Re: [Idnet] [Nmlrg] Uses case for ML -- automating traffic prioritisation

grenville armitage <> Thu, 30 March 2017 00:09 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id E031E124D68 for <>; Wed, 29 Mar 2017 17:09:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: -4.2
X-Spam-Status: No, score=-4.2 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, RP_MATCHES_RCVD=-0.001, URIBL_BLOCKED=0.001] autolearn=ham autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id S0_069zYIvGv for <>; Wed, 29 Mar 2017 17:09:29 -0700 (PDT)
Received: from ( []) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 39206126D05 for <>; Wed, 29 Mar 2017 17:09:28 -0700 (PDT)
Received: from [] ( []) by (8.14.3/8.14.3) with ESMTP id v2U09FvI010405 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES128-SHA bits=128 verify=NO); Thu, 30 Mar 2017 11:09:15 +1100
To: "Liubing (Leo)" <>
References: <> <>
From: grenville armitage <>
Message-ID: <>
Date: Thu, 30 Mar 2017 11:09:15 +1100
User-Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:38.0) Gecko/20100101 Thunderbird/38.7.1
MIME-Version: 1.0
In-Reply-To: <>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <>
Subject: Re: [Idnet] [Nmlrg] Uses case for ML -- automating traffic prioritisation
X-Mailman-Version: 2.1.22
Precedence: list
List-Id: "The IDNet \(Intelligence-Defined Network\) " <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 30 Mar 2017 00:09:31 -0000

(Belatedly following up on a question I failed to answer on nmrlg@ last year... I figure idnet@ is the appropriate place for a mea culpa and followup...)

On 10/30/2015 13:50, Liubing (Leo) wrote:
> Hi Grenville,
> Thanks for sharing your use case. I think it is a valuable use case in networking.
> You mentioned the PoC:
>> A proof-of-concept implemented in OpenWRT:
> Does it imply that your study was specific to home networks scenarios, and the algorithms were specifically designed to adapt those low performance embedded devices?

Our specific demonstration use-case was home networking, yes. But I don't recall that we specifically made algorithm choices for low performance embedded devices. Of course, when a Professor says "my group" did something, it means the student(s) at the time were the ones who really understood the finer details :-)

> I personally expect your study was not limited in home network scenarios, because the flow classification is also an important use case in carrier networks.

Agreed. My particular focus was on simplifying the remote configuration of home gateways so they could automagically prioritise interactive traffic without being pre-loaded with N*thousands of static QoS rules for all possible VoIP, game, etc, services. But the idea generalises.