possible.
> 
> The text isn't perfect yet, but I wanted to get the basic behavior proposal across.
> 
> Regards,
> 
> Steve
> 
> 5.5.4 Agent Considerations
> 
> As discussed in section x.x.x, agents can take on the role of reporting node and reacting node.
> 
> Agent as reporting node
> 
> A DOC agent will take on the role of a reporting node any time that there is a downstream DOIC association.
> 
> For the report types supporting in this document, an agent should only originate an overload report in two cases:
> 
> - When the agent is acting as overload authority for a set of servers.  In this case, requests are sent to the destination-host of the agent and the agent is responsible for reporting on the overload state when the server farm represented by that destination-host value is entering an overloaded condition.

Makes sense and is inline what I originally thought.

> - When the agent is acting as overload authority for a realm.  In this case, the agent is responsible for inserting realm based overload reports into answer messages from servers in that realm.

Ok.

> Agent as reacting node
> 
> An agent will take on the role of a reacting node whenever there is an upstream DOIC association.  In this case, the agent will be reacting to host overload reports.
> 
> The behavior of the agent acting as a reacting node depends on whether or not there is a downstream association.
> 
> If there is a downstream DOIC association then the DOC agent should pass any overload report on to the downstream Diameter node without taking any further action.
> 
>   Note: This is based on the assumption that it is best to handle the overload instance as close to the source of the Diameter transaction as possible.
> 
>   Note: This is not a must because there could be operator specific policies that result in handling of the overload condition in the agent.

I kind of agree Ulrich's concern here. If the two
associations have different supported features some
modifications may be needed on the OLRs. However, I
also think the "the DOC agent should pass any" is
enough to cover the case Ulrich points at, assuming
we note the case where downstream and upstream
associations have different capabilities.

> 
> If there is no downstream DOIC association then the DOC agent is responsible for handling the overload condition.  In this case the DOC agent must throttle requests targeted for the host or realm indicated in the overload report.  The method used should be consistent with the considerations outlined in section 5.5.2.

Ok.

- Jouni

> 
> When a request message is selected for throttling, the DOC agent must generate the appropriate error response message.
> 
> Editors note: the error message to be sent is TBD.
> _______________________________________________
> DiME mailing list
> DiME@ietf.org
> https://www.ietf.org/mailman/listinfo/dime