Re: [Roll] Child count in parent selection

Chenyang JI <chenyang.ji@imt-atlantique.net> Mon, 06 November 2017 09:50 UTC

Return-Path: <chenyang.ji@imt-atlantique.net>
X-Original-To: roll@ietfa.amsl.com
Delivered-To: roll@ietfa.amsl.com
Received: from localhost (localhost [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 2916E13FC13 for <roll@ietfa.amsl.com>; Mon, 6 Nov 2017 01:50:51 -0800 (PST)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.001
X-Spam-Level:
X-Spam-Status: No, score=-2.001 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, SPF_PASS=-0.001] autolearn=ham autolearn_force=no
Authentication-Results: ietfa.amsl.com (amavisd-new); dkim=pass (1024-bit key) header.d=imt-atlantique.net
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 nW99Qa7C2A7C for <roll@ietfa.amsl.com>; Mon, 6 Nov 2017 01:50:48 -0800 (PST)
Received: from zproxy110.enst.fr (zproxy110.enst.fr [IPv6:2001:660:330f:2::c0]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 197DB13FB8C for <roll@ietf.org>; Mon, 6 Nov 2017 01:50:47 -0800 (PST)
Received: from localhost (localhost [IPv6:::1]) by zproxy110.enst.fr (Postfix) with ESMTP id 82C43808A3; Mon, 6 Nov 2017 10:50:45 +0100 (CET)
Received: from zproxy110.enst.fr ([IPv6:::1]) by localhost (zproxy110.enst.fr [IPv6:::1]) (amavisd-new, port 10032) with ESMTP id 1lb1ezRdDM5c; Mon, 6 Nov 2017 10:50:43 +0100 (CET)
Received: from localhost (localhost [IPv6:::1]) by zproxy110.enst.fr (Postfix) with ESMTP id DAD67808C8; Mon, 6 Nov 2017 10:50:43 +0100 (CET)
DKIM-Filter: OpenDKIM Filter v2.10.3 zproxy110.enst.fr DAD67808C8
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=imt-atlantique.net; s=6A0CDB44-C782-11E6-82EC-91BDBA474D24; t=1509961843; bh=FtDr+MVNeaElq5jOZB330lv2pXpJe7KPjfQo7Lh8Hvs=; h=Date:From:To:Message-ID:MIME-Version; b=g0EiPLn6TBMkSjIp7uF6Aose+NugVIt42MRTGTnX1w0/5gSAuCp3brTuWFpmRQzTM lsXFZghZpwLhUAcXfXFOAwiRLMNYvTryhlPoPgmZV5PpvW8jSyyIu44kYpkruZtfIS 2kqdwNdg/MAsACPsIcVVqstf8C9gpGSpCjX68y9Y=
X-Virus-Scanned: amavisd-new at zproxy110.enst.fr
Received: from zproxy110.enst.fr ([IPv6:::1]) by localhost (zproxy110.enst.fr [IPv6:::1]) (amavisd-new, port 10026) with ESMTP id I6lhXeIX3L6v; Mon, 6 Nov 2017 10:50:43 +0100 (CET)
Received: from zmail131.enst.fr (zmail131.enst.fr [137.194.2.203]) by zproxy110.enst.fr (Postfix) with ESMTP id B7739806D5; Mon, 6 Nov 2017 10:50:43 +0100 (CET)
Date: Mon, 06 Nov 2017 10:50:43 +0100
From: Chenyang JI <chenyang.ji@imt-atlantique.net>
To: Houjianqiang <houjianqiang@huawei.com>
Cc: roll <roll@ietf.org>
Message-ID: <272885943.5182366.1509961843464.JavaMail.zimbra@imt-atlantique.net>
In-Reply-To: <DD0A994E4D6B3F4080662703C8C7C086A8E6A6@DGGEMM506-MBS.china.huawei.com>
References: <DD0A994E4D6B3F4080662703C8C7C086A8C793@DGGEMM506-MBS.china.huawei.com> <544130816.4872551.1509729703443.JavaMail.zimbra@imt-atlantique.net> <DD0A994E4D6B3F4080662703C8C7C086A8E6A6@DGGEMM506-MBS.china.huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
X-Originating-IP: [2001:660:7301:51:c126:3f76:249e:b974]
X-Mailer: Zimbra 8.7.11_GA_1854 (ZimbraWebClient - FF56 (Linux)/8.7.11_GA_1854)
Thread-Topic: Child count in parent selection
Thread-Index: AQHTVMhD4PR5R6hXSke6aZFwTvCsNqMHBNlQDhNm1yE=
Archived-At: <https://mailarchive.ietf.org/arch/msg/roll/-y2rFjPREUWc651m-ishQQ5xW2M>
Subject: Re: [Roll] Child count in parent selection
X-BeenThere: roll@ietf.org
X-Mailman-Version: 2.1.22
Precedence: list
List-Id: Routing Over Low power and Lossy networks <roll.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/roll>, <mailto:roll-request@ietf.org?subject=unsubscribe>
List-Archive: <https://mailarchive.ietf.org/arch/browse/roll/>
List-Post: <mailto:roll@ietf.org>
List-Help: <mailto:roll-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/roll>, <mailto:roll-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 06 Nov 2017 09:50:51 -0000

Hello,

Thank you for the answer.

Best regards,
Chenyang Ji



----- Original Message -----
From: "Houjianqiang" <houjianqiang@huawei.com>
To: "Chenyang JI" <chenyang.ji@imt-atlantique.net>, "roll" <roll@ietf.org>
Sent: Monday, November 6, 2017 10:29:35 AM
Subject: RE: [Roll] Child count in parent selection

Hi Chenyang,

Yes, we want the 'P' flag of the CNC metric to be set in every DIO in non-storing mode.

In storing mode, DAO/NP-DAO can be used for child registration and de-registration, thus there is no need of a parent address in DIO. But in non-storing mode, DAO cannot be used for counting the number of children, so we use DIO(+ parent address) instead.

Regards,
Jianqiang Hou (Derek) 

-----Original Message-----
From: Chenyang JI [mailto:chenyang.ji@imt-atlantique.net] 
Sent: Saturday, November 04, 2017 1:22 AM
To: Houjianqiang (Derek) <houjianqiang@huawei.com>; roll <roll@ietf.org>
Subject: Re: [Roll] Child count in parent selection

Hello,

Thank you for the answer and I apologize for my late response.
Today I realized that you had a new version of draft draft-qasem-roll-rpl-load-balancing-02 and I also have some questions on this.
The description about LB-OF is:

 In LB-OF algorithm, the received DIO from the child node is counted by the preferred parent node. Each DIO contains the IP  address of the chosen preferred parent as detailed in section 4.3. Thus,for each received DIO, the node matches its own IP  address with the preferred parent IP address which is inserted in the DIO message, thenincrements the number of children by  ONE for this node if there is a matching.

And the format of DIO metric container object is:

        0                   1                   2                   3
        0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |   Flags     |P|     CNC       |    CNC_MAX    |               |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+               +
       |                                                               |
       +                                                               +
       |                                                               |
       +                        Parent Address                         +
       |                                                               |
       +                                               +-+-+-+-+-+-+-+-+
       |                                               |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

which Parent Address field is optional and is set if 'p'flag is set.
Also,it says that:

Thus, to minimize traffic load, the Parent Address field in the CNC object should not be present in the storing mode.

That means the 'p'flag is set for every DIO message sent in Non-storing mode?

Thank you,
Chenyang ji

https://www.ietf.org/id/draft-qasem-roll-rpl-load-balancing-02.txt

----- Original Message -----
From: "Houjianqiang (Derek)" <houjianqiang@huawei.com>
To: "chenyang ji" <chenyang.ji@imt-atlantique.net>
Cc: "roll" <roll@ietf.org>, "jichenyang920" <jichenyang920@gmail.com>, "Qasem, Mamoun" <M.Qasem@napier.ac.uk>
Sent: Tuesday, October 31, 2017 2:42:19 AM
Subject: Re: [Roll] Child count in parent selection

Hi Chenyang



Many thanks for your interest in our draft. I am happy to answer your questions. Please see my response in line.



Best regards,

Jianqiang Hou (Derek)



Date: Mon, 30 Oct 2017 08:40:02 +0100 (CET)

From: Chenyang JI <chenyang.ji@imt-atlantique.net<mailto:chenyang.ji@imt-atlantique.net>>

To: roll@ietf.org<mailto:roll@ietf.org>

Cc: jichenyang920 <jichenyang920@gmail.com<mailto:jichenyang920@gmail.com>>

Subject: [Roll] Child count in parent selection

Message-ID:

                <961578433.3782382.1509349202587.JavaMail.zimbra@imt-atlantique.net<mailto:961578433.3782382.1509349202587.JavaMail.zimbra@imt-atlantique.net>>

Content-Type: text/plain; charset=utf-8



Hello,



I have some questions about draft-hou-roll-rpl-parent-selection-00 and draft-qasem-roll-rpl-load-balancing-01.

In the draft,it describes a new type of metric which has this format:



CNC: 8 bits. The Child Node Count is encoded in 8 bits in unsigned integer format, expressed in number count, representing the number of child nodes.



MAX_CNC: 8 bits. The Maximum Child Node Count is encoded in 8 bits in unsigned integer format, expressed in number count, representing the maximum number of child nodes allowed in the neighbor cache.



but it does not clarify if this number of child nodes refers to the number of direct child or the subtree size.

[Jianqiang] My mistake... This number of child nodes refers to the number of direct child.



In draft-qasem-roll-rpl-load-balancing-01:



In LB-OF algorithm, the received DIO from the child node is counted by the preferred parent node. Each DIO contains the IP address of the chosen preferred parent as detailed in section 4.3. Thus, for each received DIO, the node matches its own IP address with the preferred parent IP address which is inserted in the DIO message, then increments the number of children by ONE for this node if there is a matching.



My question is if that mean it will store the children's address and if yes,how will the node stores its children.

[Jianqiang] There are two possible ways of storing its children. One is to store the children's address in a new table with extra cache, the other way is to manage the neighbor cache entry (after all the direct children set is a subset of the neighbor set).



The first draft

https://tools.ietf.org/html/draft-hou-roll-rpl-parent-selection-00



The second draft

https://datatracker.ietf.org/doc/html/draft-qasem-roll-rpl-load-balancing



Best regards,

Chenyang Ji