Re: [alto] Mandatory and optional in alto protocol

Robert Varga <robert.varga@pantheon.sk> Wed, 24 August 2011 12:32 UTC

Return-Path: <robert.varga@pantheon.sk>
X-Original-To: alto@ietfa.amsl.com
Delivered-To: alto@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id B0A5221F8785 for <alto@ietfa.amsl.com>; Wed, 24 Aug 2011 05:32:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.694
X-Spam-Level:
X-Spam-Status: No, score=-0.694 tagged_above=-999 required=5 tests=[BAYES_00=-2.599, HELO_EQ_SK=1.35, HOST_EQ_SK=0.555]
Received: from mail.ietf.org ([12.22.58.30]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ha1ggrh3hz5i for <alto@ietfa.amsl.com>; Wed, 24 Aug 2011 05:32:59 -0700 (PDT)
Received: from amalka.pantheon.sk (amalka.pantheon.sk [81.89.59.174]) by ietfa.amsl.com (Postfix) with ESMTP id 1D00921F8569 for <alto@ietf.org>; Wed, 24 Aug 2011 05:32:58 -0700 (PDT)
Received: from [10.0.2.15] (nitebug.pantheon.local [172.16.4.171]) by amalka.pantheon.sk (Postfix) with ESMTPSA id 66B6D20258 for <alto@ietf.org>; Wed, 24 Aug 2011 14:34:08 +0200 (CEST)
Message-ID: <4E54EFBE.10807@pantheon.sk>
Date: Wed, 24 Aug 2011 14:34:06 +0200
From: Robert Varga <robert.varga@pantheon.sk>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:6.0) Gecko/20110812 Thunderbird/6.0
MIME-Version: 1.0
To: alto@ietf.org
References: <CA5EC1A5.1D7F7%manbhard@cisco.com>
In-Reply-To: <CA5EC1A5.1D7F7%manbhard@cisco.com>
Content-Type: text/plain; charset="ISO-8859-1"; format="flowed"
Content-Transfer-Encoding: 7bit
Subject: Re: [alto] Mandatory and optional in alto protocol
X-BeenThere: alto@ietf.org
X-Mailman-Version: 2.1.12
Precedence: list
List-Id: "Application-Layer Traffic Optimization \(alto\) WG mailing list" <alto.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/alto>, <mailto:alto-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/alto>
List-Post: <mailto:alto@ietf.org>
List-Help: <mailto:alto-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/alto>, <mailto:alto-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 24 Aug 2011 12:32:59 -0000

Hi,

On 08/03/11 18:13, manbhard wrote:
> I agree with the opinion in the meeting that all services should be
> optional. A client should use an ALTO server that advertises the service it
> needs via the directory service. As mentioned below, downloading full
> network and cost maps (and maintaining them) might be a disincentive for
> certain clients. Similarly it might be a disincentive to servers run by SPs
> who do not want to export their prefix list, and if forced to will do quite
> a lot of aggregation which undermines the effectiveness of ALTO.
>
> An approach that would satisfy the client and server constraints above would
> be all the services that only provide information (properties, costs) for
> that client only.

I second this opinion. In the pre-08 revisions of the draft, when 
services resided on fixed URIs, the notion of mandatory and optional 
services made sense, in that the client could make a blind query as soon 
as it (somehow) got a server address.

With the latest draft, though, the client needs to find a service which 
is able to execute its query. Since current ALTO discovery deals only 
with servers, I think it is safe to assume the result of the discovery 
will be an URI pointing to a server IRD. From the IRD the client needs 
to select the appropriate service, and that service may reside on a 
completely different host than the one which hosts the IRD.

This has the interesting consequence that a 'server' as resolved from 
the discovery service may actually be a set of logical servers, tied 
together by the IRD. Each of them is an ALTO server in the sense they 
speak the protocol, but in my mind that does not mean that each of them 
has to provide a Network Map service -- or anything beyond the service 
advertised by the IRD.

To sum this up: I think an 'RFC-compliant ALTO server' should be 
required to provide the IRD service. All other services should be purely 
optional. That way each deployment site is free to combine multiple ALTO 
server implementations to provide a mix of services which are most 
appropriate for that particular site, but at the same time 
implementations are not limited in what services they can provide.

Bye,
Robert