From l2vpn-bounces@ietf.org  Sat Aug  2 17:28:29 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 576343A6C14;
	Sat,  2 Aug 2008 17:28:29 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 4D9013A6C14
	for <l2vpn@core3.amsl.com>; Sat,  2 Aug 2008 17:28:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id PUmJtYrAk8S6 for <l2vpn@core3.amsl.com>;
	Sat,  2 Aug 2008 17:28:27 -0700 (PDT)
Received: from exprod7og110.obsmtp.com (exprod7og110.obsmtp.com [64.18.2.173])
	by core3.amsl.com (Postfix) with ESMTP id 5D60B3A6C0F
	for <l2vpn@ietf.org>; Sat,  2 Aug 2008 17:28:27 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob110.postini.com
	([64.18.6.12]) with SMTP; Sat, 02 Aug 2008 17:28:03 PDT
Received: from emailcorp3.jnpr.net ([66.129.254.13]) by p-emsmtp02.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Sat, 2 Aug 2008 17:24:20 -0700
Received: from 172.23.0.45 ([172.23.0.45]) by emailcorp3.jnpr.net
	([66.129.254.13]) with Microsoft Exchange Server HTTP-DAV ; 
	Sun,  3 Aug 2008 00:24:19 +0000
User-Agent: Microsoft-Entourage/11.3.6.070618
Date: Sat, 02 Aug 2008 17:24:20 -0700
Subject: Re: l2vpn-vpls-multi-homing-01.txt
From: Bhupesh Kothari <bhupesh@juniper.net>
To: "UTTARO, JAMES, ATTLABS" <uttaro@att.com>,
	<l2vpn@ietf.org>
Message-ID: <C4BA48C4.10E52%bhupesh@juniper.net>
Thread-Topic: l2vpn-vpls-multi-homing-01.txt
Thread-Index: AcjyEvaDOhOLxQlwSf6GgIV0bnIwJwC7E7+o
In-Reply-To: <1477DEAE19DD884CB004730D0FD77FD708CBC6@misout7msgusr7e.ugd.att.com>
Mime-version: 1.0
Content-type: text/plain;
	charset="US-ASCII"
Content-transfer-encoding: 7bit
X-OriginalArrivalTime: 03 Aug 2008 00:24:20.0343 (UTC)
	FILETIME=[45B5A870:01C8F4FF]
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Jim,

UTTARO, JAMES, ATTLABS <uttaro@att.com> wrote:

> The current approach being used to accomplish leaking "multi-homed"
> information between AS's is too tag the primary with some CV and
> then take action as it crosses the AS boundary ( set LP )..

This won't work for VPLS inter-AS option (b) if different providers
set different LP for an advertisement that belongs to a particular
site.  This would only work if each ASBR, in all ASes, is configured
to set the LP of each VPLS advertisement to what is configured on the
originating PE.  This would require a lot of provisioning.

Note that setting of LP on an ASBR is not a local AS matter as each
VPLS PE in all ASes need to see the exact same LP for an
advertisement.  This is a bit different than what you would usually do
with LP on an ASBR, say for internet routes.


> The draft seems to take the approach of mapping the LP to a VE
> preference for a route ( VE-ID) in the local AS ( AS1). The
> advantage of the first solution is that a route may be leaked across
> multiple AS boundaries. As the CV remains ( transitive ) I can
> always key off of it to set LP in the current AS domain. How can I
> accomplish this with VE pref? Is it kept as the route transits
> through multiple AS? I assume that VE pref is transitive and not
> reset upon re-advertisement from remote AS2 to remote AS3.

Yes, VE preference is transitive as it is carried in L2 Info community
and VE preference is not changed when advertisements are carried
across AS boundaries.  There is no reason to change the VE preference
on an ASBR and path selection can break if different PEs see different
VE pref. values.

The ASBR is translating LP to VE and VE to LP for backwards
compatibility only.

>
> It states for "BGP Path Selection" that ADs with the same exact <RD,
> VE-ID, VBO> are candidates for BGP Path Selection.. In our network
> we use unique RDs to reduce convergence time for primary/backup. Our
> approach is to ensure that both the primary and backup sites are
> opaque to the BGP topology. In this way both routes are sent to all
> PEs participating in the L2VPN. So how would a receiving PE treat to
> routes that have the same VE-ID but different RDs? The draft states
> the they would not be candidates for "BGP Path Selection". This
> implies I cannot manipulate LP to select a primary when I use unique
> RDs. This definitely is in conflict with our fast convergence goals
> for our L2VPN customers.

The draft states that such routes are not candidates for path
selection only for BGP path selection, and not for VPLS path
selection.  Such routes will be seen as different routes by RR as RD
is different, and that is what you want.  However, on the VPLS PE the
routes are candidates for VPLS path selection as VE ID is same, and
thus, VE/LP will be used to select the best route.

>
> Jim Uttaro
>

Bhupesh




From l2vpn-bounces@ietf.org  Sun Aug  3 14:24:43 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id EC31728C1DE;
	Sun,  3 Aug 2008 14:24:42 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id BD72728C1DE
	for <l2vpn@core3.amsl.com>; Sun,  3 Aug 2008 14:24:41 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.117
X-Spam-Level: 
X-Spam-Status: No, score=-1.117 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, DNS_FROM_RFC_BOGUSMX=1.482]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id tZmOFbu9GZeB for <l2vpn@core3.amsl.com>;
	Sun,  3 Aug 2008 14:24:41 -0700 (PDT)
Received: from dog.tcb.net (dog.tcb.net [64.78.150.133])
	by core3.amsl.com (Postfix) with ESMTP id 08E2B28C1D6
	for <l2vpn@ietf.org>; Sun,  3 Aug 2008 14:24:41 -0700 (PDT)
Received: by dog.tcb.net (Postfix, from userid 0)
	id 69B082684EA; Sun,  3 Aug 2008 15:25:07 -0600 (MDT)
Received: from pmac.castlepoint.net (71-215-75-236.hlrn.qwest.net
	[71.215.75.236]) (authenticated-user smtp)
	(TLSv1/SSLv3 AES128-SHA 128/128) by dog.tcb.net with SMTP;
	for l2vpn@ietf.org; Sun, 03 Aug 2008 15:25:07 -0600 (MDT)
	(envelope-from shane@castlepoint.net)
X-Avenger: version=0.7.8; receiver=dog.tcb.net; client-ip=71.215.75.236;
	client-port=1818; syn-fingerprint=65535:55:1:64:M1460,N,W3,N,N,T,S MacOS
	10.4.8; data-bytes=0
Message-Id: <EEF56227-BA2D-4016-885C-57D3F216FC06@castlepoint.net>
From: Shane Amante <shane@castlepoint.net>
To: l2vpn@ietf.org
In-Reply-To: <010a01c8ecd5$071e13e0$13466297@vim.tlt.alcatel.it>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v928.1)
Subject: Re: Re-charter Take 2: Virtual Private Multicast Service (VPMS)
Date: Sun, 3 Aug 2008 15:24:49 -0600
References: <1AB02CAF-DA50-4F0F-BD3A-EF1400616939@castlepoint.net>
	<010a01c8ecd5$071e13e0$13466297@vim.tlt.alcatel.it>
X-Mailer: Apple Mail (2.928.1)
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Folks,

This e-mail concludes the re-charter to add VPMS.  We have deemed that  
there is substantial interest in adding VPMS to the charter.  We will  
now forward on the new charter to the IESG.

-shane & Vach


On Jul 23, 2008, at 9:01 AM, Italo Busi wrote:

> I agree including VPMS within the L2VPN charter
>
> Italo
>
>> -----Original Message-----
>> From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org]
>> On Behalf Of Shane Amante
>> Sent: Wednesday, July 09, 2008 5:17 AM
>> To: l2vpn@ietf.org
>> Subject: Re-charter Take 2: Virtual Private Multicast Service (VPMS)
>>
>> Hi all,
>>
>> It seems that the PWE3 WG has completed their re-charter activity,
>> which now includes defining P2MP PW's:
>> http://www.ietf.org/html.charters/pwe3-charter.html
>>
>> As a result, we would like to conduct a short call for interest to
>> include VPMS (likely utilizing P2MP PW's) into the new L2VPN WG
>> charter, before submitting it to the IESG.  Please indicate to the
>> mailing list whether you agree that it should be included in the new
>> L2VPN charter.  (Silence will not count as support).  Also, please
>> speak up if you have comments or clarifications with respect to the
>> proposed wording of VPMS.
>>
>> Copied below is a full copy of the proposed charter, including VPMS.
>>
>> We will close discussion and gauge consensus on this issue in
>> 2 weeks,
>> on Wednesday, 7/23, at 17:00 MDT, UTC -6.
>>
>> Finally, this call *is not* re-opening the entire charter for
>> discussion, rather this is narrowly focused strictly on whether, or
>> not, to add VPMS to the previously agreed upon L2VPN charter.
>>
>> -shane & Vach
>>
>>
>



From l2vpn-bounces@ietf.org  Sun Aug  3 14:27:47 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id C4F2228C234;
	Sun,  3 Aug 2008 14:27:47 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id BB41328C236
	for <l2vpn@core3.amsl.com>; Sun,  3 Aug 2008 14:27:46 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.117
X-Spam-Level: 
X-Spam-Status: No, score=-1.117 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, DNS_FROM_RFC_BOGUSMX=1.482]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id AlKG1koI8EYb for <l2vpn@core3.amsl.com>;
	Sun,  3 Aug 2008 14:27:46 -0700 (PDT)
Received: from dog.tcb.net (dog.tcb.net [64.78.150.133])
	by core3.amsl.com (Postfix) with ESMTP id 2025E28C234
	for <l2vpn@ietf.org>; Sun,  3 Aug 2008 14:27:46 -0700 (PDT)
Received: by dog.tcb.net (Postfix, from userid 0)
	id BA4A82684EA; Sun,  3 Aug 2008 15:28:12 -0600 (MDT)
Received: from pmac.castlepoint.net (71-215-75-236.hlrn.qwest.net
	[71.215.75.236]) (authenticated-user smtp)
	(TLSv1/SSLv3 AES128-SHA 128/128) by dog.tcb.net with SMTP;
	for l2vpn@ietf.org; Sun, 03 Aug 2008 15:28:12 -0600 (MDT)
	(envelope-from shane@castlepoint.net)
X-Avenger: version=0.7.8; receiver=dog.tcb.net; client-ip=71.215.75.236;
	client-port=1967; syn-fingerprint=65535:55:1:64:M1460,N,W3,N,N,T,S MacOS
	10.4.8; data-bytes=0
Message-Id: <EF31ACE3-485A-45AD-836C-8F626DB26408@castlepoint.net>
From: Shane Amante <shane@castlepoint.net>
To: l2vpn@ietf.org
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v928.1)
Subject: IETF 72: DRAFT L2VPN WG Meeting Minutes
Date: Sun, 3 Aug 2008 15:27:57 -0600
X-Mailer: Apple Mail (2.928.1)
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

The following are the DRAFT meeting minutes of the IETF 72 L2VPN WG  
meeting in Dublin, Ireland:
http://www.ietf.org/proceedings/08jul/minutes/l2vpn.txt

Please take a look and mail the chairs if there are changes.  Changes  
are due by August 22, 2008.

-shane & Vach


From l2vpn-bounces@ietf.org  Sun Aug  3 14:32:26 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id AEF5528C23A;
	Sun,  3 Aug 2008 14:32:26 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id AE1A228C231
	for <l2vpn@core3.amsl.com>; Sun,  3 Aug 2008 14:32:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.117
X-Spam-Level: 
X-Spam-Status: No, score=-1.117 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, DNS_FROM_RFC_BOGUSMX=1.482]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id gil9OQFLSrYN for <l2vpn@core3.amsl.com>;
	Sun,  3 Aug 2008 14:32:24 -0700 (PDT)
Received: from dog.tcb.net (dog.tcb.net [64.78.150.133])
	by core3.amsl.com (Postfix) with ESMTP id 163AF28C229
	for <l2vpn@ietf.org>; Sun,  3 Aug 2008 14:32:24 -0700 (PDT)
Received: by dog.tcb.net (Postfix, from userid 0)
	id B29102684EA; Sun,  3 Aug 2008 15:32:50 -0600 (MDT)
Received: from pmac.castlepoint.net (71-215-75-236.hlrn.qwest.net
	[71.215.75.236]) (authenticated-user smtp)
	(TLSv1/SSLv3 AES128-SHA 128/128) by dog.tcb.net with SMTP;
	for l2vpn@ietf.org; Sun, 03 Aug 2008 15:32:50 -0600 (MDT)
	(envelope-from shane@castlepoint.net)
X-Avenger: version=0.7.8; receiver=dog.tcb.net; client-ip=71.215.75.236;
	client-port=2332; syn-fingerprint=65535:55:1:64:M1460,N,W3,N,N,T,S MacOS
	10.4.8; data-bytes=0
Message-Id: <34C46351-53E9-43BE-A196-CB98FE263541@castlepoint.net>
From: Shane Amante <shane@castlepoint.net>
To: l2vpn@ietf.org
In-Reply-To: <EF31ACE3-485A-45AD-836C-8F626DB26408@castlepoint.net>
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7bit
Mime-Version: 1.0 (Apple Message framework v928.1)
Subject: Re: IETF 72: DRAFT L2VPN WG Meeting Minutes
Date: Sun, 3 Aug 2008 15:32:34 -0600
References: <EF31ACE3-485A-45AD-836C-8F626DB26408@castlepoint.net>
X-Mailer: Apple Mail (2.928.1)
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Sorry for replying to self.

I forgot to mention that we wish to express much gratitude to both  
Dave Wilson & Lieven Levrau for taking the meeting minutes.  Thanks  
guys!

-shane & Vach


On Aug 3, 2008, at 3:27 PM, Shane Amante wrote:

> The following are the DRAFT meeting minutes of the IETF 72 L2VPN WG  
> meeting in Dublin, Ireland:
> http://www.ietf.org/proceedings/08jul/minutes/l2vpn.txt
>
> Please take a look and mail the chairs if there are changes.   
> Changes are due by August 22, 2008.
>
> -shane & Vach



From l2vpn-bounces@ietf.org  Sun Aug  3 22:22:37 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id B5AA93A6C36;
	Sun,  3 Aug 2008 22:22:37 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 1379C3A68F0
	for <l2vpn@core3.amsl.com>; Sun,  3 Aug 2008 22:22:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id Q0gh4lj22MF0 for <l2vpn@core3.amsl.com>;
	Sun,  3 Aug 2008 22:22:36 -0700 (PDT)
Received: from mu-out-0910.google.com (mu-out-0910.google.com [209.85.134.184])
	by core3.amsl.com (Postfix) with ESMTP id 163763A6C49
	for <l2vpn@ietf.org>; Sun,  3 Aug 2008 22:22:35 -0700 (PDT)
Received: by mu-out-0910.google.com with SMTP id g7so1780090muf.9
	for <l2vpn@ietf.org>; Sun, 03 Aug 2008 22:23:02 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma;
	h=domainkey-signature:received:received:message-id:date:from:to
	:subject:in-reply-to:mime-version:content-type
	:content-transfer-encoding:content-disposition:references;
	bh=f/4uuo2WBdIgYwNgYO/P9eV3CghMlD/dwZBBl5uk0xY=;
	b=bbaNn1UYLmhssxZsbYtVUneSmXBZXkj+aVzDLk4/Vqxe75eo8A6GaWO/BMUQg5cI7z
	kMlIQrRynHKi7Sf364g3YQvpnZXhGgwa6m96MF9ZMHwCLBmLpuY2hF47/siyVLSj1kJz
	XSW247Fn31Y5HQ+zjmUm4vcNnyJwsZgWCWLms=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=message-id:date:from:to:subject:in-reply-to:mime-version
	:content-type:content-transfer-encoding:content-disposition
	:references;
	b=JSX54d27IS3G3iuIgGhoyVDZuabl9Tsha2Lu9kmbWbez60u/eZkVFpyOkEWMxFwlb4
	wueOPGx4gkWH5mZWqJtpykH9gXE8dLGstaHBJlcFU9cNuuCvul9PiNSI8fmktTrCPOD+
	T9t4UE8Is13Ls2sh1psfekb1NQODqBDzi00yk=
Received: by 10.103.213.10 with SMTP id p10mr5306921muq.46.1217827381931;
	Sun, 03 Aug 2008 22:23:01 -0700 (PDT)
Received: by 10.103.246.7 with HTTP; Sun, 3 Aug 2008 22:23:01 -0700 (PDT)
Message-ID: <715756490808032223g3d85cb36s461e82d9c6e5f9a1@mail.gmail.com>
Date: Mon, 4 Aug 2008 10:53:01 +0530
From: "venkatesan mahalingam" <venkatflex@gmail.com>
To: l2vpn@ietf.org
Subject: VPLS terms clarification
In-Reply-To: <715756490808031302y6e67ba6cqa0a9b4c0e3bba20a@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
References: <715756490808031302y6e67ba6cqa0a9b4c0e3bba20a@mail.gmail.com>
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

L2 VPN id, VPLS instance and VSI how these three terms associated/used in
VPLS?

Can we have more than one VPLS instances in a VSI at AC side?

Where L2 VPN id will be useful?

How this VSI (L2 context id) mapped in VPLS mib?


Thanks in advance and
Best Regards,
Venkatesan Mahalingam


From l2vpn-bounces@ietf.org  Mon Aug  4 11:51:20 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 0BE813A68E8;
	Mon,  4 Aug 2008 11:51:20 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 0C8B83A6850
	for <l2vpn@core3.amsl.com>; Mon,  4 Aug 2008 11:51:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id Tqhc0QlsOfrm for <l2vpn@core3.amsl.com>;
	Mon,  4 Aug 2008 11:51:17 -0700 (PDT)
Received: from exprod7og102.obsmtp.com (exprod7og102.obsmtp.com [64.18.2.157])
	by core3.amsl.com (Postfix) with ESMTP id 1272F3A680F
	for <l2vpn@ietf.org>; Mon,  4 Aug 2008 11:51:13 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob102.postini.com
	([64.18.6.12]) with SMTP; Mon, 04 Aug 2008 11:51:26 PDT
Received: from p-emlb01-sac.jnpr.net ([66.129.254.46]) by p-emsmtp01.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 11:51:28 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb01-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 11:51:27 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 11:51:21 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m74IpLu28112;
	Mon, 4 Aug 2008 11:51:21 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m74IpLiE021223;
	Mon, 4 Aug 2008 11:51:21 -0700
To: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Date: Mon, 04 Aug 2008 11:51:21 -0700
Message-ID: <21222.1217875881@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 04 Aug 2008 18:51:21.0984 (UTC)
	FILETIME=[16818000:01C8F663]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:

> During the session I expressed concerns regarding 2 dual homes CE(s)
> on 2 PE(s) with multi-homing. Here is the scenario I had in mind.
>=20
>                             (VE1, VE2)
>                 ____ PE1=E2=80=A6=E2=80=A6=E2=80=A6=E2=80=A6=E2=80=A6=E2=
=80=A6..
>               /  ____/  :                               :
>         __/   /         :                                PE3
> CE1 __   /          :                                 :
>             \ /             :                              :
>              / \_____ PE2 =E2=80=A6=E2=80=A6=E2=80=A6=E2=80=A6=E2=80=A6=
=E2=80=A6
>        __/               /  (VE1, VE2)
> CE2 _________/
>=20

> PE3 picks a designated forwarder lets say PE1 for both CE1 and
> CE2. After discussion what I understood is that when CE1 fails PE3
> reelects the designated forwarder for both CE1 and CE2 and PE3 also
> flushes the Mac table.

When CE1--PE1 link fails, all PEs will reelect who is the designated
PE for VE1 (assuming VE1 is for CE1).  Since the designated PE in this
case will move from PE1 to PE2, PE3 needs to flush the MACs learned
from PE1.

>=20
> This should be addressed as an impact on CE1 should not impact other
> CE(s) according to me.
>=20

In VPLS, there is only one PW between a pair of PEs as the ACs are not
tied to particular remote neighbors as in the case of P2P L2VPNs.
Since PWs are not tied to ACs, when a AC goes down there is no way to
flush just the MACs associated with a particular AC other than
explicitly sending a list of MACs to be flushed.  In your example, PE1
can send a list of MACs (learned from CE1) to be flushed to PE3.  The
capability to send a list of MACs to be flushed is not yet there, but
can be added.


> Cheers,
> Wim


Thanks
Bhupesh



From l2vpn-bounces@ietf.org  Mon Aug  4 12:17:18 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 97D7F3A67EC;
	Mon,  4 Aug 2008 12:17:18 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 7E6C73A67EC
	for <l2vpn@core3.amsl.com>; Mon,  4 Aug 2008 12:17:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.04
X-Spam-Level: 
X-Spam-Status: No, score=-2.04 tagged_above=-999 required=5 tests=[AWL=0.209, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id Lx7Q9Yg+iwU5 for <l2vpn@core3.amsl.com>;
	Mon,  4 Aug 2008 12:17:16 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [64.208.49.27])
	by core3.amsl.com (Postfix) with ESMTP id 789E23A6783
	for <l2vpn@ietf.org>; Mon,  4 Aug 2008 12:17:16 -0700 (PDT)
Received: from FRVELSBHS03.ad2.ad.alcatel.com (frvelsbhs03.ad2.ad.alcatel.com
	[155.132.6.75])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m74JHc7V024683; 
	Mon, 4 Aug 2008 21:17:41 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS03.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Mon, 4 Aug 2008 21:17:39 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: draft-kompella-l2vpn-vpls-multihoming-01
Date: Mon, 4 Aug 2008 21:17:41 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <21222.1217875881@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2YyX9O49JHZHyTC+i4wQpI8TaEwAAwj8g
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
	<21222.1217875881@bhupesh-f8.jnpr.net>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 04 Aug 2008 19:17:39.0483 (UTC)
	FILETIME=[C2C4BEB0:01C8F666]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Bhupesh,

Thanks, In this case does the solution allows 2 different designated
forwarder for 2 different VE(s) on the same PE as per the example below.
I would assume it does. PE3 would have to maintain a PW to each PE (PE1
and PE2 in the example).

As you mentioned below we should extend the solution with a MAC flush
capability like LDP does today.

Cheers,
Wim=20

-----Original Message-----
From: Bhupesh Kothari [mailto:bhupesh@juniper.net]=20
Sent: maandag 4 augustus 2008 20:51
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:

> During the session I expressed concerns regarding 2 dual homes CE(s)=20
> on 2 PE(s) with multi-homing. Here is the scenario I had in mind.
>=20
>                             (VE1, VE2)
>                 ____ PE1....................
>               /  ____/  :                               :
>         __/   /         :                                PE3
> CE1 __   /          :                                 :
>             \ /             :                              :
>              / \_____ PE2 ..................
>        __/               /  (VE1, VE2)
> CE2 _________/
>=20

> PE3 picks a designated forwarder lets say PE1 for both CE1 and CE2.=20
> After discussion what I understood is that when CE1 fails PE3 reelects

> the designated forwarder for both CE1 and CE2 and PE3 also flushes the

> Mac table.

When CE1--PE1 link fails, all PEs will reelect who is the designated PE
for VE1 (assuming VE1 is for CE1).  Since the designated PE in this case
will move from PE1 to PE2, PE3 needs to flush the MACs learned from PE1.

>=20
> This should be addressed as an impact on CE1 should not impact other
> CE(s) according to me.
>=20

In VPLS, there is only one PW between a pair of PEs as the ACs are not
tied to particular remote neighbors as in the case of P2P L2VPNs.
Since PWs are not tied to ACs, when a AC goes down there is no way to
flush just the MACs associated with a particular AC other than
explicitly sending a list of MACs to be flushed.  In your example, PE1
can send a list of MACs (learned from CE1) to be flushed to PE3.  The
capability to send a list of MACs to be flushed is not yet there, but
can be added.


> Cheers,
> Wim


Thanks
Bhupesh



From l2vpn-bounces@ietf.org  Mon Aug  4 12:21:12 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 7C1AB3A6A97;
	Mon,  4 Aug 2008 12:21:12 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id A8A783A6A92
	for <l2vpn@core3.amsl.com>; Mon,  4 Aug 2008 12:21:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id SgDCWPggSZOP for <l2vpn@core3.amsl.com>;
	Mon,  4 Aug 2008 12:21:10 -0700 (PDT)
Received: from exprod7og114.obsmtp.com (exprod7ob114.obsmtp.com [64.18.2.214])
	by core3.amsl.com (Postfix) with ESMTP id E42A23A68A3
	for <l2vpn@ietf.org>; Mon,  4 Aug 2008 12:21:05 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob114.postini.com
	([64.18.6.12]) with SMTP; Mon, 04 Aug 2008 12:20:51 PDT
Received: from p-emlb02-sac.jnpr.net ([66.129.254.47]) by p-emsmtp03.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 12:20:51 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb02-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 12:20:50 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 12:20:50 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m74JKnu38396;
	Mon, 4 Aug 2008 12:20:49 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m74JKnA6022991;
	Mon, 4 Aug 2008 12:20:49 -0700
To: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
	<4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Date: Mon, 04 Aug 2008 12:20:49 -0700
Message-ID: <22989.1217877649@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 04 Aug 2008 19:20:50.0001 (UTC)
	FILETIME=[34537C10:01C8F667]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

BALUS Florin <Florin.Balus@alcatel-lucent.com> wrote:

> I have some other comments, questions I want to raise on this draft:
>=20
> -          Let=E2=80=99s assume PE1 is chosen by PE3 over PE2 in the diag=
ram below
>=20
> -          The distribution model for BGP VPLS information (through BGP) =
is one to many with no individual p2p
> communication between PE2-PE3.
>=20
> -          How does PE2 know that he was not chosen by PE3? PE3 needs to =
know that to bring down its own data plane
> towards PE2 or otherwise I see some problems (e.g. PE3 sending BUM traffi=
c to PE2 wasting bandwidth in the core, even
> confusion at PE3 for the PHP case when receiving traffic from both PE2 an=
d PE1 from the same VE ID).


It's explained in Section 3 of the draft.  In brief, path selection on
all VPLS PEs (and RRs if any) will deterministically elect a single
designated PE for a multi-homed customer site.  In your case, path
selection on PE1, PE2 and PE2 will result in PE1 as the winner
(assuming PE1 is better than PE2 -- what is considered better is
specified in the draft).

>=20
> Adding to the comments raised below by Wim
>=20
> -          I think the case of a single homed CE mixed with the dual-home=
d ones seems even more problematic: i.e. the
> single-homed CE could be left hanging, isolated from the other VPLS sites.
>=20

A single homed customer site is expected to be assigned a different VE
ID than a multi-homed site on the same PE.


> -          I was under the impression this could be done by assigning sep=
arate VE IDs & label blocks for each
> dual-homed CEs/single-homed CE group (a complex solution, wasting a lot o=
f label block resources).  I understood from
> hallway discussions that the authors said the separate VE IDs are not any=
how the solution. Can we get a confirmation
> about how this is addressed?
>=20

It must be a misunderstanding.  As I explained above, assigning
different VE IDs is the right approach.



> Thanks,
>=20
> Florin


Thanks
Bhupesh


From l2vpn-bounces@ietf.org  Mon Aug  4 13:37:30 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 7F6FE3A6C15;
	Mon,  4 Aug 2008 13:37:30 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 24B2D3A6C15
	for <l2vpn@core3.amsl.com>; Mon,  4 Aug 2008 13:37:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id FyOgZqN76ke6 for <l2vpn@core3.amsl.com>;
	Mon,  4 Aug 2008 13:37:28 -0700 (PDT)
Received: from exprod7og116.obsmtp.com (exprod7ob116.obsmtp.com [64.18.2.218])
	by core3.amsl.com (Postfix) with ESMTP id 7F3423A6A34
	for <l2vpn@ietf.org>; Mon,  4 Aug 2008 13:37:26 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob116.postini.com
	([64.18.6.12]) with SMTP; Mon, 04 Aug 2008 13:37:54 PDT
Received: from p-emlb01-sac.jnpr.net ([66.129.254.46]) by p-emsmtp01.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 13:37:24 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb01-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 13:37:24 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 13:37:24 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m74KbOu62283;
	Mon, 4 Aug 2008 13:37:24 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m74KbKf0027925;
	Mon, 4 Aug 2008 13:37:24 -0700
To: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
	<21222.1217875881@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
Date: Mon, 04 Aug 2008 13:37:20 -0700
Message-ID: <27924.1217882240@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 04 Aug 2008 20:37:24.0290 (UTC)
	FILETIME=[E6BC5A20:01C8F671]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Wim,

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:

> Bhupesh,
> 
> Thanks, In this case does the solution allows 2 different designated
> forwarder for 2 different VE(s) on the same PE as per the example below.
> I would assume it does. PE3 would have to maintain a PW to each PE (PE1
> and PE2 in the example).

Not necessary. The existence of PWs is independent of how we will
flush MACs.  Consider that VE1 is VE ID 1, VE2 is VE ID 2 and VE3 for
PE3 is VE ID
3.  

Assumes that PE1 is "better" than PE2 for both VE1 and VE2 (based on
preference).  In this case:

-  PE1 will have one PW to PE3 (between IDs 1 and 3). 
-  PE2 will have no PW.
-  PE3 will have one PW to PE1 (between IDs 3 and 1). 

In this case, for both customer sites, CE1 and CE2, PE1 is the
designated forwarder. 


If PE2 is better for VE2 and PE1 is better for VE1, then:

- PE1 (ID 1) will have a PW to PE2 (ID 2) and PE3 (ID 3)
- PE2 (ID 2) will have a PW to PE1 (ID 1) and PE3 (ID 3)
- PE3 (ID 3) will have a PW to PE1 (ID 1) and PE2 (ID 2)

In this case, CE1 traffic is through PE1 and CE2 traffic is through
PE2.

In either case, the goal is to have an implicit or an explicit flush
that will flush only the MACs that need to be.  Note that BGP VPLS has
implicit flush semantics as it is aware of multi-homed sites.


> As you mentioned below we should extend the solution with a MAC flush
> capability like LDP does today.

Yes, we'll add the explicit flush.  The explicit flush capability
provides optimization at the expense of more state in the control
plane as now the PEs need to track the MACs learned in forwarding so
that those can be withdrawn later on.


> 
> Cheers,
> Wim 

Thanks
Bhupesh


From l2vpn-bounces@ietf.org  Mon Aug  4 15:41:42 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 5ADFA3A68AB;
	Mon,  4 Aug 2008 15:41:42 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id AF6993A6858
	for <l2vpn@core3.amsl.com>; Mon,  4 Aug 2008 15:41:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.576
X-Spam-Level: 
X-Spam-Status: No, score=-2.576 tagged_above=-999 required=5 tests=[AWL=0.023, 
	BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id B7yiXyo2LMWt for <l2vpn@core3.amsl.com>;
	Mon,  4 Aug 2008 15:41:39 -0700 (PDT)
Received: from audl952.usa.alcatel.com (audl952.usa.alcatel.com
	[143.209.238.159])
	by core3.amsl.com (Postfix) with ESMTP id C6B5C3A682A
	for <l2vpn@ietf.org>; Mon,  4 Aug 2008 15:41:39 -0700 (PDT)
Received: from usdalsbhs01.ad3.ad.alcatel.com (usdalsbhs01.usa.alcatel.com
	[172.22.216.19])
	by audl952.usa.alcatel.com (ALCANET) with ESMTP id m74Mg79F008614;
	Mon, 4 Aug 2008 17:42:07 -0500
Received: from USDALSMBS03.ad3.ad.alcatel.com ([172.22.216.8]) by
	usdalsbhs01.ad3.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Mon, 4 Aug 2008 17:42:07 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="utf-8"
Content-Transfer-Encoding: base64
Subject: RE: draft-kompella-l2vpn-vpls-multihoming-01
Date: Mon, 4 Aug 2008 17:42:05 -0500
Message-ID: <4A5028372622294A99B8FFF6BD06EB7B048244DF@USDALSMBS03.ad3.ad.alcatel.com>
In-reply-to: <22989.1217877649@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2Z3oLyMQUAZ1xQ+S/FhBMNfAtIAAGGqwg
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
	<22989.1217877649@bhupesh-f8.jnpr.net>
From: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 04 Aug 2008 22:42:07.0291 (UTC)
	FILETIME=[52F3B0B0:01C8F683]
X-Scanned-By: MIMEDefang 2.64 on 143.209.238.34
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org
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From l2vpn-bounces@ietf.org  Mon Aug  4 17:30:02 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id C81923A6926;
	Mon,  4 Aug 2008 17:30:02 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 3A6F03A6926
	for <l2vpn@core3.amsl.com>; Mon,  4 Aug 2008 17:30:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.201
X-Spam-Level: 
X-Spam-Status: No, score=-6.201 tagged_above=-999 required=5 tests=[AWL=0.398, 
	BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id Khju0V3nGvXg for <l2vpn@core3.amsl.com>;
	Mon,  4 Aug 2008 17:30:01 -0700 (PDT)
Received: from exprod7og110.obsmtp.com (exprod7og110.obsmtp.com [64.18.2.173])
	by core3.amsl.com (Postfix) with ESMTP id 2DD343A67CC
	for <l2vpn@ietf.org>; Mon,  4 Aug 2008 17:29:58 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob110.postini.com
	([64.18.6.12]) with SMTP; Mon, 04 Aug 2008 17:30:10 PDT
Received: from p-emlb01-sac.jnpr.net ([66.129.254.46]) by p-emsmtp03.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 17:29:31 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb01-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 17:29:31 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Mon, 4 Aug 2008 17:29:30 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m750TUu34856;
	Mon, 4 Aug 2008 17:29:30 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m750TU42019847;
	Mon, 4 Aug 2008 17:29:30 -0700
To: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <4A5028372622294A99B8FFF6BD06EB7B048244DF@USDALSMBS03.ad3.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
	<22989.1217877649@bhupesh-f8.jnpr.net>
	<4A5028372622294A99B8FFF6BD06EB7B048244DF@USDALSMBS03.ad3.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
Date: Mon, 04 Aug 2008 17:29:30 -0700
Message-ID: <19846.1217896170@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 05 Aug 2008 00:29:30.0896 (UTC)
	FILETIME=[53A3C900:01C8F692]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Florin,

> > >
> > > -          I think the case of a single homed CE mixed with the dual-
> > homed ones seems even more problematic: i.e. the
> > > single-homed CE could be left hanging, isolated from the other VPLS
> > sites.
> > >
> > 
> > A single homed customer site is expected to be assigned a different VE
> > ID than a multi-homed site on the same PE.
> > 
> 
> [[FB]] Then it means that for each "active/preferred" PE and for
> each VPLS instance there will be at least two set of PWs: one for
> the VE ID assigned for the Single-Homed sites, another one for the
> dual-homed CE1, even if they are part of the same customer VPLS
> service.  Otherwise the failure of CE1-PE1 link will bring down all
> the PWs to that VPLS switching instance on PE1, isolating all the
> single-homed CEs.


No, there is only one PW between a pair of PEs and so only one PW is
created even though two customer sites are connected to PE1.  When the
CE1-PE1 link goes down and if the PW goes down (depends on the VE ID),
then another PW will be reestablished, for the single homed site as it
has a unique VE ID in the VPLS instance, and hence, the site will
never be isolated from the VPLS domain.  The remote peers already know
the label to use to reach that single homed site.

 
> Also assuming more than one dual-homed sites are present on PE1 in a
> VPLS instance (& assuming for simplicity PE1 is the preferred PE for
> all of them), then you will also need a different VE ID for each of
> these CEs and a different set of PWs for each of them: i.e. one for
> dual-homed CE2,... one for dual-homed CEn, even if they are part of
> the same customer VPLS?  Otherwise failure of CE1-PE1 link will
> produce interruption of service for the other dual-homed CEs
> (CE2-CEn).
> 
> Did I understand correctly?


A different VE ID will be required, but PW is not created for each.


I'll add more text in the draft to explain operations in such
scenarios. 


> 
> Florin


Bhupesh


From l2vpn-bounces@ietf.org  Tue Aug  5 00:54:37 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id B402A3A69DA;
	Tue,  5 Aug 2008 00:54:37 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 574E93A69DA
	for <l2vpn@core3.amsl.com>; Tue,  5 Aug 2008 00:54:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.063
X-Spam-Level: 
X-Spam-Status: No, score=-2.063 tagged_above=-999 required=5 tests=[AWL=0.186, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id t+BGCio-3BZG for <l2vpn@core3.amsl.com>;
	Tue,  5 Aug 2008 00:54:35 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [64.208.49.27])
	by core3.amsl.com (Postfix) with ESMTP id B6A493A695A
	for <l2vpn@ietf.org>; Tue,  5 Aug 2008 00:54:34 -0700 (PDT)
Received: from FRVELSBHS03.ad2.ad.alcatel.com (frvelsbhs03.ad2.ad.alcatel.com
	[155.132.6.75])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m757t1F1007156; 
	Tue, 5 Aug 2008 09:55:01 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS03.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Tue, 5 Aug 2008 09:55:01 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: draft-kompella-l2vpn-vpls-multihoming-01
Date: Tue, 5 Aug 2008 09:55:01 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <27924.1217882240@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2cfusbpP5APx1QhmSP06sW6dyZwATRPJQ
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
	<21222.1217875881@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com>
	<27924.1217882240@bhupesh-f8.jnpr.net>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 05 Aug 2008 07:55:01.0364 (UTC)
	FILETIME=[903D8340:01C8F6D0]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Bhupesh, thx more in-line

-----Original Message-----
From: Bhupesh Kothari [mailto:bhupesh@juniper.net]=20
Sent: maandag 4 augustus 2008 22:37
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01

Hi Wim,

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:

> Bhupesh,
>=20
> Thanks, In this case does the solution allows 2 different designated=20
> forwarder for 2 different VE(s) on the same PE as per the example
below.
> I would assume it does. PE3 would have to maintain a PW to each PE=20
> (PE1 and PE2 in the example).

Not necessary. The existence of PWs is independent of how we will flush
MACs.  Consider that VE1 is VE ID 1, VE2 is VE ID 2 and VE3 for
PE3 is VE ID
3. =20

Assumes that PE1 is "better" than PE2 for both VE1 and VE2 (based on
preference).  In this case:

-  PE1 will have one PW to PE3 (between IDs 1 and 3).=20
-  PE2 will have no PW.
-  PE3 will have one PW to PE1 (between IDs 3 and 1).=20

In this case, for both customer sites, CE1 and CE2, PE1 is the
designated forwarder.=20


If PE2 is better for VE2 and PE1 is better for VE1, then:

- PE1 (ID 1) will have a PW to PE2 (ID 2) and PE3 (ID 3)
- PE2 (ID 2) will have a PW to PE1 (ID 1) and PE3 (ID 3)
- PE3 (ID 3) will have a PW to PE1 (ID 1) and PE2 (ID 2)

In this case, CE1 traffic is through PE1 and CE2 traffic is through PE2.

In either case, the goal is to have an implicit or an explicit flush
that will flush only the MACs that need to be.  Note that BGP VPLS has
implicit flush semantics as it is aware of multi-homed sites.

WH> Indeed this is what I meant. Thx for clarifying


> As you mentioned below we should extend the solution with a MAC flush=20
> capability like LDP does today.

Yes, we'll add the explicit flush.  The explicit flush capability
provides optimization at the expense of more state in the control plane
as now the PEs need to track the MACs learned in forwarding so that
those can be withdrawn later on.


WH> How would this work? E.g. in the example we use PE3 has no awareness
of the Mac @ belonging to which CE on a given PE.

>=20
> Cheers,
> Wim

Thanks
Bhupesh


From l2vpn-bounces@ietf.org  Tue Aug  5 05:30:03 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id C87C628C291;
	Tue,  5 Aug 2008 05:30:03 -0700 (PDT)
X-Original-To: l2vpn@ietf.org
Delivered-To: l2vpn@core3.amsl.com
Received: by core3.amsl.com (Postfix, from userid 0)
	id 8AF3128C270; Tue,  5 Aug 2008 05:30:01 -0700 (PDT)
From: Internet-Drafts@ietf.org
To: i-d-announce@ietf.org
Subject: I-D Action:draft-ietf-l2vpn-vpls-mcast-reqts-06.txt 
Content-Type: Multipart/Mixed; Boundary="NextPart"
Mime-Version: 1.0
Message-Id: <20080805123001.8AF3128C270@core3.amsl.com>
Date: Tue,  5 Aug 2008 05:30:01 -0700 (PDT)
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org


--NextPart

A New Internet-Draft is available from the on-line Internet-Drafts directories.
This draft is a work item of the Layer 2 Virtual Private Networks Working Group of the IETF.


	Title           : Requirements for Multicast Support in Virtual Private LAN Services
	Author(s)       : Y. Kamite, et al.
	Filename        : draft-ietf-l2vpn-vpls-mcast-reqts-06.txt
	Pages           : 30
	Date            : 2008-08-05

This document provides functional requirements for network solutions
that support multicast over Virtual Private LAN Service (VPLS).  It
specifies requirements both from the end user and service provider
standpoints.  It is intended that potential solutions will use these
requirements as guidelines.

A URL for this Internet-Draft is:
http://www.ietf.org/internet-drafts/draft-ietf-l2vpn-vpls-mcast-reqts-06.txt

Internet-Drafts are also available by anonymous FTP at:
ftp://ftp.ietf.org/internet-drafts/

Below is the data which will enable a MIME compliant mail reader
implementation to automatically retrieve the ASCII version of the
Internet-Draft.

--NextPart
Content-Type: Message/External-body;
	name="draft-ietf-l2vpn-vpls-mcast-reqts-06.txt";
	site="ftp.ietf.org";
	access-type="anon-ftp";
	directory="internet-drafts"

Content-Type: text/plain
Content-ID: <2008-08-05052037.I-D@ietf.org>


--NextPart--


From l2vpn-bounces@ietf.org  Tue Aug  5 15:07:37 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id C368F28C2A1;
	Tue,  5 Aug 2008 15:07:37 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id EDA0A3A68FD
	for <l2vpn@core3.amsl.com>; Tue,  5 Aug 2008 15:07:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.579
X-Spam-Level: 
X-Spam-Status: No, score=-2.579 tagged_above=-999 required=5 tests=[AWL=0.020, 
	BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id wFx-MWszf4oy for <l2vpn@core3.amsl.com>;
	Tue,  5 Aug 2008 15:07:35 -0700 (PDT)
Received: from audl751.usa.alcatel.com (audl751.usa.alcatel.com
	[143.209.238.164])
	by core3.amsl.com (Postfix) with ESMTP id AC75128C328
	for <l2vpn@ietf.org>; Tue,  5 Aug 2008 15:07:33 -0700 (PDT)
Received: from usdalsbhs01.ad3.ad.alcatel.com (usdalsbhs01.usa.alcatel.com
	[172.22.216.19])
	by audl751.usa.alcatel.com (ALCANET) with ESMTP id m75M84J0028071;
	Tue, 5 Aug 2008 17:08:04 -0500
Received: from USDALSMBS03.ad3.ad.alcatel.com ([172.22.216.8]) by
	usdalsbhs01.ad3.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Tue, 5 Aug 2008 17:08:04 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: draft-kompella-l2vpn-vpls-multihoming-01
Date: Tue, 5 Aug 2008 17:08:00 -0500
Message-ID: <4A5028372622294A99B8FFF6BD06EB7B048246F6@USDALSMBS03.ad3.ad.alcatel.com>
In-reply-to: <19846.1217896170@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2knipPLqu8dV0T6CazOFp/XkRLQAAJVYA
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
	<22989.1217877649@bhupesh-f8.jnpr.net>
	<4A5028372622294A99B8FFF6BD06EB7B048244DF@USDALSMBS03.ad3.ad.alcatel.com>
	<19846.1217896170@bhupesh-f8.jnpr.net>
From: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 05 Aug 2008 22:08:04.0400 (UTC)
	FILETIME=[BBB50300:01C8F747]
X-Scanned-By: MIMEDefang 2.51 on 143.209.238.34
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org


Hi Bhupesh,
Thanks for the answers.=20

>From what you explain below the failure of the CE1-PE1 access link
affects the PW infrastructure and implicitly the communication between
all the other CEs connected to the same VPLS on PE1. I do not think this
is a good approach as it extends the scope of the failure.

More in-line for technicalities...

> -----Original Message-----
> From: Bhupesh Kothari [mailto:bhupesh@juniper.net]
> Sent: Monday, August 04, 2008 5:30 PM
> To: BALUS Florin
> Cc: l2vpn@ietf.org
> Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
>=20
> Hi Florin,
>=20
> > > >
> > > > -          I think the case of a single homed CE mixed with the
> dual-
> > > homed ones seems even more problematic: i.e. the
> > > > single-homed CE could be left hanging, isolated from the other
> VPLS
> > > sites.
> > > >
> > >
> > > A single homed customer site is expected to be assigned a
different
> VE
> > > ID than a multi-homed site on the same PE.
> > >
> >
> > [[FB]] Then it means that for each "active/preferred" PE and for
> > each VPLS instance there will be at least two set of PWs: one for
> > the VE ID assigned for the Single-Homed sites, another one for the
> > dual-homed CE1, even if they are part of the same customer VPLS
> > service.  Otherwise the failure of CE1-PE1 link will bring down all
> > the PWs to that VPLS switching instance on PE1, isolating all the
> > single-homed CEs.
>=20
>=20
> No, there is only one PW between a pair of PEs and so only one PW is
> created even though two customer sites are connected to PE1.  When the
> CE1-PE1 link goes down and if the PW goes down (depends on the VE ID),
> then another PW will be reestablished, for the single homed site as it
> has a unique VE ID in the VPLS instance, and hence, the site will
> never be isolated from the VPLS domain. =20

[[FB]] From what you are saying here it looks like the dual-homed CE1
and the single-homed CEs on PE1 share initially the same PW. Then if
CE1-PE1 link fails the initial PW will be torn down and a new PW will be
re-established towards PE2 in order to reach CE1. Also another PW to the
site corresponding to the single-homed CEs on PE1 will be
reestablished/activated.=20

This results in a service interruption for single-homed sites due to
failure of CE1-PE1 access link. I think this is not acceptable as it
increases the scope of the failure unnecessarily.=20

>The remote peers already know
> the label to use to reach that single homed site.
>=20
[[FB]] From what you are saying the service labels/PWs assigned for the
dual-homed site/VE ID are used for all the other sites/VE IDs in the
VPLS on PE1? In other words the remote PE peers do not instantiate in
the data path the service labels for the single-homed site/VE ID from
PE1?=20

What if there are multiple dual-homed sites on the same PEs, in the same
VPLS? Which VE ID's PW is chosen and based on what? (I am talking about
the case of different sites/VE IDs in the same VPLS)


Thanks,
Florin

>=20
> > Also assuming more than one dual-homed sites are present on PE1 in a
> > VPLS instance (& assuming for simplicity PE1 is the preferred PE for
> > all of them), then you will also need a different VE ID for each of
> > these CEs and a different set of PWs for each of them: i.e. one for
> > dual-homed CE2,... one for dual-homed CEn, even if they are part of
> > the same customer VPLS?  Otherwise failure of CE1-PE1 link will
> > produce interruption of service for the other dual-homed CEs
> > (CE2-CEn).
> >
> > Did I understand correctly?
>=20
>=20
> A different VE ID will be required, but PW is not created for each.
>=20
>=20
> I'll add more text in the draft to explain operations in such
> scenarios.
>=20
>=20
> >
> > Florin
>=20
>=20
> Bhupesh


From l2vpn-bounces@ietf.org  Wed Aug  6 02:29:49 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id D581C3A6A59;
	Wed,  6 Aug 2008 02:29:49 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 5B8363A6A59
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 02:29:49 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.836
X-Spam-Level: 
X-Spam-Status: No, score=-0.836 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, HTML_MESSAGE=0.001, MISSING_SUBJECT=1.762]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id mOQCHFV1cW9P for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 02:29:48 -0700 (PDT)
Received: from szxga02-in.huawei.com (szxga02-in.huawei.com [119.145.14.65])
	by core3.amsl.com (Postfix) with ESMTP id E731B3A6890
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 02:29:47 -0700 (PDT)
Received: from huawei.com (szxga02-in [172.24.2.6])
	by szxga02-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5600219ADSVD@szxga02-in.huawei.com> for
	l2vpn@ietf.org; Wed, 06 Aug 2008 17:29:53 +0800 (CST)
Received: from huawei.com ([172.24.1.24])
	by szxga02-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5600G1GADSMD@szxga02-in.huawei.com> for
	l2vpn@ietf.org; Wed, 06 Aug 2008 17:29:52 +0800 (CST)
Received: from j59929 ([10.70.77.93])
	by szxml04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5600030ADRTK@szxml04-in.huawei.com> for
	l2vpn@ietf.org; Wed, 06 Aug 2008 17:29:52 +0800 (CST)
Date: Wed, 06 Aug 2008 17:29:51 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: 
To: sajassi@cisco.com
Message-id: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_okbp+uoXR6Z41gCnv5GBWw)"
X-Priority: 3
X-MSMail-priority: Normal
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_okbp+uoXR6Z41gCnv5GBWw)
Content-type: text/plain; charset=gb2312
Content-transfer-encoding: 7BIT

Hi, Ali:

Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
Following are the two questions:

1. What do you think will be faster convergence? When you flush MAC addresses, 
the learning is done in data plane at wire speed. Why does that prolong convergence time?


-Faster convergence means the time to associate the MACs with the "exact" out port
will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
the wire speed, but also on the host processing time and delays on all the intermediate routers.
For this reason, flooding may also last for a rather long time.


2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
 distributed among multiple PE's, how do you take care of that?

-I believe the most commonly deployed scenario is one backup PE for a primary PE.
In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
only one PE out of them will be elected as the working PE when the primary PE failed, 
rather than distribute the traffic over multiple backup PEs,
otherwise, loops may be introduced to the attached customer site.
BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
add them to the future drafts. 

cheers,
Yuanlong

--Boundary_(ID_okbp+uoXR6Z41gCnv5GBWw)
Content-type: text/html; charset=gb2312
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=gb2312">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">Thanks for your comments in the last L2VPN WG 
meeting, I will try to give you some reasons here.</FONT></DIV>
<DIV><FONT face="Times New Roman">Following are the&nbsp;two 
questions:</FONT></DIV>
<DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">1. What do you think will be faster 
convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
<DIV><FONT face="Times New Roman">the learning is done in data plane at wire 
speed. Why does that prolong convergence time?<BR></FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">-Faster convergence means the time to 
associate the MACs with the "exact" out port</FONT></DIV>
<DIV><FONT face="Times New Roman">will be shorter. Though learning is done in 
data plane at wire speed, the time for a frame to arrive </FONT></DIV>
<DIV><FONT face="Times New Roman">at the end host and the response time for a 
host to send a reverse frame&nbsp;accross the PE (on which the </FONT></DIV>
<DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather long 
time. So the learning process is not only determined solely by </FONT></DIV>
<DIV><FONT face="Times New Roman">the wire </FONT><FONT 
face="Times New Roman">speed, but also on the host processing time and delays on 
all the intermediate routers.</FONT></DIV>
<DIV><FONT face="Times New Roman">For this reason, flooding may also last for a 
rather long time.</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
<DIV><FONT face="Times New Roman">2. You consider a simple case with simple 
backup for primary.&nbsp;In cases where primary fails and traffic 
gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
that?</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">-I believe the most commonly 
deployed&nbsp;scenario is&nbsp;one backup PE for a primary PE.</FONT></DIV>
<DIV><FONT face="Times New Roman">In my opinion, for the scenario of multiple 
backup PEs protecting a primary PE, </FONT></DIV>
<DIV><FONT face="Times New Roman">only one PE out of them will be elected as the 
working PE when the primary PE failed, </FONT></DIV>
<DIV><FONT face="Times New Roman">rather than distribute the traffic over 
multiple backup PEs,</FONT></DIV>
<DIV><FONT face="Times New Roman">otherwise, loops may be introduced to the 
attached customer site.</FONT></DIV>
<DIV><FONT face="Times New Roman">BTW, in a similar way, for PW redundancy 
mechanism, only 1 secondary is selected as the working PW (see </FONT></DIV>
<DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
in&nbsp;PWE3 WG)</FONT></DIV>
<DIV><FONT face="Times New Roman">Anyway, if there is any such requirements 
confirmed, I think it will be OK for us to find some solutions and </FONT></DIV>
<DIV><FONT face="Times New Roman">add them to&nbsp;the future 
drafts.&nbsp;</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
<DIV><FONT face="Times New Roman">Yuanlong<BR></FONT></DIV></BODY></HTML>

--Boundary_(ID_okbp+uoXR6Z41gCnv5GBWw)--


From l2vpn-bounces@ietf.org  Wed Aug  6 02:37:28 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 4916D28C143;
	Wed,  6 Aug 2008 02:37:28 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 9C67128C15A
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 02:37:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.665
X-Spam-Level: 
X-Spam-Status: No, score=-0.665 tagged_above=-999 required=5
	tests=[AWL=-0.171, BAYES_00=-2.599, FH_RELAY_NODNS=1.451,
	HELO_MISMATCH_COM=0.553, HTML_MESSAGE=0.001, RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id 7poRRXM+f3+Y for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 02:37:26 -0700 (PDT)
Received: from szxga04-in.huawei.com (unknown [119.145.14.67])
	by core3.amsl.com (Postfix) with ESMTP id 7393928C138
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 02:37:26 -0700 (PDT)
Received: from huawei.com (szxga04-in [172.24.2.12])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5600511AKQLY@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Wed, 06 Aug 2008 17:34:02 +0800 (CST)
Received: from huawei.com ([172.24.1.18])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5600KF9AKQ4F@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Wed, 06 Aug 2008 17:34:02 +0800 (CST)
Received: from j59929 ([10.70.77.93])
	by szxml03-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5600IBWAKPGN@szxml03-in.huawei.com> for
	l2vpn@ietf.org; Wed, 06 Aug 2008 17:34:01 +0800 (CST)
Date: Wed, 06 Aug 2008 17:34:01 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Reply to your comments on the draft
	"Optimized MAC Address Operations in VPLS with Redundancy"
To: sajassi@cisco.com
Message-id: <002601c8f7a7$8f532890$5d4d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_xxpfOJidU37IrrIwGzKuPw)"
X-Priority: 3
X-MSMail-priority: Normal
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_xxpfOJidU37IrrIwGzKuPw)
Content-type: text/plain; charset=gb2312
Content-transfer-encoding: 7BIT

Hi, Ali:

Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
Following are the two questions:

1. What do you think will be faster convergence? When you flush MAC addresses, 
the learning is done in data plane at wire speed. Why does that prolong convergence time?


-Faster convergence means the time to associate the MACs with the "exact" out port
will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
the wire speed, but also on the host processing time and delays on all the intermediate routers.
For this reason, flooding may also last for a rather long time.


2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
 distributed among multiple PE's, how do you take care of that?

-I believe the most commonly deployed scenario is one backup PE for a primary PE.
In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
only one PE out of them will be elected as the working PE when the primary PE failed, 
rather than distribute the traffic over multiple backup PEs,
otherwise, loops may be introduced to the attached customer site.
BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
add them to the future drafts. 

cheers,
Yuanlong

--Boundary_(ID_xxpfOJidU37IrrIwGzKuPw)
Content-type: text/html; charset=gb2312
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=gb2312">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">Thanks for your comments in the last L2VPN WG 
meeting, I will try to give you some reasons here.</FONT></DIV>
<DIV><FONT face="Times New Roman">Following are the&nbsp;two 
questions:</FONT></DIV>
<DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">1. What do you think will be faster 
convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
<DIV><FONT face="Times New Roman">the learning is done in data plane at wire 
speed. Why does that prolong convergence time?<BR></FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">-Faster convergence means the time to 
associate the MACs with the "exact" out port</FONT></DIV>
<DIV><FONT face="Times New Roman">will be shorter. Though learning is done in 
data plane at wire speed, the time for a frame to arrive </FONT></DIV>
<DIV><FONT face="Times New Roman">at the end host and the response time for a 
host to send a reverse frame&nbsp;accross the PE (on which the </FONT></DIV>
<DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather long 
time. So the learning process is not only determined solely by </FONT></DIV>
<DIV><FONT face="Times New Roman">the wire </FONT><FONT 
face="Times New Roman">speed, but also on the host processing time and delays on 
all the intermediate routers.</FONT></DIV>
<DIV><FONT face="Times New Roman">For this reason, flooding may also last for a 
rather long time.</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
<DIV><FONT face="Times New Roman">2. You consider a simple case with simple 
backup for primary.&nbsp;In cases where primary fails and traffic 
gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
that?</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">-I believe the most commonly 
deployed&nbsp;scenario is&nbsp;one backup PE for a primary PE.</FONT></DIV>
<DIV><FONT face="Times New Roman">In my opinion, for the scenario of multiple 
backup PEs protecting a primary PE, </FONT></DIV>
<DIV><FONT face="Times New Roman">only one PE out of them will be elected as the 
working PE when the primary PE failed, </FONT></DIV>
<DIV><FONT face="Times New Roman">rather than distribute the traffic over 
multiple backup PEs,</FONT></DIV>
<DIV><FONT face="Times New Roman">otherwise, loops may be introduced to the 
attached customer site.</FONT></DIV>
<DIV><FONT face="Times New Roman">BTW, in a similar way, for PW redundancy 
mechanism, only 1 secondary is selected as the working PW (see </FONT></DIV>
<DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
in&nbsp;PWE3 WG)</FONT></DIV>
<DIV><FONT face="Times New Roman">Anyway, if there is any such requirements 
confirmed, I think it will be OK for us to find some solutions and </FONT></DIV>
<DIV><FONT face="Times New Roman">add them to&nbsp;the future 
drafts.&nbsp;</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
<DIV><FONT face="Times New Roman">Yuanlong<BR></FONT></DIV></BODY></HTML>

--Boundary_(ID_xxpfOJidU37IrrIwGzKuPw)--


From l2vpn-bounces@ietf.org  Wed Aug  6 12:08:38 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 9D6D33A6B05;
	Wed,  6 Aug 2008 12:08:38 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 52C603A6972
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 12:08:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.267
X-Spam-Level: 
X-Spam-Status: No, score=-6.267 tagged_above=-999 required=5 tests=[AWL=0.332, 
	BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id vY1iKdCf6iE1 for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 12:08:36 -0700 (PDT)
Received: from exprod7og104.obsmtp.com (exprod7og104.obsmtp.com [64.18.2.161])
	by core3.amsl.com (Postfix) with ESMTP id 0202F3A6B26
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 12:08:31 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob104.postini.com
	([64.18.6.12]) with SMTP; Wed, 06 Aug 2008 12:07:47 PDT
Received: from p-emlb02-sac.jnpr.net ([66.129.254.47]) by p-emsmtp03.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 12:09:04 -0700
Received: from emailsmtp55.jnpr.net ([172.24.18.132]) by p-emlb02-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 12:09:03 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp55.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.1830); 
	Wed, 6 Aug 2008 12:09:03 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m76J93u24282;
	Wed, 6 Aug 2008 12:09:03 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m76J92ud002019;
	Wed, 6 Aug 2008 12:09:03 -0700
To: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
	<21222.1217875881@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com>
	<27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
Date: Wed, 06 Aug 2008 12:09:02 -0700
Message-ID: <2018.1218049742@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 06 Aug 2008 19:09:03.0727 (UTC)
	FILETIME=[E42E1FF0:01C8F7F7]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:
> > As you mentioned below we should extend the solution with a MAC flush 
> > capability like LDP does today.
> 
> Yes, we'll add the explicit flush.  The explicit flush capability
> provides optimization at the expense of more state in the control plane
> as now the PEs need to track the MACs learned in forwarding so that
> those can be withdrawn later on.
> 
> 
> WH> How would this work? E.g. in the example we use PE3 has no awareness
> of the Mac @ belonging to which CE on a given PE.
> 

I should have been more clear.  A MAC flush message from PE1 containing a list
of MACs will be sufficient for this purpose.

Bhupesh


From l2vpn-bounces@ietf.org  Wed Aug  6 13:18:31 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 59C4A28C13D;
	Wed,  6 Aug 2008 13:18:31 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 1188928C13D
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 13:18:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.081
X-Spam-Level: 
X-Spam-Status: No, score=-2.081 tagged_above=-999 required=5 tests=[AWL=0.167, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id eAXP8CDgb+Ak for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 13:18:29 -0700 (PDT)
Received: from smail6.alcatel.fr (colt-na5.alcatel.fr [62.23.212.5])
	by core3.amsl.com (Postfix) with ESMTP id A644E28C0D6
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 13:18:28 -0700 (PDT)
Received: from FRVELSBHS06.ad2.ad.alcatel.com (frvelsbhs06.ad2.ad.alcatel.com
	[155.132.6.78])
	by smail6.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m76KIgOQ022294; 
	Wed, 6 Aug 2008 22:18:42 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS06.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Wed, 6 Aug 2008 22:18:42 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8F801.9E460E44"
Subject: RE: 
Date: Wed, 6 Aug 2008 22:18:41 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Index: Acj3pxaLn+anxK1pSba3lz+DjqzWJQAR5hOQ
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Jiang Yuan-long" <yljiang@huawei.com>, <sajassi@cisco.com>
X-OriginalArrivalTime: 06 Aug 2008 20:18:42.0220 (UTC)
	FILETIME=[9EC18EC0:01C8F801]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.84
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8F801.9E460E44
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Yuanlong,
=20
The difficulty with your draft is that it requires a PE to do different
things depending in which situation the PE is in, and this makes it very
cumbersome in case changes happen to topologies or configurations.
=20
My view is that we need a simple MAC flush which we have today in LDP.
=20
Cheers,
WIm

________________________________

From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf
Of Jiang Yuan-long
Sent: woensdag 6 augustus 2008 11:30
To: sajassi@cisco.com
Cc: l2vpn@ietf.org
Subject:=20


Hi, Ali:
=20
Thanks for your comments in the last L2VPN WG meeting, I will try to
give you some reasons here.
Following are the two questions:
=20
1. What do you think will be faster convergence? When you flush MAC
addresses,=20
the learning is done in data plane at wire speed. Why does that prolong
convergence time?

=20
-Faster convergence means the time to associate the MACs with the
"exact" out port
will be shorter. Though learning is done in data plane at wire speed,
the time for a frame to arrive=20
at the end host and the response time for a host to send a reverse frame
accross the PE (on which the=20
MACs is flushed) will be a rather long time. So the learning process is
not only determined solely by=20
the wire speed, but also on the host processing time and delays on all
the intermediate routers.
For this reason, flooding may also last for a rather long time.
=20
=20
2. You consider a simple case with simple backup for primary. In cases
where primary fails and traffic gets
 distributed among multiple PE's, how do you take care of that?
=20
-I believe the most commonly deployed scenario is one backup PE for a
primary PE.
In my opinion, for the scenario of multiple backup PEs protecting a
primary PE,=20
only one PE out of them will be elected as the working PE when the
primary PE failed,=20
rather than distribute the traffic over multiple backup PEs,
otherwise, loops may be introduced to the attached customer site.
BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is
selected as the working PW (see=20
draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
Anyway, if there is any such requirements confirmed, I think it will be
OK for us to find some solutions and=20
add them to the future drafts.=20
=20
cheers,
Yuanlong


------_=_NextPart_001_01C8F801.9E460E44
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Dus-ascii">
<META content=3D"MSHTML 6.00.2900.3354" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV dir=3Dltr align=3Dleft>Yuanlong<SPAN =
class=3D924080318-06082008>,</SPAN></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN =
class=3D924080318-06082008></SPAN>&nbsp;</DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2>The difficulty with your draft is that it =
requires a PE to=20
do different things depending in which situation the PE is in, and this =
makes it=20
very cumbersome in case changes happen to topologies or=20
configurations.</FONT></SPAN></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2>My view is that we need a simple MAC flush =
which we have=20
today in LDP.</FONT></SPAN></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2>Cheers,</FONT></SPAN></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2>WIm</FONT></SPAN></DIV><BR>
<DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
<HR tabIndex=3D-1>
<FONT face=3DTahoma size=3D2><B>From:</B> l2vpn-bounces@ietf.org=20
[mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang=20
Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B>=20
sajassi@cisco.com<BR><B>Cc:</B> l2vpn@ietf.org<BR><B>Subject:</B>=20
<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV><FONT face=3D"Times New Roman">Hi, Ali:</FONT></DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face=3D"Times New Roman">Thanks for your comments in the last =
L2VPN WG=20
meeting, I will try to give you some reasons here.</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">Following are the&nbsp;two=20
questions:</FONT></DIV>
<DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face=3D"Times New Roman">1. What do you think will be faster=20
convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
<DIV><FONT face=3D"Times New Roman">the learning is done in data plane =
at wire=20
speed. Why does that prolong convergence time?<BR></FONT></DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face=3D"Times New Roman">-Faster convergence means the time =
to=20
associate the MACs with the "exact" out port</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">will be shorter. Though learning is =
done in=20
data plane at wire speed, the time for a frame to arrive </FONT></DIV>
<DIV><FONT face=3D"Times New Roman">at the end host and the response =
time for a=20
host to send a reverse frame&nbsp;accross the PE (on which the =
</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">MACs&nbsp;is flushed) will be a =
rather long=20
time. So the learning process is not only determined solely by =
</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">the wire </FONT><FONT=20
face=3D"Times New Roman">speed, but also on the host processing time and =
delays on=20
all the intermediate routers.</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">For this reason, flooding may also =
last for a=20
rather long time.</FONT></DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV></DIV>
<DIV><FONT face=3D"Times New Roman">2. You consider a simple case with =
simple=20
backup for primary.&nbsp;In cases where primary fails and traffic=20
gets<BR>&nbsp;distributed among multiple PE's, how do you take care of=20
that?</FONT></DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face=3D"Times New Roman">-I believe the most commonly=20
deployed&nbsp;scenario is&nbsp;one backup PE for a primary =
PE.</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">In my opinion, for the scenario of =
multiple=20
backup PEs protecting a primary PE, </FONT></DIV>
<DIV><FONT face=3D"Times New Roman">only one PE out of them will be =
elected as the=20
working PE when the primary PE failed, </FONT></DIV>
<DIV><FONT face=3D"Times New Roman">rather than distribute the traffic =
over=20
multiple backup PEs,</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">otherwise, loops may be introduced =
to the=20
attached customer site.</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">BTW, in a similar way, for PW =
redundancy=20
mechanism, only 1 secondary is selected as the working PW (see =
</FONT></DIV>
<DIV><FONT face=3D"Times New =
Roman">draft-ietf-pwe3-redundancy-bit-00.txt=20
in&nbsp;PWE3 WG)</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">Anyway, if there is any such =
requirements=20
confirmed, I think it will be OK for us to find some solutions and =
</FONT></DIV>
<DIV><FONT face=3D"Times New Roman">add them to&nbsp;the future=20
drafts.&nbsp;</FONT></DIV>
<DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face=3D"Times New Roman">cheers,</FONT></DIV>
<DIV><FONT face=3D"Times New =
Roman">Yuanlong<BR></FONT></DIV></BODY></HTML>

------_=_NextPart_001_01C8F801.9E460E44--


From l2vpn-bounces@ietf.org  Wed Aug  6 14:54:58 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id F01603A68AE;
	Wed,  6 Aug 2008 14:54:57 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id DFB0D3A67DD
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 14:54:56 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.315
X-Spam-Level: 
X-Spam-Status: No, score=-6.315 tagged_above=-999 required=5 tests=[AWL=0.284, 
	BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id N4cWl40QT1aM for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 14:54:55 -0700 (PDT)
Received: from exprod7og107.obsmtp.com (exprod7og107.obsmtp.com [64.18.2.167])
	by core3.amsl.com (Postfix) with ESMTP id D2A603A689B
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 14:54:38 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob107.postini.com
	([64.18.6.12]) with SMTP; Wed, 06 Aug 2008 14:52:17 PDT
Received: from p-emlb02-sac.jnpr.net ([66.129.254.47]) by p-emsmtp02.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 14:54:14 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb02-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 14:54:14 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 14:54:12 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m76LsCu02528;
	Wed, 6 Aug 2008 14:54:12 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m76LsCj1013003;
	Wed, 6 Aug 2008 14:54:12 -0700
To: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <4A5028372622294A99B8FFF6BD06EB7B048246F6@USDALSMBS03.ad3.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
	<22989.1217877649@bhupesh-f8.jnpr.net>
	<4A5028372622294A99B8FFF6BD06EB7B048244DF@USDALSMBS03.ad3.ad.alcatel.com>
	<19846.1217896170@bhupesh-f8.jnpr.net>
	<4A5028372622294A99B8FFF6BD06EB7B048246F6@USDALSMBS03.ad3.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
Date: Wed, 06 Aug 2008 14:54:12 -0700
Message-ID: <13002.1218059652@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 06 Aug 2008 21:54:12.0452 (UTC)
	FILETIME=[F63D9240:01C8F80E]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Florin,

BALUS Florin <Florin.Balus@alcatel-lucent.com> wrote:
> 
> Hi Bhupesh,
> Thanks for the answers. 
> 
> From what you explain below the failure of the CE1-PE1 access link
> affects the PW infrastructure and implicitly the communication between
> all the other CEs connected to the same VPLS on PE1. I do not think this
> is a good approach as it extends the scope of the failure.

The scope of the failure is not extended to other CEs.  See below for
more details.


> 
> More in-line for technicalities...
> 
> > -----Original Message-----
> > From: Bhupesh Kothari [mailto:bhupesh@juniper.net]
> > Sent: Monday, August 04, 2008 5:30 PM
> > To: BALUS Florin
> > Cc: l2vpn@ietf.org
> > Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
> > 
> > Hi Florin,
> > 
> > > > >
> > > > > -          I think the case of a single homed CE mixed with the
> > dual-
> > > > homed ones seems even more problematic: i.e. the
> > > > > single-homed CE could be left hanging, isolated from the other
> > VPLS
> > > > sites.
> > > > >
> > > >
> > > > A single homed customer site is expected to be assigned a
> different
> > VE
> > > > ID than a multi-homed site on the same PE.
> > > >
> > >
> > > [[FB]] Then it means that for each "active/preferred" PE and for
> > > each VPLS instance there will be at least two set of PWs: one for
> > > the VE ID assigned for the Single-Homed sites, another one for the
> > > dual-homed CE1, even if they are part of the same customer VPLS
> > > service.  Otherwise the failure of CE1-PE1 link will bring down all
> > > the PWs to that VPLS switching instance on PE1, isolating all the
> > > single-homed CEs.
> > 
> > 
> > No, there is only one PW between a pair of PEs and so only one PW is
> > created even though two customer sites are connected to PE1.  When the
> > CE1-PE1 link goes down and if the PW goes down (depends on the VE ID),
> > then another PW will be reestablished, for the single homed site as it
> > has a unique VE ID in the VPLS instance, and hence, the site will
> > never be isolated from the VPLS domain.  
> 
> [[FB]] From what you are saying here it looks like the dual-homed CE1
> and the single-homed CEs on PE1 share initially the same PW. Then if
> CE1-PE1 link fails the initial PW will be torn down and a new PW will be
> re-established towards PE2 in order to reach CE1.

Correct.


> Also another PW to the site corresponding to the single-homed CEs on
> PE1 will be reestablished/activated.
> 
> This results in a service interruption for single-homed sites due to
> failure of CE1-PE1 access link. I think this is not acceptable as it
> increases the scope of the failure unnecessarily. 


Your understanding is correct about the PW activation, but this does
not necessarily mean traffic disruption for single-homed site.  See
below for details.


> 
> >The remote peers already know
> > the label to use to reach that single homed site.
> > 
> [[FB]] From what you are saying the service labels/PWs assigned for the
> dual-homed site/VE ID are used for all the other sites/VE IDs in the
> VPLS on PE1?

Yes.

> In other words the remote PE peers do not instantiate in the data
> path the service labels for the single-homed site/VE ID from PE1?
>

Yes.

 
> What if there are multiple dual-homed sites on the same PEs, in the same
> VPLS? Which VE ID's PW is chosen and based on what? (I am talking about
> the case of different sites/VE IDs in the same VPLS)

Irrespective of whether multiple sites are all multi-homed or only
some are multi-homed and others single-homed, ONLY one site is
"selected" for PW establishment that will carry traffic.  The site
with the lowest site ID is selected for this purpose.  So, if the
multi-homed site ID is 1 and single-homed site ID is 2, just for the
PW establishment, ID 1 is picked by all remote PEs.

Its an implementation detail, but here is a general explanation of why
there shouldn't be any traffic loss.  For clarity, consider:


              CE4 -------    ...............
                         \  .               .    ___ CE2
                      ___ PE1                .  /
                     /    :                  PE3
                  __/    :       Service      :
              CE1 __     :       Provider    PE4
                    \     :                   : \___ CE3
                     \___ PE2                .
                            .               .
                             ...............


PE1 has two sites, one multi-homed and other single-homed.  Site IDs
are 1 (for CE1) and 4 (for CE4). 

On PE2, there is just one site, with ID 1 for CE1 (multi-homed).

On PE3, there is just one site, with ID 2 for CE2

For simplicity, lets just consider one remote PE, PE3, for
explanation.  Also assume that PE1 has higher preference than PE2.

PE3 will establish just one PW to PE1.  For this PW, PE1 programs its
FIB with labels it advertises for site ID 1.  For this PW, PE3 selects
site ID 1 from the list of sites on PE1, and so the labels advertised
by PE1 for site ID 1 are used.  In other words, there is one PW
between PE1 and PE3, and just for the PW establishment, site 1 and 2
are picked on PE1 and PE3 ends.  PE3 will send traffic for both CE1
and CE4 over this PW.

Note that PE3 also has NLRIs for site ID 4 from PE1.  PE3 simply keeps
those in its control plane state and only programs its FIB with labels
for site ID 1.  

When a failure occurs, all PEs should do "make-before-break" as the
labels are already known.

If CE1-PE1 link goes, PE1 has to advertise the NLRI for site 1 with D
bit set.  Before it advertises this, it programs its FIB with labels
it advertises for site ID 4 (for CE4).  It also does NOT tear down
it's end of the old PW (which used site 1 labels).  When PE3 learns
about the D bit for site 1, it knows that there is additional site,
site 4, on PE1, and so it should not simply tear down the old PW.  It
should program the labels that it received from site 4, then it should
switch traffic to the new PW and only then it should tear down it's
end of the old PW.  The question of when PE1 tears down it's end of
the old PW is again an implementationd detail, but it could be dumb
timer based or better heuristic based.

Also note that the path selection that PE3 does on receiving the D bit
for site 1 from PE1 is independent of what is explained above.  The
path selection will elect PE2 as the forwarder for site 1 and so a new
PW to PE2 will be set up.


> Thanks,
> Florin

Hope it helps.

Thanks,
Bhupesh



From l2vpn-bounces@ietf.org  Wed Aug  6 15:11:40 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 77D4B3A6890;
	Wed,  6 Aug 2008 15:11:40 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 9C6EA3A682D
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 15:11:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.35
X-Spam-Level: 
X-Spam-Status: No, score=-6.35 tagged_above=-999 required=5 tests=[AWL=0.249, 
	BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id fXtCxckyU4G0 for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 15:11:37 -0700 (PDT)
Received: from exprod7og110.obsmtp.com (exprod7og110.obsmtp.com [64.18.2.173])
	by core3.amsl.com (Postfix) with ESMTP id 16F523A6890
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 15:10:52 -0700 (PDT)
Received: from source ([66.129.228.6]) by exprod7ob110.postini.com
	([64.18.6.12]) with SMTP; Wed, 06 Aug 2008 15:10:33 PDT
Received: from p-emlb01-sac.jnpr.net ([66.129.254.46]) by p-emsmtp02.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 15:09:35 -0700
Received: from emailsmtp55.jnpr.net ([172.24.18.132]) by p-emlb01-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Wed, 6 Aug 2008 15:09:35 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp55.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.1830); 
	Wed, 6 Aug 2008 15:09:34 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m76M9Yu07663;
	Wed, 6 Aug 2008 15:09:34 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m76M9Xlk014372;
	Wed, 6 Aug 2008 15:09:34 -0700
To: "Mallette\, Edwin" <Edwin.Mallette@bhnis.com>
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <DA0D6E76EA1D4D4F87F523AAF9E0CA5E373C65C2F5@CNEMAIL.corp.local>
References: <DA0D6E76EA1D4D4F87F523AAF9E0CA5E373C65C2F5@CNEMAIL.corp.local>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Date: Wed, 06 Aug 2008 15:09:33 -0700
Message-ID: <14371.1218060573@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 06 Aug 2008 22:09:34.0686 (UTC)
	FILETIME=[1BEF4FE0:01C8F811]
Cc: "l2vpn@ietf.org" <l2vpn@ietf.org>
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Ed,

Mallette, Edwin <Edwin.Mallette@bhnis.com> wrote:

> Hi Bhupesh,
>=20
> My comments are based on scenario 1:
>=20
>                              ...............
>=20
>                             .               .    ___ CE2
>=20
>                       ___ PE1                .  /
>=20
>                      /    :                  PE3
>=20
>                   __/    :       Service      :
>=20
>               CE1 __     :       Provider    PE4
>=20
>                     \     :                   : \___ CE3
>=20
>                      \___ PE2                .
>=20
>                             .               .
>=20
>                              ...............
>=20
> It doesn=E2=80=99t appear that the draft includes failure conditions
> accounted for.  As such, it strikes me the only the reasonable
> failure case in which PE1 becomes aware of the attachment circuit
> failure is when the PE1-to-CE1 AC loses link.  Is this an accurate
> assertion?

No.=20=20

>=20
> As it=E2=80=99s common to have some multi-access layer between PE1 and CE1
> that essentially hides any sort of link loss unless the link between
> PE1 and the multi-access device fails, how would we account for
> that?  With the spanning-tree model, there are some opportunities
> for control messages that would verify the forwarding path between
> PE1 and CE1.  I like to see some verbiage about how to determine PE1
> to CE1 and PE2 to CE1 path health in addition to the loop mitigation
> methods described in this draft.  Some sort of 802.1ag or 802.3ah
> OAM messaging between PE1 and CE1 and PE2 and CE1 to verify path
> health would seem to be a potential fit here.=20=20

Both 802.1ag and 802.3ah are a fit.  How you detect failure is
independent of what you do (path selection in this case) on detecting
the failure, and thats why its outside the scope of this draft.=20=20

> I=E2=80=99m not sure if adding that to this draft is appropriate, though I
> would think it=E2=80=99s reasonable to call out specific applicable use
> cases or to exclude some use cases based on the last-mile.

I am not clear.  Are you suggesting that we provide a list 802.1ag,
802.3ah, etc. that can be used for failure detection and what would
happen if there is lack of such mechanisms?

>=20
> Regards,
>=20
> Ed
>=20

Thanks
Bhupesh


From l2vpn-bounces@ietf.org  Wed Aug  6 20:36:09 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 4C46728C120;
	Wed,  6 Aug 2008 20:36:09 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 7A11A28C0FB
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 20:36:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.58
X-Spam-Level: 
X-Spam-Status: No, score=-0.58 tagged_above=-999 required=5 tests=[AWL=-0.086, 
	BAYES_00=-2.599, FH_RELAY_NODNS=1.451, HELO_MISMATCH_COM=0.553, 
	HTML_MESSAGE=0.001, RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id Lk0kC8Ni+RF7 for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 20:36:07 -0700 (PDT)
Received: from szxga04-in.huawei.com (unknown [119.145.14.67])
	by core3.amsl.com (Postfix) with ESMTP id 1BB3A28C120
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 20:36:07 -0700 (PDT)
Received: from huawei.com (szxga04-in [172.24.2.12])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5700HK0OOV0U@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 11:36:31 +0800 (CST)
Received: from huawei.com ([172.24.1.12])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5700096OOVOF@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 11:36:31 +0800 (CST)
Received: from j59929 ([10.70.77.93])
	by szxml05-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K57008OYOOV4L@szxml05-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 11:36:31 +0800 (CST)
Date: Thu, 07 Aug 2008 11:36:27 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Some reasonings on the draft
	"Optimized MAC Address Operations in VPLS with Redundancy"
To: mlasserre@alcatel-lucent.com
Message-id: <005001c8f83e$c67a1710$5d4d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_bDnpAoSqauct4nUCK7PNcA)"
X-Priority: 3
X-MSMail-priority: Normal
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_bDnpAoSqauct4nUCK7PNcA)
Content-type: text/plain; charset=gb2312
Content-transfer-encoding: 7BIT

Hi, Marc:

Thank you for your comments on the IETF 72 meeting, here I will give some more reasonings concerning 
with your comments.

The following is the excerpt from the 72nd IETF L2VPN WG minutes:
- Marc Lassere: There is already an existing proposal,
          (draft-pdutta-l2vpn-vpls-ldp-mac-opt-03), so I'm not sure what the
          benefits are vs. what has been proposed already to only flush relevant
          MAC addresses.
- Marc Lassere: In the proposal made, the PE-id is used to find out what
          the MAC addresses are to be flushed.  So, instead of blindly flushing,
          it's not an empty list, you only flush the ones behind the PE that
          failed.  This is as optimized as can be.  The second point is I'm a bit
          worried that you're suggesting to switch MAC addresses.  You don't want
          to create a new learning engine in VPLS, you want to learn from data
          plane as they come.  You don't want two FIB's, because you got some
          LDP-based message.
-end of the excerpt

With regard to draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, we have introduced it in brief in our draft.
Firstly, IMO, it applies only to a certain hub-and-spoke network topology, if CE is dual-homed to 
the PEs by ACs, it will be out of the capability of the draft.
Secondly, for the customer sites with hundreds of MACs, still a lot of MAC learning will be needed.
Thirdly, for the customer site with few MACs but large bandwidth, although few MACs are flushed, but 
as the time for the responsive frame to arrive at the PEs where MACs are flushed is not trivial, the
total frames flooded accross the core may be added to a high volume.
Therefore, draft-jiang-l2vpn-vpls-mac-operation-00 tries to eliminate the MAC leaning and subsequent 
flooding procedures due to PE or PW switching in a broader way.

For the 2nd question, I don't think we need two FIBs to provide MAC switching function.
For what I know, the original MAC address withdrawal process defined in RFC4762 first provides
the LDP-based message to flush specified MACs or all MACs except those associated with certain PWs 
in the FIB, so maybe there is no problem to pass MACs and PW information to the dataplane.

The only difference lies in the processing logic of the dataplane, which depends on the implementation.
Except "flush all MACs" blindly, selectively flushing or any other optimizations will somehow introduce 
complexities to the logic. IMO, from the view of computation comlexity, MAC switching is just the same
as RFC 4762.

The following are some snippets for the comparisons of the processing logic :
In RFC 4762: 
1) for each MAC in a set of MACs,
        lookup the FIB for this MAC and flush the corresponding FIB item.

2) Given PW1, for each item in the FIB, 
        if the item's outport is not the same as this PW, then flush it.

In our draft, 
1) Given PW1, for each MAC in a set of MACs,
        lookup the FIB for this MAC and change its outport to PW1.
2) Given PW1, PW2, for each item in the FIB, 
        if the item's outport is the same as PW1, then change it to PW2.


cheers,

Yuanlong


--Boundary_(ID_bDnpAoSqauct4nUCK7PNcA)
Content-type: text/html; charset=gb2312
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=gb2312">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT size=2>Hi, Marc:</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>Thank you for your comments on the IETF 72 meeting, here I 
will give some more reasonings concerning </FONT></DIV>
<DIV><FONT size=2>with your comments.</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV>
<DIV><FONT size=2>The following is the excerpt from the 72nd IETF L2VPN WG 
minutes:</FONT></DIV>
<DIV><FONT size=2>- Marc Lassere: There is already an existing 
proposal,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; (<A 
href="http://tools.ietf.org/html?draft=draft-pdutta-l2vpn-vpls-ldp-mac-opt-03">draft-pdutta-l2vpn-vpls-ldp-mac-opt-03</A>), 
so I'm not sure what 
the<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; benefits are vs. 
what has been proposed already to only flush 
relevant<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; MAC 
addresses.<BR>- Marc Lassere: In the proposal made, the PE-id is used to find 
out what<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; the MAC 
addresses are to be flushed.&nbsp; So, instead of blindly 
flushing,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; it's not an 
empty list, you only flush the ones behind the PE 
that<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; failed.&nbsp; 
This is as optimized as can be.&nbsp; The second point is I'm a 
bit<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; worried that 
you're suggesting to switch MAC addresses.&nbsp; You don't 
want<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to create a new 
learning engine in VPLS, you want to learn from 
data<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; plane as they 
come.&nbsp; You don't want two FIB's, because you got 
some<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; LDP-based 
message.<BR>-end of the excerpt</DIV></FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>With regard to draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, we have 
introduced it in brief in our draft.</FONT></DIV>
<DIV><FONT size=2>Firstly, IMO, it applies only&nbsp;to a certain hub-and-spoke 
network topology, if CE is </FONT><FONT size=2>dual-homed&nbsp;to </FONT></DIV>
<DIV><FONT size=2>the PEs by ACs, it will be out of the capability of the 
draft.</FONT></DIV>
<DIV><FONT size=2>Secondly, for the customer sites&nbsp;with hundreds of MACs, 
still a lot of MAC learning will be needed.</FONT></DIV>
<DIV><FONT size=2>Thirdly, for the customer site with few MACs but large 
bandwidth, although few MACs are flushed, but </FONT></DIV>
<DIV><FONT size=2>as </FONT><FONT size=2>the&nbsp;time for the responsive frame 
to arrive at the PEs where MACs are flushed&nbsp;is not trivial, 
the</FONT></DIV>
<DIV><FONT size=2>total frames flooded accross the core may be added to a high 
volume.</FONT></DIV>
<DIV><FONT size=2>Therefore, draft-jiang-l2vpn-vpls-mac-operation-00 tries to 
eliminate the MAC leaning and subsequent </FONT></DIV>
<DIV><FONT size=2>flooding </FONT><FONT size=2>procedures due to PE or 
</FONT><FONT size=2>PW </FONT><FONT size=2>switching in a broader 
way.</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>For the 2nd question, I don't think we need two FIBs to 
provide MAC switching function.</FONT></DIV>
<DIV><FONT size=2>For what I know, the original MAC address withdrawal process 
defined in RFC4762 first provides</FONT></DIV>
<DIV><FONT size=2>the LDP-based message to flush specified MACs or all MACs 
except those associated with certain PWs </FONT></DIV>
<DIV><FONT size=2>in </FONT><FONT size=2>the FIB, so maybe there is no problem 
to pass MACs </FONT><FONT size=2>and PW information to the 
dataplane.</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>The only difference lies in the processing logic of&nbsp;the 
dataplane, which depends on the implementation.</FONT></DIV>
<DIV><FONT size=2>Except "flush all MACs" blindly, selectively flushing or any 
other optimizations will somehow introduce </FONT></DIV>
<DIV><FONT size=2>complexities to the logic.&nbsp;IMO,&nbsp;from the view of 
computation comlexity, MAC switching is just the same</FONT></DIV>
<DIV><FONT size=2>as RFC 4762.</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>The following&nbsp;are some snippets for the comparisons of 
the processing logic&nbsp;:</FONT></DIV>
<DIV><FONT size=2>In RFC 4762: </FONT></DIV>
<DIV><FONT size=2>1) for each MAC in a set of MACs,</FONT></DIV>
<DIV><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;lookup the FIB 
for this MAC and flush the corresponding FIB item.</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>
<DIV><FONT size=2>2) Given PW1, for each&nbsp;item in the FIB, </FONT></DIV>
<DIV><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if 
the&nbsp;item's outport&nbsp;is not the same as this PW,&nbsp;then flush 
it.</FONT></DIV></FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>In our draft, </FONT></DIV>
<DIV><FONT size=2>1) Given PW1, for each MAC in a set of MACs,</FONT></DIV>
<DIV><FONT size=2>&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp; lookup the FIB for this 
MAC and change its outport to&nbsp;PW1.</FONT></DIV>
<DIV><FONT size=2>2) Given PW1, PW2,</FONT><FONT size=2> for each&nbsp;item in 
the FIB, </FONT></DIV>
<DIV><FONT size=2>
<DIV><FONT size=2>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if 
the&nbsp;item's outport&nbsp;is the same as PW1, </FONT>then change&nbsp;it 
to&nbsp;PW2.</FONT></DIV></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>cheers,</FONT></DIV>
<DIV><FONT size=2></FONT>&nbsp;</DIV>
<DIV><FONT size=2>Yuanlong</FONT></DIV>
<DIV>&nbsp;</DIV>
<DIV><FONT size=2>&nbsp;</DIV></FONT></BODY></HTML>

--Boundary_(ID_bDnpAoSqauct4nUCK7PNcA)--


From l2vpn-bounces@ietf.org  Wed Aug  6 20:59:48 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 03AF03A6BB3;
	Wed,  6 Aug 2008 20:59:48 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 24A523A6BB3
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 20:59:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.109
X-Spam-Level: 
X-Spam-Status: No, score=-0.109 tagged_above=-999 required=5
	tests=[AWL=-0.499, BAYES_00=-2.599, FH_RELAY_NODNS=1.451,
	HELO_MISMATCH_COM=0.553, HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001, 
	RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id x6rLjviIO59y for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 20:59:46 -0700 (PDT)
Received: from szxga04-in.huawei.com (unknown [119.145.14.67])
	by core3.amsl.com (Postfix) with ESMTP id BDA423A6B6B
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 20:59:45 -0700 (PDT)
Received: from huawei.com (szxga04-in [172.24.2.12])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5700H68PSA0U@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 12:00:10 +0800 (CST)
Received: from huawei.com ([172.24.1.12])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5700BBAPSA6F@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 12:00:10 +0800 (CST)
Received: from j59929 ([10.70.77.93])
	by szxml05-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K57008X6PS94L@szxml05-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 12:00:10 +0800 (CST)
Date: Thu, 07 Aug 2008 12:00:05 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE:
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <006c01c8f842$1347c620$5d4d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_yI5PvhRkAakbohs7oZOkXw)"
X-Priority: 3
X-MSMail-priority: Normal
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_yI5PvhRkAakbohs7oZOkXw)
Content-type: text/plain; charset=iso-8859-1
Content-transfer-encoding: 7BIT


Wim:

Thanks for you comment. My draft is not a standlone protocol, it can be deployed with other protocol suites.
For what I know, "PW redundancy" is being standardized in PWE3 WG, which can synchronize the behavior of
the PEs and PWs in PW switching. So it is not such a difficult task to do PW switching, and do MAC switching
subsequently. 

The MAC withdrawal process defined in RFC 4762 is not so simple as you think, it maybe not just "flush all MACs" 
blindly, but can selectively flush MACs, which computation comlexity is on the same level as my draft.
  ----- Original Message ----- 
  From: HENDERICKX Wim 
  To: Jiang Yuan-long ; sajassi@cisco.com 
  Cc: l2vpn@ietf.org 
  Sent: Thursday, August 07, 2008 4:18 AM
  Subject: RE:


  Yuanlong,

  The difficulty with your draft is that it requires a PE to do different things depending in which situation the PE is in, and this makes it very cumbersome in case changes happen to topologies or configurations.

  My view is that we need a simple MAC flush which we have today in LDP.

  Cheers,
  WIm



------------------------------------------------------------------------------
  From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
  Sent: woensdag 6 augustus 2008 11:30
  To: sajassi@cisco.com
  Cc: l2vpn@ietf.org
  Subject: 


  Hi, Ali:

  Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
  Following are the two questions:

  1. What do you think will be faster convergence? When you flush MAC addresses, 
  the learning is done in data plane at wire speed. Why does that prolong convergence time?


  -Faster convergence means the time to associate the MACs with the "exact" out port
  will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
  at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
  MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
  the wire speed, but also on the host processing time and delays on all the intermediate routers.
  For this reason, flooding may also last for a rather long time.


  2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
   distributed among multiple PE's, how do you take care of that?

  -I believe the most commonly deployed scenario is one backup PE for a primary PE.
  In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
  only one PE out of them will be elected as the working PE when the primary PE failed, 
  rather than distribute the traffic over multiple backup PEs,
  otherwise, loops may be introduced to the attached customer site.
  BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
  draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
  Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
  add them to the future drafts. 

  cheers,
  Yuanlong

--Boundary_(ID_yI5PvhRkAakbohs7oZOkXw)
Content-type: text/html; charset=iso-8859-1
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2>Wim:</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2>Thanks for you comment. My draft is not a standlone 
protocol, it can be deployed with other protocol suites.</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>For what I know, "PW redundancy" is being standardized 
in PWE3 WG, which can synchronize the behavior of</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>the PEs and PWs in PW switching. So it is 
not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC 
switching</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>subsequently. </FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2>The MAC withdrawal process defined in RFC 4762 is not 
so simple as you think,&nbsp;it maybe not just<FONT size=2>&nbsp;"flush all 
MACs" </FONT></FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2><FONT size=2>blindly, but can selectively flush MACs, 
which computation comlexity is on the same level as my draft</FONT><FONT 
size=2>.</FONT></DIV></FONT>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A 
  title=sajassi@cisco.com href="mailto:sajassi@cisco.com">sajassi@cisco.com</A> 
  </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, August 07, 2008 4:18 AM</DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE:</DIV>
  <DIV><BR></DIV>
  <DIV dir=ltr align=left>Yuanlong<SPAN class=924080318-06082008>,</SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008></SPAN>&nbsp;</DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2>The difficulty with your draft is that it requires a PE 
  to do different things depending in which situation the PE is in, and this 
  makes it very cumbersome in case changes happen to topologies or 
  configurations.</FONT></SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2>My view is that we need a simple MAC flush which we have 
  today in LDP.</FONT></SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2>Cheers,</FONT></SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2>WIm</FONT></SPAN></DIV><BR>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> <A 
  href="mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A> 
  [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang 
  Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> <A 
  href="mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> <A 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> 
  <BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face="Times New Roman">Thanks for your comments in the last L2VPN 
  WG meeting, I will try to give you some reasons here.</FONT></DIV>
  <DIV><FONT face="Times New Roman">Following are the&nbsp;two 
  questions:</FONT></DIV>
  <DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face="Times New Roman">1. What do you think will be faster 
  convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
  <DIV><FONT face="Times New Roman">the learning is done in data plane at wire 
  speed. Why does that prolong convergence time?<BR></FONT></DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face="Times New Roman">-Faster convergence means the time to 
  associate the MACs with the "exact" out port</FONT></DIV>
  <DIV><FONT face="Times New Roman">will be shorter. Though learning is done in 
  data plane at wire speed, the time for a frame to arrive </FONT></DIV>
  <DIV><FONT face="Times New Roman">at the end host and the response time for a 
  host to send a reverse frame&nbsp;accross the PE (on which the </FONT></DIV>
  <DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather long 
  time. So the learning process is not only determined solely by </FONT></DIV>
  <DIV><FONT face="Times New Roman">the wire </FONT><FONT 
  face="Times New Roman">speed, but also on the host processing time and delays 
  on all the intermediate routers.</FONT></DIV>
  <DIV><FONT face="Times New Roman">For this reason, flooding may also last for 
  a rather long time.</FONT></DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
  <DIV><FONT face="Times New Roman">2. You consider a simple case with simple 
  backup for primary.&nbsp;In cases where primary fails and traffic 
  gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
  that?</FONT></DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face="Times New Roman">-I believe the most commonly 
  deployed&nbsp;scenario is&nbsp;one backup PE for a primary PE.</FONT></DIV>
  <DIV><FONT face="Times New Roman">In my opinion, for the scenario of multiple 
  backup PEs protecting a primary PE, </FONT></DIV>
  <DIV><FONT face="Times New Roman">only one PE out of them will be elected as 
  the working PE when the primary PE failed, </FONT></DIV>
  <DIV><FONT face="Times New Roman">rather than distribute the traffic over 
  multiple backup PEs,</FONT></DIV>
  <DIV><FONT face="Times New Roman">otherwise, loops may be introduced to the 
  attached customer site.</FONT></DIV>
  <DIV><FONT face="Times New Roman">BTW, in a similar way, for PW redundancy 
  mechanism, only 1 secondary is selected as the working PW (see </FONT></DIV>
  <DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
  in&nbsp;PWE3 WG)</FONT></DIV>
  <DIV><FONT face="Times New Roman">Anyway, if there is any such requirements 
  confirmed, I think it will be OK for us to find some solutions and 
  </FONT></DIV>
  <DIV><FONT face="Times New Roman">add them to&nbsp;the future 
  drafts.&nbsp;</FONT></DIV>
  <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
  <DIV><FONT 
face="Times New Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BODY></HTML>

--Boundary_(ID_yI5PvhRkAakbohs7oZOkXw)--


From l2vpn-bounces@ietf.org  Wed Aug  6 21:48:24 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 24DAD3A6A73;
	Wed,  6 Aug 2008 21:48:24 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 8EC893A6A73
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 21:48:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.097
X-Spam-Level: 
X-Spam-Status: No, score=-2.097 tagged_above=-999 required=5 tests=[AWL=0.152, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id or6o16pIRGYN for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 21:48:21 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [62.23.212.27])
	by core3.amsl.com (Postfix) with ESMTP id 4C2943A6A10
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 21:48:20 -0700 (PDT)
Received: from FRVELSBHS04.ad2.ad.alcatel.com (frvelsbhs04.ad2.ad.alcatel.com
	[155.132.6.76])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m774mTqr022760; 
	Thu, 7 Aug 2008 06:48:32 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS04.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Thu, 7 Aug 2008 06:48:29 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: draft-kompella-l2vpn-vpls-multihoming-01
Date: Thu, 7 Aug 2008 06:48:33 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C61A@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <2018.1218049742@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj39+vcpnJgjlgNRXSgNizg3I220gAUNXmA
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com>
	<21222.1217875881@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com>
	<27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<2018.1218049742@bhupesh-f8.jnpr.net>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 07 Aug 2008 04:48:29.0007 (UTC)
	FILETIME=[D5E6DDF0:01C8F848]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Thx for clarification, which BGP message do you envisage to use for this
purpose ?

-----Original Message-----
From: Bhupesh Kothari [mailto:bhupesh@juniper.net]=20
Sent: woensdag 6 augustus 2008 21:09
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:
> > As you mentioned below we should extend the solution with a MAC=20
> > flush capability like LDP does today.
>=20
> Yes, we'll add the explicit flush.  The explicit flush capability=20
> provides optimization at the expense of more state in the control=20
> plane as now the PEs need to track the MACs learned in forwarding so=20
> that those can be withdrawn later on.
>=20
>=20
> WH> How would this work? E.g. in the example we use PE3 has no=20
> WH> awareness
> of the Mac @ belonging to which CE on a given PE.
>=20

I should have been more clear.  A MAC flush message from PE1 containing
a list of MACs will be sufficient for this purpose.

Bhupesh


From l2vpn-bounces@ietf.org  Wed Aug  6 22:24:27 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 279283A679F;
	Wed,  6 Aug 2008 22:24:27 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id F3F893A689D
	for <l2vpn@core3.amsl.com>; Wed,  6 Aug 2008 22:24:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.667
X-Spam-Level: 
X-Spam-Status: No, score=-1.667 tagged_above=-999 required=5
	tests=[AWL=-0.303, BAYES_00=-2.599, HELO_EQ_FR=0.35,
	HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id Qy5FxA99X26Y for <l2vpn@core3.amsl.com>;
	Wed,  6 Aug 2008 22:24:24 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [62.23.212.27])
	by core3.amsl.com (Postfix) with ESMTP id 5757D3A679F
	for <l2vpn@ietf.org>; Wed,  6 Aug 2008 22:24:23 -0700 (PDT)
Received: from FRVELSBHS02.ad2.ad.alcatel.com (frvelsbhs02.ad2.ad.alcatel.com
	[155.132.6.74])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m775OhBM032660; 
	Thu, 7 Aug 2008 07:24:43 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS02.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Thu, 7 Aug 2008 07:24:43 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8F84D.E562EC2B"
Subject: RE: RE:
Date: Thu, 7 Aug 2008 07:24:47 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <006c01c8f842$1347c620$5d4d460a@china.huawei.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: RE:
Thread-Index: Acj4R7X4MAYybe/iR7ClTp8XTfblXgABWarg
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Jiang Yuan-long" <yljiang@huawei.com>
X-OriginalArrivalTime: 07 Aug 2008 05:24:43.0183 (UTC)
	FILETIME=[E5CFCBF0:01C8F84D]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8F84D.E562EC2B
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Yuanlong,
=20
My point is that how are you going to define the behavior on each PE
when the topology is adapted. I hope you are not going to change the
configuration on each PE in the VPLS each time.
=20
On top, by doing selective MAC flush you need to go one by one to the
items and hence the processing to withdraw the MAC(s) will be longer.
When you use the existing mechanisms you can flush without the extra
lookups and hence the performance will be faster. FLushing a MAC will
result in re-learning which means you do not have traffic impact, just a
small flooding until the MAC address is re-learned.
=20
Cheers,
Wim

________________________________

From: Jiang Yuan-long [mailto:yljiang@huawei.com]=20
Sent: donderdag 7 augustus 2008 6:00
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: RE:


=20
Wim:
=20
Thanks for you comment. My draft is not a standlone protocol, it can be
deployed with other protocol suites.
For what I know, "PW redundancy" is being standardized in PWE3 WG, which
can synchronize the behavior of
the PEs and PWs in PW switching. So it is not such a difficult task to
do PW switching, and do MAC switching
subsequently.=20
=20
The MAC withdrawal process defined in RFC 4762 is not so simple as you
think, it maybe not just "flush all MACs"=20
blindly, but can selectively flush MACs, which computation comlexity is
on the same level as my draft.

	----- Original Message -----=20
	From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be> =20
	To: Jiang Yuan-long <mailto:yljiang@huawei.com>  ;
sajassi@cisco.com=20
	Cc: l2vpn@ietf.org=20
	Sent: Thursday, August 07, 2008 4:18 AM
	Subject: RE:

	Yuanlong,
	=20
	The difficulty with your draft is that it requires a PE to do
different things depending in which situation the PE is in, and this
makes it very cumbersome in case changes happen to topologies or
configurations.
	=20
	My view is that we need a simple MAC flush which we have today
in LDP.
	=20
	Cheers,
	WIm

________________________________

	From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On
Behalf Of Jiang Yuan-long
	Sent: woensdag 6 augustus 2008 11:30
	To: sajassi@cisco.com
	Cc: l2vpn@ietf.org
	Subject:=20
=09
=09
	Hi, Ali:
	=20
	Thanks for your comments in the last L2VPN WG meeting, I will
try to give you some reasons here.
	Following are the two questions:
	=20
	1. What do you think will be faster convergence? When you flush
MAC addresses,=20
	the learning is done in data plane at wire speed. Why does that
prolong convergence time?
=09
	=20
	-Faster convergence means the time to associate the MACs with
the "exact" out port
	will be shorter. Though learning is done in data plane at wire
speed, the time for a frame to arrive=20
	at the end host and the response time for a host to send a
reverse frame accross the PE (on which the=20
	MACs is flushed) will be a rather long time. So the learning
process is not only determined solely by=20
	the wire speed, but also on the host processing time and delays
on all the intermediate routers.
	For this reason, flooding may also last for a rather long time.
	=20
	=20
	2. You consider a simple case with simple backup for primary. In
cases where primary fails and traffic gets
	 distributed among multiple PE's, how do you take care of that?
	=20
	-I believe the most commonly deployed scenario is one backup PE
for a primary PE.
	In my opinion, for the scenario of multiple backup PEs
protecting a primary PE,=20
	only one PE out of them will be elected as the working PE when
the primary PE failed,=20
	rather than distribute the traffic over multiple backup PEs,
	otherwise, loops may be introduced to the attached customer
site.
	BTW, in a similar way, for PW redundancy mechanism, only 1
secondary is selected as the working PW (see=20
	draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
	Anyway, if there is any such requirements confirmed, I think it
will be OK for us to find some solutions and=20
	add them to the future drafts.=20
	=20
	cheers,
	Yuanlong
=09


------_=_NextPart_001_01C8F84D.E562EC2B
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Dus-ascii">
<META content=3D"MSHTML 6.00.2900.3354" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D770051905-07082008>
<DIV dir=3Dltr align=3Dleft><FONT face=3DArial><FONT =
color=3D#0000ff><FONT=20
size=3D2>Yuanlong<SPAN=20
class=3D924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><SPAN=20
class=3D770051905-07082008><FONT face=3DArial color=3D#0000ff =
size=3D2>My point is that=20
how are you going to define the behavior on each PE when the topology is =

adapted. I hope you are not going to change the configuration on each PE =
in the=20
VPLS each time.</FONT></SPAN></SPAN></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D770051905-07082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2>On top, by doing selective MAC flush you need =
to go one by=20
one to the items and hence the processing to withdraw the MAC(s) will be =
longer.=20
When you use the existing mechanisms you can flush without the extra =
lookups and=20
hence the performance will be faster. FLushing a MAC will result in =
re-learning=20
which means you do not have traffic impact, just a small flooding until =
the MAC=20
address is re-learned.</FONT></SPAN></DIV>
<DIV>&nbsp;</DIV>
<DIV><SPAN class=3D770051905-07082008></SPAN><FONT face=3DArial><FONT=20
color=3D#0000ff><FONT size=3D2>C<SPAN=20
class=3D770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
<DIV><FONT><FONT color=3D#0000ff><FONT size=3D2><SPAN=20
class=3D770051905-07082008></SPAN></FONT></FONT></FONT><SPAN=20
class=3D770051905-07082008></SPAN><FONT face=3DArial><FONT =
color=3D#0000ff><FONT=20
size=3D2>W<SPAN =
class=3D770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
<DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
<HR tabIndex=3D-1>
<FONT face=3DTahoma size=3D2><B>From:</B> Jiang Yuan-long=20
[mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008=20
6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B>=20
l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>Wim:</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>Thanks for you comment. My =
draft is not a standlone=20
protocol, it can be deployed with other protocol suites.</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>For what I know, "PW =
redundancy" is being standardized=20
in PWE3 WG, which can synchronize the behavior of</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>the PEs and PWs in PW =
switching. So it is=20
not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC=20
switching</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>subsequently. </FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>The MAC withdrawal process =
defined in RFC 4762 is not=20
so simple as you think,&nbsp;it maybe not just<FONT =
size=3D2>&nbsp;"flush all=20
MACs" </FONT></FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2><FONT size=3D2>blindly, but =
can selectively flush MACs,=20
which computation comlexity is on the same level as my draft</FONT><FONT =

size=3D2>.</FONT></DIV></FONT>
<BLOCKQUOTE dir=3Dltr=20
style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; =
BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style=3D"FONT: 9pt &#23435;&#20307;">----- Original Message ----- =
</DIV>
  <DIV style=3D"BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; =
font-color: black"><B>From:</B>=20
  <A title=3Dwim.henderickx@alcatel-lucent.be=20
  href=3D"mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> =
</DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>To:</B> <A =
title=3Dyljiang@huawei.com=20
  href=3D"mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A=20
  title=3Dsajassi@cisco.com =
href=3D"mailto:sajassi@cisco.com">sajassi@cisco.com</A>=20
  </DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A =
title=3Dl2vpn@ietf.org=20
  href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, =
August 07, 2008 4:18 AM</DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE:</DIV>
  <DIV><BR></DIV>
  <DIV dir=3Dltr align=3Dleft>Yuanlong<SPAN =
class=3D924080318-06082008>,</SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN =
class=3D924080318-06082008></SPAN>&nbsp;</DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2>The difficulty with your draft is that it =
requires a PE=20
  to do different things depending in which situation the PE is in, and =
this=20
  makes it very cumbersome in case changes happen to topologies or=20
  configurations.</FONT></SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2>My view is that we need a simple MAC flush =
which we have=20
  today in LDP.</FONT></SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2>Cheers,</FONT></SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2>WIm</FONT></SPAN></DIV><BR>
  <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
  <HR tabIndex=3D-1>
  <FONT face=3DTahoma size=3D2><B>From:</B> <A=20
  href=3D"mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A>=20
  [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang=20
  Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> =
<A=20
  href=3D"mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> =
<A=20
  href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B>=20
  <BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=3D"Times New Roman">Hi, Ali:</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D"Times New Roman">Thanks for your comments in the =
last L2VPN=20
  WG meeting, I will try to give you some reasons here.</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">Following are the&nbsp;two=20
  questions:</FONT></DIV>
  <DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D"Times New Roman">1. What do you think will be =
faster=20
  convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">the learning is done in data plane =
at wire=20
  speed. Why does that prolong convergence time?<BR></FONT></DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D"Times New Roman">-Faster convergence means the time =
to=20
  associate the MACs with the "exact" out port</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">will be shorter. Though learning =
is done in=20
  data plane at wire speed, the time for a frame to arrive </FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">at the end host and the response =
time for a=20
  host to send a reverse frame&nbsp;accross the PE (on which the =
</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">MACs&nbsp;is flushed) will be a =
rather long=20
  time. So the learning process is not only determined solely by =
</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">the wire </FONT><FONT=20
  face=3D"Times New Roman">speed, but also on the host processing time =
and delays=20
  on all the intermediate routers.</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">For this reason, flooding may also =
last for=20
  a rather long time.</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV></DIV>
  <DIV><FONT face=3D"Times New Roman">2. You consider a simple case with =
simple=20
  backup for primary.&nbsp;In cases where primary fails and traffic=20
  gets<BR>&nbsp;distributed among multiple PE's, how do you take care of =

  that?</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D"Times New Roman">-I believe the most commonly=20
  deployed&nbsp;scenario is&nbsp;one backup PE for a primary =
PE.</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">In my opinion, for the scenario of =
multiple=20
  backup PEs protecting a primary PE, </FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">only one PE out of them will be =
elected as=20
  the working PE when the primary PE failed, </FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">rather than distribute the traffic =
over=20
  multiple backup PEs,</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">otherwise, loops may be introduced =
to the=20
  attached customer site.</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">BTW, in a similar way, for PW =
redundancy=20
  mechanism, only 1 secondary is selected as the working PW (see =
</FONT></DIV>
  <DIV><FONT face=3D"Times New =
Roman">draft-ietf-pwe3-redundancy-bit-00.txt=20
  in&nbsp;PWE3 WG)</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">Anyway, if there is any such =
requirements=20
  confirmed, I think it will be OK for us to find some solutions and=20
  </FONT></DIV>
  <DIV><FONT face=3D"Times New Roman">add them to&nbsp;the future=20
  drafts.&nbsp;</FONT></DIV>
  <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D"Times New Roman">cheers,</FONT></DIV>
  <DIV><FONT=20
face=3D"Times New =
Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C8F84D.E562EC2B--


From l2vpn-bounces@ietf.org  Thu Aug  7 00:31:22 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 74CCD3A68EA;
	Thu,  7 Aug 2008 00:31:22 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 4CD303A6823
	for <l2vpn@core3.amsl.com>; Thu,  7 Aug 2008 00:31:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.19
X-Spam-Level: 
X-Spam-Status: No, score=-0.19 tagged_above=-999 required=5 tests=[AWL=0.324, 
	BAYES_00=-2.599, HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001, 
	J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id cQemyvD+v2uv for <l2vpn@core3.amsl.com>;
	Thu,  7 Aug 2008 00:31:20 -0700 (PDT)
Received: from szxga01-in.huawei.com (szxga01-in.huawei.com [119.145.14.64])
	by core3.amsl.com (Postfix) with ESMTP id 310DF3A68ED
	for <l2vpn@ietf.org>; Thu,  7 Aug 2008 00:31:19 -0700 (PDT)
Received: from huawei.com (szxga01-in [172.24.2.3])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K57002E7ZL3Z3@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 15:31:51 +0800 (CST)
Received: from huawei.com ([172.24.1.18])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5700159ZL258@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 15:31:50 +0800 (CST)
Received: from j59929 ([10.70.77.93])
	by szxml03-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5700KCCZL2YE@szxml03-in.huawei.com> for
	l2vpn@ietf.org; Thu, 07 Aug 2008 15:31:50 +0800 (CST)
Date: Thu, 07 Aug 2008 15:31:25 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE:
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <007d01c8f85f$99578350$5d4d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_Fa5B1bWAoXNikxirtdTWPg)"
X-Priority: 3
X-MSMail-priority: Normal
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_Fa5B1bWAoXNikxirtdTWPg)
Content-type: text/plain; charset=iso-8859-1
Content-transfer-encoding: 7BIT

Wim:

In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group, 
such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their 
respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective 
PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes 
working PW).
The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other 
configurations need to be changed, can you give me some more clues?


RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except 
those associated with specific PWs. So what you mean may be that all the existing implementations only provide 
flushing "all" MACs blindly in a VPLS instance, is that true?
As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.


Best regards,

Yuanlong
  ----- Original Message ----- 
  From: HENDERICKX Wim 
  To: Jiang Yuan-long 
  Cc: l2vpn@ietf.org 
  Sent: Thursday, August 07, 2008 1:24 PM
  Subject: RE: RE:


  Yuanlong,

  My point is that how are you going to define the behavior on each PE when the topology is adapted. I hope you are not going to change the configuration on each PE in the VPLS each time.

  On top, by doing selective MAC flush you need to go one by one to the items and hence the processing to withdraw the MAC(s) will be longer. When you use the existing mechanisms you can flush without the extra lookups and hence the performance will be faster. FLushing a MAC will result in re-learning which means you do not have traffic impact, just a small flooding until the MAC address is re-learned.

  Cheers,
  Wim


------------------------------------------------------------------------------
  From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
  Sent: donderdag 7 augustus 2008 6:00
  To: HENDERICKX Wim
  Cc: l2vpn@ietf.org
  Subject: Re: RE:



  Wim:

  Thanks for you comment. My draft is not a standlone protocol, it can be deployed with other protocol suites.
  For what I know, "PW redundancy" is being standardized in PWE3 WG, which can synchronize the behavior of
  the PEs and PWs in PW switching. So it is not such a difficult task to do PW switching, and do MAC switching
  subsequently. 

  The MAC withdrawal process defined in RFC 4762 is not so simple as you think, it maybe not just "flush all MACs" 
  blindly, but can selectively flush MACs, which computation comlexity is on the same level as my draft.
    ----- Original Message ----- 
    From: HENDERICKX Wim 
    To: Jiang Yuan-long ; sajassi@cisco.com 
    Cc: l2vpn@ietf.org 
    Sent: Thursday, August 07, 2008 4:18 AM
    Subject: RE:


    Yuanlong,

    The difficulty with your draft is that it requires a PE to do different things depending in which situation the PE is in, and this makes it very cumbersome in case changes happen to topologies or configurations.

    My view is that we need a simple MAC flush which we have today in LDP.

    Cheers,
    WIm



----------------------------------------------------------------------------
    From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
    Sent: woensdag 6 augustus 2008 11:30
    To: sajassi@cisco.com
    Cc: l2vpn@ietf.org
    Subject: 


    Hi, Ali:

    Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
    Following are the two questions:

    1. What do you think will be faster convergence? When you flush MAC addresses, 
    the learning is done in data plane at wire speed. Why does that prolong convergence time?


    -Faster convergence means the time to associate the MACs with the "exact" out port
    will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
    at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
    MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
    the wire speed, but also on the host processing time and delays on all the intermediate routers.
    For this reason, flooding may also last for a rather long time.


    2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
     distributed among multiple PE's, how do you take care of that?

    -I believe the most commonly deployed scenario is one backup PE for a primary PE.
    In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
    only one PE out of them will be elected as the working PE when the primary PE failed, 
    rather than distribute the traffic over multiple backup PEs,
    otherwise, loops may be introduced to the attached customer site.
    BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
    draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
    Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
    add them to the future drafts. 

    cheers,
    Yuanlong

--Boundary_(ID_Fa5B1bWAoXNikxirtdTWPg)
Content-type: text/html; charset=iso-8859-1
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=&#23435;&#20307; size=2>Wim:</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2>In VPLS with dual-homed topology, we 
must&nbsp;pre-configure the working PE and backup PE to form a protection 
group,&nbsp;</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>such as </FONT><FONT face=&#23435;&#20307; size=2>described in ICCP 
(defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT face=&#23435;&#20307; size=2>and 
pre-configure working PWs with their </FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>respective backup PWs, </FONT><FONT face=&#23435;&#20307; size=2>as 
</FONT><FONT face=&#23435;&#20307; size=2>described in "PW 
redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>Anyway, these configurations are prerequesite to 
provide sevice resiliency for VPLS, and the roles of the respective 
</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>PE or PW will </FONT><FONT face=&#23435;&#20307; size=2>change 
automatically after PE or PW switching (ie. backup PE becomes working PE, backup 
PW becomes </FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>working PW).</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>The mechanism in&nbsp;my draft&nbsp;is based on such 
a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure what other 
</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>configurations need to </FONT><FONT face=&#23435;&#20307; size=2>be 
changed, can you give me some more clues?</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2>RFC 4762 already provides some selective flushing 
mechansms, such as flushing some specified MACs or all MACs except </FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>those associated with </FONT><FONT face=&#23435;&#20307; 
size=2>specific PWs. So what</FONT><FONT face=&#23435;&#20307; size=2> you mean may be that 
all the existing implementations only&nbsp;provide </FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>flushing </FONT><FONT face=&#23435;&#20307; size=2>"all" MACs 
blindly in a VPLS instance, </FONT><FONT face=&#23435;&#20307; size=2>is that 
true?</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>As our anlysis and other sources(such as 
draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>impacts all sites in a VPLS instance, even their 
topology is not </FONT><FONT face=&#23435;&#20307; size=2>affected, the MACs must be re-leaned 
may be numerous,</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>and the number of frames flooded may be not so small. 
Thus blindly flushing is unacceptable to a carrier grade VPLS </FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2>service,</FONT><FONT face=&#23435;&#20307; size=2>and&nbsp;for 
this&nbsp;reason&nbsp;</FONT><FONT face=&#23435;&#20307; size=2>we proposed some optimized MAC 
operations such as MAC </FONT><FONT face=&#23435;&#20307; size=2>Switching and MAC 
notification.</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><BR><FONT face=&#23435;&#20307; size=2>Best 
regards,</FONT></DIV>
<DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
<DIV><FONT face=&#23435;&#20307; size=2>Yuanlong</FONT></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, August 07, 2008 1:24 PM</DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE: RE:</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT><BR></DIV>
  <DIV dir=ltr align=left><SPAN class=770051905-07082008>
  <DIV dir=ltr align=left><FONT face=Arial><FONT color=#0000ff><FONT 
  size=2>Yuanlong<SPAN 
  class=924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><SPAN 
  class=770051905-07082008><FONT face=Arial color=#0000ff size=2>My point is 
  that how are you going to define the behavior on each PE when the topology is 
  adapted. I hope you are not going to change the configuration on each PE in 
  the VPLS each time.</FONT></SPAN></SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
  color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=770051905-07082008><FONT face=Arial 
  color=#0000ff size=2>On top, by doing selective MAC flush you need to go one 
  by one to the items and hence the processing to withdraw the MAC(s) will be 
  longer. When you use the existing mechanisms you can flush without the extra 
  lookups and hence the performance will be faster. FLushing a MAC will result 
  in re-learning which means you do not have traffic impact, just a small 
  flooding until the MAC address is re-learned.</FONT></SPAN></DIV>
  <DIV>&nbsp;</DIV>
  <DIV><SPAN class=770051905-07082008></SPAN><FONT face=Arial><FONT 
  color=#0000ff><FONT size=2>C<SPAN 
  class=770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
  <DIV><FONT size=+0><FONT color=#0000ff><FONT size=2><SPAN 
  class=770051905-07082008></SPAN></FONT></FONT></FONT><SPAN 
  class=770051905-07082008></SPAN><FONT face=Arial><FONT color=#0000ff><FONT 
  size=2>W<SPAN 
class=770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
  6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
  l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>Wim:</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>Thanks for you comment. My draft is not a standlone 
  protocol, it can be deployed with other protocol suites.</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>For what I know, "PW redundancy" is being 
  standardized in PWE3 WG, which can synchronize the behavior of</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>the PEs and PWs in PW switching. So it is 
  not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC 
  switching</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>subsequently. </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>The MAC withdrawal process defined in RFC 4762 is 
  not so simple as you think,&nbsp;it maybe not just<FONT size=2>&nbsp;"flush 
  all MACs" </FONT></FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2><FONT size=2>blindly, but can selectively flush 
  MACs, which computation comlexity is on the same level as my draft</FONT><FONT 
  size=2>.</FONT></DIV></FONT>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
    <DIV 
    style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> <A 
    title=wim.henderickx@alcatel-lucent.be 
    href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
    href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A 
    title=sajassi@cisco.com 
    href="mailto:sajassi@cisco.com">sajassi@cisco.com</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, August 07, 2008 4:18 
    AM</DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE:</DIV>
    <DIV><BR></DIV>
    <DIV dir=ltr align=left>Yuanlong<SPAN 
class=924080318-06082008>,</SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008></SPAN>&nbsp;</DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2>The difficulty with your draft is that it requires a PE 
    to do different things depending in which situation the PE is in, and this 
    makes it very cumbersome in case changes happen to topologies or 
    configurations.</FONT></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2>My view is that we need a simple MAC flush which we 
    have today in LDP.</FONT></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2>Cheers,</FONT></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2>WIm</FONT></SPAN></DIV><BR>
    <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
    <HR tabIndex=-1>
    <FONT face=Tahoma size=2><B>From:</B> <A 
    href="mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A> 
    [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang 
    Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> <A 
    href="mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> <A 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> 
    <BR></FONT><BR></DIV>
    <DIV></DIV>
    <DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman">Thanks for your comments in the last L2VPN 
    WG meeting, I will try to give you some reasons here.</FONT></DIV>
    <DIV><FONT face="Times New Roman">Following are the&nbsp;two 
    questions:</FONT></DIV>
    <DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman">1. What do you think will be faster 
    convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
    <DIV><FONT face="Times New Roman">the learning is done in data plane at wire 
    speed. Why does that prolong convergence time?<BR></FONT></DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman">-Faster convergence means the time to 
    associate the MACs with the "exact" out port</FONT></DIV>
    <DIV><FONT face="Times New Roman">will be shorter. Though learning is done 
    in data plane at wire speed, the time for a frame to arrive </FONT></DIV>
    <DIV><FONT face="Times New Roman">at the end host and the response time for 
    a host to send a reverse frame&nbsp;accross the PE (on which the 
    </FONT></DIV>
    <DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather 
    long time. So the learning process is not only determined solely by 
    </FONT></DIV>
    <DIV><FONT face="Times New Roman">the wire </FONT><FONT 
    face="Times New Roman">speed, but also on the host processing time and 
    delays on all the intermediate routers.</FONT></DIV>
    <DIV><FONT face="Times New Roman">For this reason, flooding may also last 
    for a rather long time.</FONT></DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
    <DIV><FONT face="Times New Roman">2. You consider a simple case with simple 
    backup for primary.&nbsp;In cases where primary fails and traffic 
    gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
    that?</FONT></DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman">-I believe the most commonly 
    deployed&nbsp;scenario is&nbsp;one backup PE for a primary PE.</FONT></DIV>
    <DIV><FONT face="Times New Roman">In my opinion, for the scenario of 
    multiple backup PEs protecting a primary PE, </FONT></DIV>
    <DIV><FONT face="Times New Roman">only one PE out of them will be elected as 
    the working PE when the primary PE failed, </FONT></DIV>
    <DIV><FONT face="Times New Roman">rather than distribute the traffic over 
    multiple backup PEs,</FONT></DIV>
    <DIV><FONT face="Times New Roman">otherwise, loops may be introduced to the 
    attached customer site.</FONT></DIV>
    <DIV><FONT face="Times New Roman">BTW, in a similar way, for PW redundancy 
    mechanism, only 1 secondary is selected as the working PW (see </FONT></DIV>
    <DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
    in&nbsp;PWE3 WG)</FONT></DIV>
    <DIV><FONT face="Times New Roman">Anyway, if there is any such requirements 
    confirmed, I think it will be OK for us to find some solutions and 
    </FONT></DIV>
    <DIV><FONT face="Times New Roman">add them to&nbsp;the future 
    drafts.&nbsp;</FONT></DIV>
    <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
    <DIV><FONT 
face="Times New Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

--Boundary_(ID_Fa5B1bWAoXNikxirtdTWPg)--


From l2vpn-bounces@ietf.org  Thu Aug  7 12:39:48 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id D2CE83A67EB;
	Thu,  7 Aug 2008 12:39:48 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id B57893A6783
	for <l2vpn@core3.amsl.com>; Thu,  7 Aug 2008 12:39:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.757
X-Spam-Level: **
X-Spam-Status: No, score=2.757 tagged_above=-999 required=5
	tests=[HELO_EQ_FR=0.35, HTML_FONT_FACE_BAD=0.606, HTML_MESSAGE=0.001,
	J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id fX9XjYN5H-bx for <l2vpn@core3.amsl.com>;
	Thu,  7 Aug 2008 12:39:46 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [64.208.49.27])
	by core3.amsl.com (Postfix) with ESMTP id A2C783A6765
	for <l2vpn@ietf.org>; Thu,  7 Aug 2008 12:39:45 -0700 (PDT)
Received: from FRVELSBHS06.ad2.ad.alcatel.com (frvelsbhs06.ad2.ad.alcatel.com
	[155.132.6.78])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m77Jdd6m023194; 
	Thu, 7 Aug 2008 21:39:39 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS06.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Thu, 7 Aug 2008 21:39:39 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8F8C5.548D181E"
Subject: RE: RE:
Date: Thu, 7 Aug 2008 21:39:45 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C877@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <007d01c8f85f$99578350$5d4d460a@china.huawei.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: RE:
Thread-Index: Acj4ZUidRCnQXfUDSqeiUJLUWrrw0QAX3vBg
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
	<007d01c8f85f$99578350$5d4d460a@china.huawei.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Jiang Yuan-long" <yljiang@huawei.com>
X-OriginalArrivalTime: 07 Aug 2008 19:39:39.0780 (UTC)
	FILETIME=[54F75440:01C8F8C5]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8F8C5.548D181E
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Yuanlong, see in-line

________________________________

From: Jiang Yuan-long [mailto:yljiang@huawei.com]=20
Sent: donderdag 7 augustus 2008 9:31
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: RE:


Wim:
=20
In VPLS with dual-homed topology, we must pre-configure the working PE
and backup PE to form a protection group,=20
such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and
pre-configure working PWs with their=20
respective backup PWs, as described in "PW
redundancy"(draft-ietf-pwe3-redundancy-bit-00).
Anyway, these configurations are prerequesite to provide sevice
resiliency for VPLS, and the roles of the respective=20
PE or PW will change automatically after PE or PW switching (ie. backup
PE becomes working PE, backup PW becomes=20
working PW).
The mechanism in my draft is based on such a protection relationship of
PEs and PWs, I am not sure what other=20
configurations need to be changed, can you give me some more clues?
=20
WH> We don;t need any of this for PW resilience in VPLS according to me
as PW status together with the existing MAC flushes do the job for these
topologies.
=20
=20
RFC 4762 already provides some selective flushing mechansms, such as
flushing some specified MACs or all MACs except=20
those associated with specific PWs. So what you mean may be that all the
existing implementations only provide=20
flushing "all" MACs blindly in a VPLS instance, is that true?=20
=20
WH> There are certain cases where we need to extend them like for PBB
where we have C-MAC ties to B-MACs.
=20
As our anlysis and other sources(such as
draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing
adversely=20
impacts all sites in a VPLS instance, even their topology is not
affected, the MACs must be re-leaned may be numerous,
and the number of frames flooded may be not so small. Thus blindly
flushing is unacceptable to a carrier grade VPLS=20
service,and for this reason we proposed some optimized MAC operations
such as MAC Switching and MAC notification.
=20
WH> The optimizations which are suggested in the mentioned drafts are
very simple, whilst your proposal is fairly complex and depends on the
topology you propose but might not be efficient in case we change the
service topology.

Best regards,
=20
Yuanlong

	----- Original Message -----=20
	From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be> =20
	To: Jiang Yuan-long <mailto:yljiang@huawei.com> =20
	Cc: l2vpn@ietf.org=20
	Sent: Thursday, August 07, 2008 1:24 PM
	Subject: RE: RE:
=09
=09
=09
	Yuanlong,
	=20
	My point is that how are you going to define the behavior on
each PE when the topology is adapted. I hope you are not going to change
the configuration on each PE in the VPLS each time.
	=20
	On top, by doing selective MAC flush you need to go one by one
to the items and hence the processing to withdraw the MAC(s) will be
longer. When you use the existing mechanisms you can flush without the
extra lookups and hence the performance will be faster. FLushing a MAC
will result in re-learning which means you do not have traffic impact,
just a small flooding until the MAC address is re-learned.
	=20
	Cheers,
	Wim
=09
________________________________

	From: Jiang Yuan-long [mailto:yljiang@huawei.com]=20
	Sent: donderdag 7 augustus 2008 6:00
	To: HENDERICKX Wim
	Cc: l2vpn@ietf.org
	Subject: Re: RE:
=09
=09
	=20
	Wim:
	=20
	Thanks for you comment. My draft is not a standlone protocol, it
can be deployed with other protocol suites.
	For what I know, "PW redundancy" is being standardized in PWE3
WG, which can synchronize the behavior of
	the PEs and PWs in PW switching. So it is not such a difficult
task to do PW switching, and do MAC switching
	subsequently.=20
	=20
	The MAC withdrawal process defined in RFC 4762 is not so simple
as you think, it maybe not just "flush all MACs"=20
	blindly, but can selectively flush MACs, which computation
comlexity is on the same level as my draft.

		----- Original Message -----=20
		From: HENDERICKX Wim
<mailto:wim.henderickx@alcatel-lucent.be> =20
		To: Jiang Yuan-long <mailto:yljiang@huawei.com>  ;
sajassi@cisco.com=20
		Cc: l2vpn@ietf.org=20
		Sent: Thursday, August 07, 2008 4:18 AM
		Subject: RE:

		Yuanlong,
		=20
		The difficulty with your draft is that it requires a PE
to do different things depending in which situation the PE is in, and
this makes it very cumbersome in case changes happen to topologies or
configurations.
		=20
		My view is that we need a simple MAC flush which we have
today in LDP.
		=20
		Cheers,
		WIm

________________________________

		From: l2vpn-bounces@ietf.org
[mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
		Sent: woensdag 6 augustus 2008 11:30
		To: sajassi@cisco.com
		Cc: l2vpn@ietf.org
		Subject:=20
	=09
	=09
		Hi, Ali:
		=20
		Thanks for your comments in the last L2VPN WG meeting, I
will try to give you some reasons here.
		Following are the two questions:
		=20
		1. What do you think will be faster convergence? When
you flush MAC addresses,=20
		the learning is done in data plane at wire speed. Why
does that prolong convergence time?
	=09
		=20
		-Faster convergence means the time to associate the MACs
with the "exact" out port
		will be shorter. Though learning is done in data plane
at wire speed, the time for a frame to arrive=20
		at the end host and the response time for a host to send
a reverse frame accross the PE (on which the=20
		MACs is flushed) will be a rather long time. So the
learning process is not only determined solely by=20
		the wire speed, but also on the host processing time and
delays on all the intermediate routers.
		For this reason, flooding may also last for a rather
long time.
		=20
		=20
		2. You consider a simple case with simple backup for
primary. In cases where primary fails and traffic gets
		 distributed among multiple PE's, how do you take care
of that?
		=20
		-I believe the most commonly deployed scenario is one
backup PE for a primary PE.
		In my opinion, for the scenario of multiple backup PEs
protecting a primary PE,=20
		only one PE out of them will be elected as the working
PE when the primary PE failed,=20
		rather than distribute the traffic over multiple backup
PEs,
		otherwise, loops may be introduced to the attached
customer site.
		BTW, in a similar way, for PW redundancy mechanism, only
1 secondary is selected as the working PW (see=20
		draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
		Anyway, if there is any such requirements confirmed, I
think it will be OK for us to find some solutions and=20
		add them to the future drafts.=20
		=20
		cheers,
		Yuanlong
	=09


------_=_NextPart_001_01C8F8C5.548D181E
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Dus-ascii">
<META content=3D"MSHTML 6.00.2900.3354" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV dir=3Dltr align=3Dleft><FONT face=3DArial>
<DIV><FONT face=3DSimSun><FONT color=3D#0000ff><FONT =
size=3D2>Yuanlong<SPAN=20
class=3D289373519-07082008>, see=20
in-line</SPAN></FONT></FONT></FONT></DIV></FONT></DIV><BR>
<DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
<HR tabIndex=3D-1>
<FONT face=3DTahoma size=3D2><B>From:</B> Jiang Yuan-long=20
[mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008=20
9:31<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B>=20
l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>Wim:</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>In VPLS with dual-homed =
topology, we=20
must&nbsp;pre-configure the working PE and backup PE to form a =
protection=20
group,&nbsp;</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>such as </FONT><FONT =
face=3D&#23435;&#20307; size=3D2>described in ICCP=20
(defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT =
face=3D&#23435;&#20307; size=3D2>and=20
pre-configure working PWs with their </FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>respective backup PWs, =
</FONT><FONT face=3D&#23435;&#20307; size=3D2>as=20
</FONT><FONT face=3D&#23435;&#20307; size=3D2>described in "PW=20
redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>Anyway, these configurations =
are prerequesite to=20
provide sevice resiliency for VPLS, and the roles of the respective=20
</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>PE or PW will </FONT><FONT =
face=3D&#23435;&#20307; size=3D2>change=20
automatically after PE or PW switching (ie. backup PE becomes working =
PE, backup=20
PW becomes </FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>working PW).</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>The mechanism in&nbsp;my =
draft&nbsp;is based on such=20
a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure what =
other=20
</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>configurations need to =
</FONT><FONT face=3D&#23435;&#20307; size=3D2>be=20
changed, can you give me some more clues?</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3DSimSun size=3D2><SPAN =
class=3D289373519-07082008>WH&gt; We don;t=20
need any of this for PW resilience in VPLS according to me as PW status =
together=20
with the existing MAC flushes do the job for these=20
topologies.</SPAN></FONT></DIV>
<DIV><FONT face=3DSimSun size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>RFC 4762 already provides =
some selective flushing=20
mechansms, such as flushing some specified MACs or all MACs except =
</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>those associated with =
</FONT><FONT face=3D&#23435;&#20307;=20
size=3D2>specific PWs. So what</FONT><FONT face=3D&#23435;&#20307; =
size=3D2> you mean may be that=20
all the existing implementations only&nbsp;provide </FONT></DIV>
<DIV><FONT size=3D2><FONT face=3D&#23435;&#20307;>flushing </FONT><FONT =
face=3D&#23435;&#20307;>"all" MACs blindly=20
in a VPLS instance, </FONT><FONT face=3D&#23435;&#20307;>is that =
true?<SPAN=20
class=3D289373519-07082008><FONT face=3DArial=20
color=3D#0000ff>&nbsp;</FONT></SPAN></FONT></FONT></DIV>
<DIV><FONT size=3D2><FONT face=3D&#23435;&#20307;><SPAN=20
class=3D289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
<DIV><FONT size=3D2><FONT face=3D&#23435;&#20307;><SPAN =
class=3D289373519-07082008>WH&gt; There are=20
certain cases where we need to extend them like for PBB where we have =
C-MAC ties=20
to B-MACs.</SPAN></FONT></FONT></DIV>
<DIV><FONT size=3D2><FONT face=3D&#23435;&#20307;><SPAN=20
class=3D289373519-07082008>&nbsp;</SPAN></FONT></FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>As our anlysis and other =
sources(such as=20
draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing =
adversely=20
</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>impacts all sites in a VPLS =
instance, even their=20
topology is not </FONT><FONT face=3D&#23435;&#20307; size=3D2>affected, =
the MACs must be re-leaned=20
may be numerous,</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>and the number of frames =
flooded may be not so small.=20
Thus blindly flushing is unacceptable to a carrier grade VPLS =
</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>service,</FONT><FONT =
face=3D&#23435;&#20307; size=3D2>and&nbsp;for=20
this&nbsp;reason&nbsp;</FONT><FONT face=3D&#23435;&#20307; size=3D2>we =
proposed some optimized MAC=20
operations such as MAC </FONT><FONT face=3D&#23435;&#20307; =
size=3D2>Switching and MAC=20
notification.</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><SPAN class=3D289373519-07082008><FONT face=3DSimSun =
size=3D2>WH&gt; The=20
optimizations which are suggested in the mentioned drafts are very =
simple,=20
whilst your proposal is fairly complex and depends on the topology you =
propose=20
but might not be efficient in case we change the service=20
topology.</FONT></SPAN></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT><FONT =
face=3D&#23435;&#20307; size=3D2></FONT><FONT face=3D&#23435;&#20307;=20
size=3D2></FONT><FONT face=3D&#23435;&#20307; size=3D2></FONT><FONT =
face=3D&#23435;&#20307; size=3D2></FONT><FONT=20
face=3D&#23435;&#20307; size=3D2></FONT><FONT face=3D&#23435;&#20307; =
size=3D2></FONT><FONT face=3D&#23435;&#20307;=20
size=3D2></FONT><FONT face=3D&#23435;&#20307; size=3D2></FONT><FONT =
face=3DSimSun=20
size=3D2></FONT><BR><FONT face=3D&#23435;&#20307; size=3D2>Best =
regards,</FONT></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>Yuanlong</FONT></DIV>
<BLOCKQUOTE dir=3Dltr=20
style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; =
BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style=3D"FONT: 9pt &#23435;&#20307;">----- Original Message ----- =
</DIV>
  <DIV style=3D"BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; =
font-color: black"><B>From:</B>=20
  <A title=3Dwim.henderickx@alcatel-lucent.be=20
  href=3D"mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> =
</DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>To:</B> <A =
title=3Dyljiang@huawei.com=20
  href=3D"mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A =
title=3Dl2vpn@ietf.org=20
  href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, =
August 07, 2008 1:24 PM</DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE: =
RE:</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT><BR></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D770051905-07082008>
  <DIV dir=3Dltr align=3Dleft><FONT face=3DArial><FONT =
color=3D#0000ff><FONT=20
  size=3D2>Yuanlong<SPAN=20
  class=3D924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><SPAN=20
  class=3D770051905-07082008><FONT face=3DArial color=3D#0000ff =
size=3D2>My point is=20
  that how are you going to define the behavior on each PE when the =
topology is=20
  adapted. I hope you are not going to change the configuration on each =
PE in=20
  the VPLS each time.</FONT></SPAN></SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D770051905-07082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2>On top, by doing selective MAC flush you need =
to go one=20
  by one to the items and hence the processing to withdraw the MAC(s) =
will be=20
  longer. When you use the existing mechanisms you can flush without the =
extra=20
  lookups and hence the performance will be faster. FLushing a MAC will =
result=20
  in re-learning which means you do not have traffic impact, just a =
small=20
  flooding until the MAC address is re-learned.</FONT></SPAN></DIV>
  <DIV>&nbsp;</DIV>
  <DIV><SPAN class=3D770051905-07082008></SPAN><FONT face=3DArial><FONT=20
  color=3D#0000ff><FONT size=3D2>C<SPAN=20
  class=3D770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
  <DIV><FONT size=3D+0><FONT color=3D#0000ff><FONT size=3D2><SPAN=20
  class=3D770051905-07082008></SPAN></FONT></FONT></FONT><SPAN=20
  class=3D770051905-07082008></SPAN><FONT face=3DArial><FONT =
color=3D#0000ff><FONT=20
  size=3D2>W<SPAN=20
class=3D770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
  <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
  <HR tabIndex=3D-1>
  <FONT face=3DTahoma size=3D2><B>From:</B> Jiang Yuan-long=20
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 =

  6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B>=20
  l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>Wim:</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>Thanks for you comment. My =
draft is not a standlone=20
  protocol, it can be deployed with other protocol suites.</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>For what I know, "PW =
redundancy" is being=20
  standardized in PWE3 WG, which can synchronize the behavior =
of</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>the PEs and PWs in PW =
switching. So it is=20
  not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC=20
  switching</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>subsequently. =
</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>The MAC withdrawal process =
defined in RFC 4762 is=20
  not so simple as you think,&nbsp;it maybe not just<FONT =
size=3D2>&nbsp;"flush=20
  all MACs" </FONT></FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2><FONT size=3D2>blindly, =
but can selectively flush=20
  MACs, which computation comlexity is on the same level as my =
draft</FONT><FONT=20
  size=3D2>.</FONT></DIV></FONT>
  <BLOCKQUOTE dir=3Dltr=20
  style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; =
BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style=3D"FONT: 9pt &#23435;&#20307;">----- Original Message =
----- </DIV>
    <DIV=20
    style=3D"BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; =
font-color: black"><B>From:</B> <A=20
    title=3Dwim.henderickx@alcatel-lucent.be=20
    href=3D"mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> =
</DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>To:</B> <A =
title=3Dyljiang@huawei.com=20
    href=3D"mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A=20
    title=3Dsajassi@cisco.com=20
    href=3D"mailto:sajassi@cisco.com">sajassi@cisco.com</A> </DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A =
title=3Dl2vpn@ietf.org=20
    href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, =
August 07, 2008 4:18=20
    AM</DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE:</DIV>
    <DIV><BR></DIV>
    <DIV dir=3Dltr align=3Dleft>Yuanlong<SPAN=20
class=3D924080318-06082008>,</SPAN></DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN =
class=3D924080318-06082008></SPAN>&nbsp;</DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
    color=3D#0000ff size=3D2>The difficulty with your draft is that it =
requires a PE=20
    to do different things depending in which situation the PE is in, =
and this=20
    makes it very cumbersome in case changes happen to topologies or=20
    configurations.</FONT></SPAN></DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
    color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
    color=3D#0000ff size=3D2>My view is that we need a simple MAC flush =
which we=20
    have today in LDP.</FONT></SPAN></DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
    color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
    color=3D#0000ff size=3D2>Cheers,</FONT></SPAN></DIV>
    <DIV dir=3Dltr align=3Dleft><SPAN class=3D924080318-06082008><FONT =
face=3DArial=20
    color=3D#0000ff size=3D2>WIm</FONT></SPAN></DIV><BR>
    <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr =
align=3Dleft>
    <HR tabIndex=3D-1>
    <FONT face=3DTahoma size=3D2><B>From:</B> <A=20
    href=3D"mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A>=20
    [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang=20
    Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 =
11:30<BR><B>To:</B> <A=20
    =
href=3D"mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> <A =

    href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> =

    <BR></FONT><BR></DIV>
    <DIV></DIV>
    <DIV><FONT face=3D"Times New Roman">Hi, Ali:</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face=3D"Times New Roman">Thanks for your comments in the =
last L2VPN=20
    WG meeting, I will try to give you some reasons here.</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">Following are the&nbsp;two=20
    questions:</FONT></DIV>
    <DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face=3D"Times New Roman">1. What do you think will be =
faster=20
    convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">the learning is done in data =
plane at wire=20
    speed. Why does that prolong convergence time?<BR></FONT></DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face=3D"Times New Roman">-Faster convergence means the =
time to=20
    associate the MACs with the "exact" out port</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">will be shorter. Though learning =
is done=20
    in data plane at wire speed, the time for a frame to arrive =
</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">at the end host and the response =
time for=20
    a host to send a reverse frame&nbsp;accross the PE (on which the=20
    </FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">MACs&nbsp;is flushed) will be a =
rather=20
    long time. So the learning process is not only determined solely by=20
    </FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">the wire </FONT><FONT=20
    face=3D"Times New Roman">speed, but also on the host processing time =
and=20
    delays on all the intermediate routers.</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">For this reason, flooding may =
also last=20
    for a rather long time.</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV></DIV>
    <DIV><FONT face=3D"Times New Roman">2. You consider a simple case =
with simple=20
    backup for primary.&nbsp;In cases where primary fails and traffic=20
    gets<BR>&nbsp;distributed among multiple PE's, how do you take care =
of=20
    that?</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face=3D"Times New Roman">-I believe the most commonly=20
    deployed&nbsp;scenario is&nbsp;one backup PE for a primary =
PE.</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">In my opinion, for the scenario =
of=20
    multiple backup PEs protecting a primary PE, </FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">only one PE out of them will be =
elected as=20
    the working PE when the primary PE failed, </FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">rather than distribute the =
traffic over=20
    multiple backup PEs,</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">otherwise, loops may be =
introduced to the=20
    attached customer site.</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">BTW, in a similar way, for PW =
redundancy=20
    mechanism, only 1 secondary is selected as the working PW (see =
</FONT></DIV>
    <DIV><FONT face=3D"Times New =
Roman">draft-ietf-pwe3-redundancy-bit-00.txt=20
    in&nbsp;PWE3 WG)</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">Anyway, if there is any such =
requirements=20
    confirmed, I think it will be OK for us to find some solutions and=20
    </FONT></DIV>
    <DIV><FONT face=3D"Times New Roman">add them to&nbsp;the future=20
    drafts.&nbsp;</FONT></DIV>
    <DIV><FONT face=3D"Times New Roman"></FONT>&nbsp;</DIV>
    <DIV><FONT face=3D"Times New Roman">cheers,</FONT></DIV>
    <DIV><FONT=20
face=3D"Times New =
Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C8F8C5.548D181E--


From l2vpn-bounces@ietf.org  Thu Aug  7 16:06:24 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 29EE53A6C33;
	Thu,  7 Aug 2008 16:06:24 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 0D24E3A6993
	for <l2vpn@core3.amsl.com>; Thu,  7 Aug 2008 16:06:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id HIas+YVwOaZN for <l2vpn@core3.amsl.com>;
	Thu,  7 Aug 2008 16:06:22 -0700 (PDT)
Received: from audl751.usa.alcatel.com (audl751.usa.alcatel.com
	[143.209.238.164])
	by core3.amsl.com (Postfix) with ESMTP id CBEAB3A6C35
	for <l2vpn@ietf.org>; Thu,  7 Aug 2008 16:06:21 -0700 (PDT)
Received: from usdalsbhs02.ad3.ad.alcatel.com (usdalsbhs02.usa.alcatel.com
	[172.22.216.13])
	by audl751.usa.alcatel.com (ALCANET) with ESMTP id m77N6C66012218;
	Thu, 7 Aug 2008 18:06:21 -0500
Received: from USDALSMBS03.ad3.ad.alcatel.com ([172.22.216.8]) by
	usdalsbhs02.ad3.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Thu, 7 Aug 2008 18:06:14 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: draft-kompella-l2vpn-vpls-multihoming-01
Date: Thu, 7 Aug 2008 18:06:07 -0500
Message-ID: <4A5028372622294A99B8FFF6BD06EB7B04824B2E@USDALSMBS03.ad3.ad.alcatel.com>
In-reply-to: <13002.1218059652@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj4Dx9y0DnidhBQQ/+uwhwwQ94lIAAzw2Zg
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><4A5028372622294A99B8FFF6BD06EB7B0479210A@USDALSMBS03.ad3.ad.alcatel.com>
	<22989.1217877649@bhupesh-f8.jnpr.net>
	<4A5028372622294A99B8FFF6BD06EB7B048244DF@USDALSMBS03.ad3.ad.alcatel.com>
	<19846.1217896170@bhupesh-f8.jnpr.net>
	<4A5028372622294A99B8FFF6BD06EB7B048246F6@USDALSMBS03.ad3.ad.alcatel.com>
	<13002.1218059652@bhupesh-f8.jnpr.net>
From: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 07 Aug 2008 23:06:14.0420 (UTC)
	FILETIME=[30BF4D40:01C8F8E2]
X-Scanned-By: MIMEDefang 2.51 on 143.209.238.34
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org


Hi Bhupesh,

I think there is no way to guarantee that make before break for PWs with
MAC learning + MAC Flushing on top is hitless - i.e. synchronization of
different processes at the 2 PEs (label installation + switching between
old and new PWs, MAC Flush, MAC learning) depends on network condition,
PE load, traffic profile.=20

So I think there is a good chance that the scope of failure does extend
to non-affected CEs even with the make before break scheme.=20
There might be a need to look for an alternative to tearing down the
VPLS infrastructure every time something happens with one site.=20

But of course we should hear from people running networks on this one
before spending additional cycles...

Thanks for the detailed answers.
Florin

> -----Original Message-----
> From: Bhupesh Kothari [mailto:bhupesh@juniper.net]
> Sent: Wednesday, August 06, 2008 2:54 PM
> To: BALUS Florin
> Cc: l2vpn@ietf.org
> Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
>=20
> Hi Florin,
>=20
> BALUS Florin <Florin.Balus@alcatel-lucent.com> wrote:
> >
> > Hi Bhupesh,
> > Thanks for the answers.
> >
> > From what you explain below the failure of the CE1-PE1 access link
> > affects the PW infrastructure and implicitly the communication
> between
> > all the other CEs connected to the same VPLS on PE1. I do not think
> this
> > is a good approach as it extends the scope of the failure.
>=20
> The scope of the failure is not extended to other CEs.  See below for
> more details.
>=20
>=20
> >
> > More in-line for technicalities...
> >
> > > -----Original Message-----
> > > From: Bhupesh Kothari [mailto:bhupesh@juniper.net]
> > > Sent: Monday, August 04, 2008 5:30 PM
> > > To: BALUS Florin
> > > Cc: l2vpn@ietf.org
> > > Subject: Re: draft-kompella-l2vpn-vpls-multihoming-01
> > >
> > > Hi Florin,
> > >
> > > > > >
> > > > > > -          I think the case of a single homed CE mixed with
> the
> > > dual-
> > > > > homed ones seems even more problematic: i.e. the
> > > > > > single-homed CE could be left hanging, isolated from the
> other
> > > VPLS
> > > > > sites.
> > > > > >
> > > > >
> > > > > A single homed customer site is expected to be assigned a
> > different
> > > VE
> > > > > ID than a multi-homed site on the same PE.
> > > > >
> > > >
> > > > [[FB]] Then it means that for each "active/preferred" PE and for
> > > > each VPLS instance there will be at least two set of PWs: one
for
> > > > the VE ID assigned for the Single-Homed sites, another one for
> the
> > > > dual-homed CE1, even if they are part of the same customer VPLS
> > > > service.  Otherwise the failure of CE1-PE1 link will bring down
> all
> > > > the PWs to that VPLS switching instance on PE1, isolating all
the
> > > > single-homed CEs.
> > >
> > >
> > > No, there is only one PW between a pair of PEs and so only one PW
> is
> > > created even though two customer sites are connected to PE1.  When
> the
> > > CE1-PE1 link goes down and if the PW goes down (depends on the VE
> ID),
> > > then another PW will be reestablished, for the single homed site
as
> it
> > > has a unique VE ID in the VPLS instance, and hence, the site will
> > > never be isolated from the VPLS domain.
> >
> > [[FB]] From what you are saying here it looks like the dual-homed
CE1
> > and the single-homed CEs on PE1 share initially the same PW. Then if
> > CE1-PE1 link fails the initial PW will be torn down and a new PW
will
> be
> > re-established towards PE2 in order to reach CE1.
>=20
> Correct.
>=20
>=20
> > Also another PW to the site corresponding to the single-homed CEs on
> > PE1 will be reestablished/activated.
> >
> > This results in a service interruption for single-homed sites due to
> > failure of CE1-PE1 access link. I think this is not acceptable as it
> > increases the scope of the failure unnecessarily.
>=20
>=20
> Your understanding is correct about the PW activation, but this does
> not necessarily mean traffic disruption for single-homed site.  See
> below for details.
>=20
>=20
> >
> > >The remote peers already know
> > > the label to use to reach that single homed site.
> > >
> > [[FB]] From what you are saying the service labels/PWs assigned for
> the
> > dual-homed site/VE ID are used for all the other sites/VE IDs in the
> > VPLS on PE1?
>=20
> Yes.
>=20
> > In other words the remote PE peers do not instantiate in the data
> > path the service labels for the single-homed site/VE ID from PE1?
> >
>=20
> Yes.
>=20
>=20
> > What if there are multiple dual-homed sites on the same PEs, in the
> same
> > VPLS? Which VE ID's PW is chosen and based on what? (I am talking
> about
> > the case of different sites/VE IDs in the same VPLS)
>=20
> Irrespective of whether multiple sites are all multi-homed or only
> some are multi-homed and others single-homed, ONLY one site is
> "selected" for PW establishment that will carry traffic.  The site
> with the lowest site ID is selected for this purpose.  So, if the
> multi-homed site ID is 1 and single-homed site ID is 2, just for the
> PW establishment, ID 1 is picked by all remote PEs.
>=20
> Its an implementation detail, but here is a general explanation of why
> there shouldn't be any traffic loss.  For clarity, consider:
>=20
>=20
>               CE4 -------    ...............
>                          \  .               .    ___ CE2
>                       ___ PE1                .  /
>                      /    :                  PE3
>                   __/    :       Service      :
>               CE1 __     :       Provider    PE4
>                     \     :                   : \___ CE3
>                      \___ PE2                .
>                             .               .
>                              ...............
>=20
>=20
> PE1 has two sites, one multi-homed and other single-homed.  Site IDs
> are 1 (for CE1) and 4 (for CE4).
>=20
> On PE2, there is just one site, with ID 1 for CE1 (multi-homed).
>=20
> On PE3, there is just one site, with ID 2 for CE2
>=20
> For simplicity, lets just consider one remote PE, PE3, for
> explanation.  Also assume that PE1 has higher preference than PE2.
>=20
> PE3 will establish just one PW to PE1.  For this PW, PE1 programs its
> FIB with labels it advertises for site ID 1.  For this PW, PE3 selects
> site ID 1 from the list of sites on PE1, and so the labels advertised
> by PE1 for site ID 1 are used.  In other words, there is one PW
> between PE1 and PE3, and just for the PW establishment, site 1 and 2
> are picked on PE1 and PE3 ends.  PE3 will send traffic for both CE1
> and CE4 over this PW.
>=20
> Note that PE3 also has NLRIs for site ID 4 from PE1.  PE3 simply keeps
> those in its control plane state and only programs its FIB with labels
> for site ID 1.
>=20
> When a failure occurs, all PEs should do "make-before-break" as the
> labels are already known.
>=20
> If CE1-PE1 link goes, PE1 has to advertise the NLRI for site 1 with D
> bit set.  Before it advertises this, it programs its FIB with labels
> it advertises for site ID 4 (for CE4).  It also does NOT tear down
> it's end of the old PW (which used site 1 labels).  When PE3 learns
> about the D bit for site 1, it knows that there is additional site,
> site 4, on PE1, and so it should not simply tear down the old PW.  It
> should program the labels that it received from site 4, then it should
> switch traffic to the new PW and only then it should tear down it's
> end of the old PW.  The question of when PE1 tears down it's end of
> the old PW is again an implementationd detail, but it could be dumb
> timer based or better heuristic based.
>=20
> Also note that the path selection that PE3 does on receiving the D bit
> for site 1 from PE1 is independent of what is explained above.  The
> path selection will elect PE2 as the forwarder for site 1 and so a new
> PW to PE2 will be set up.
>=20
>=20
> > Thanks,
> > Florin
>=20
> Hope it helps.
>=20
> Thanks,
> Bhupesh



From l2vpn-bounces@ietf.org  Fri Aug  8 02:08:02 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id B71673A6B0A;
	Fri,  8 Aug 2008 02:08:02 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id C2AB63A693B
	for <l2vpn@core3.amsl.com>; Fri,  8 Aug 2008 02:08:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.19
X-Spam-Level: **
X-Spam-Status: No, score=2.19 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, FH_RELAY_NODNS=1.451, HELO_MISMATCH_COM=0.553, 
	HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001, J_CHICKENPOX_31=0.6,
	J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6, RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id HaL4rho8QX8G for <l2vpn@core3.amsl.com>;
	Fri,  8 Aug 2008 02:07:50 -0700 (PDT)
Received: from szxga01-in.huawei.com (unknown [119.145.14.64])
	by core3.amsl.com (Postfix) with ESMTP id 76C7F3A696E
	for <l2vpn@ietf.org>; Fri,  8 Aug 2008 02:07:49 -0700 (PDT)
Received: from huawei.com (szxga01-in [172.24.2.3])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K590059FYLCV7@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Fri, 08 Aug 2008 17:05:37 +0800 (CST)
Received: from huawei.com ([172.24.1.24])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5900BYUYLCQU@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Fri, 08 Aug 2008 17:05:36 +0800 (CST)
Received: from j59929 ([10.70.77.69])
	by szxml04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K59004UIYLC4K@szxml04-in.huawei.com> for
	l2vpn@ietf.org; Fri, 08 Aug 2008 17:05:36 +0800 (CST)
Date: Fri, 08 Aug 2008 17:05:33 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE:
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <007301c8f935$ea290980$454d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_tuc0TyZ5PqzB75CZtFEBOw)"
X-Priority: 3
X-MSMail-priority: Normal
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
	<007d01c8f85f$99578350$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C877@FRVELSMBS22.ad2.ad.alcatel.com>
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_tuc0TyZ5PqzB75CZtFEBOw)
Content-type: text/plain; charset=ISO-8859-1
Content-transfer-encoding: 7BIT

Hi,Wim,Please see my comments in line.
  ----- Original Message ----- 
  From: HENDERICKX Wim 
  To: Jiang Yuan-long 
  Cc: l2vpn@ietf.org 
  Sent: Friday, August 08, 2008 3:39 AM
  Subject: RE: RE:


  Yuanlong, see in-line



------------------------------------------------------------------------------
  From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
  Sent: donderdag 7 augustus 2008 9:31
  To: HENDERICKX Wim
  Cc: l2vpn@ietf.org
  Subject: Re: RE:


  Wim:

  In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group, 
  such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their 
  respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
  Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective 
  PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes 
  working PW).
  The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other 
  configurations need to be changed, can you give me some more clues?

  WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.

  Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails, 
  how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems. 

  RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except 
  those associated with specific PWs. So what you mean may be that all the existing implementations only provide 
  flushing "all" MACs blindly in a VPLS instance, is that true? 

  WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
  Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs.

  As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
  impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
  and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
  service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.

  WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
  Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend 
  on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed 
  to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as 
  my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms. 


  Best regards,

  Yuanlong
    ----- Original Message ----- 
    From: HENDERICKX Wim 
    To: Jiang Yuan-long 
    Cc: l2vpn@ietf.org 
    Sent: Thursday, August 07, 2008 1:24 PM
    Subject: RE: RE:


    Yuanlong,

    My point is that how are you going to define the behavior on each PE when the topology is adapted. I hope you are not going to change the configuration on each PE in the VPLS each time.

    On top, by doing selective MAC flush you need to go one by one to the items and hence the processing to withdraw the MAC(s) will be longer. When you use the existing mechanisms you can flush without the extra lookups and hence the performance will be faster. FLushing a MAC will result in re-learning which means you do not have traffic impact, just a small flooding until the MAC address is re-learned.

    Cheers,
    Wim


----------------------------------------------------------------------------
    From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
    Sent: donderdag 7 augustus 2008 6:00
    To: HENDERICKX Wim
    Cc: l2vpn@ietf.org
    Subject: Re: RE:



    Wim:

    Thanks for you comment. My draft is not a standlone protocol, it can be deployed with other protocol suites.
    For what I know, "PW redundancy" is being standardized in PWE3 WG, which can synchronize the behavior of
    the PEs and PWs in PW switching. So it is not such a difficult task to do PW switching, and do MAC switching
    subsequently. 

    The MAC withdrawal process defined in RFC 4762 is not so simple as you think, it maybe not just "flush all MACs" 
    blindly, but can selectively flush MACs, which computation comlexity is on the same level as my draft.
      ----- Original Message ----- 
      From: HENDERICKX Wim 
      To: Jiang Yuan-long ; sajassi@cisco.com 
      Cc: l2vpn@ietf.org 
      Sent: Thursday, August 07, 2008 4:18 AM
      Subject: RE:


      Yuanlong,

      The difficulty with your draft is that it requires a PE to do different things depending in which situation the PE is in, and this makes it very cumbersome in case changes happen to topologies or configurations.

      My view is that we need a simple MAC flush which we have today in LDP.

      Cheers,
      WIm



--------------------------------------------------------------------------
      From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
      Sent: woensdag 6 augustus 2008 11:30
      To: sajassi@cisco.com
      Cc: l2vpn@ietf.org
      Subject: 


      Hi, Ali:

      Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
      Following are the two questions:

      1. What do you think will be faster convergence? When you flush MAC addresses, 
      the learning is done in data plane at wire speed. Why does that prolong convergence time?


      -Faster convergence means the time to associate the MACs with the "exact" out port
      will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
      at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
      MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
      the wire speed, but also on the host processing time and delays on all the intermediate routers.
      For this reason, flooding may also last for a rather long time.


      2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
       distributed among multiple PE's, how do you take care of that?

      -I believe the most commonly deployed scenario is one backup PE for a primary PE.
      In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
      only one PE out of them will be elected as the working PE when the primary PE failed, 
      rather than distribute the traffic over multiple backup PEs,
      otherwise, loops may be introduced to the attached customer site.
      BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
      draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
      Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
      add them to the future drafts. 

      cheers,
      Yuanlong

--Boundary_(ID_tuc0TyZ5PqzB75CZtFEBOw)
Content-type: text/html; charset=ISO-8859-1
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=&#23435;&#20307; size=2>Hi&#65292;Wim&#65292;Please see my comments in line.</FONT></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Friday, August 08, 2008 3:39 AM</DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE: RE:</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT><BR></DIV>
  <DIV dir=ltr align=left><FONT face=Arial>
  <DIV><FONT face=SimSun><FONT color=#0000ff><FONT size=2>Yuanlong<SPAN 
  class=289373519-07082008>, see 
  in-line</SPAN></FONT></FONT></FONT></DIV></FONT></DIV><BR>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
  9:31<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
  RE:<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>Wim:</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>In VPLS with dual-homed topology, we 
  must&nbsp;pre-configure the working PE and backup PE to form a protection 
  group,&nbsp;</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>such as </FONT><FONT face=&#23435;&#20307; size=2>described in 
  ICCP (defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT face=&#23435;&#20307; 
  size=2>and pre-configure working PWs with their </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>respective backup PWs, </FONT><FONT face=&#23435;&#20307; 
  size=2>as </FONT><FONT face=&#23435;&#20307; size=2>described in "PW 
  redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>Anyway, these configurations are prerequesite to 
  provide sevice resiliency for VPLS, and the roles of the respective 
  </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>PE or PW will </FONT><FONT face=&#23435;&#20307; size=2>change 
  automatically after PE or PW switching (ie. backup PE becomes working PE, 
  backup PW becomes </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>working PW).</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>The mechanism in&nbsp;my draft&nbsp;is based on such 
  a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure what other 
  </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>configurations need to </FONT><FONT face=&#23435;&#20307; 
  size=2>be changed, can you give me some more clues?</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=SimSun size=2><SPAN class=289373519-07082008>WH&gt; We don;t 
  need any of this for PW resilience in VPLS according to me as PW status 
  together with the existing MAC flushes do the job for these 
  topologies.</SPAN></FONT></DIV>
  <DIV><FONT face=SimSun size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; color=#800000 size=2>Jiang: But how to set 
  the&nbsp;exact&nbsp;PW status?&nbsp;When a CE is dual-homed to PEs by attached 
  circuits, and if the working AC fails, </FONT></DIV>
  <DIV><FONT size=2><FONT face=&#23435;&#20307;><FONT color=#800000><FONT>how do the PEs 
  </FONT><FONT>coordinate their PWs?&nbsp;IMHO, RFC 4762&nbsp;is not perfect for 
  these problems.&nbsp;</FONT></FONT></FONT></FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>RFC 4762 already provides some selective flushing 
  mechansms, such as flushing some specified MACs or all MACs except 
  </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>those associated with </FONT><FONT face=&#23435;&#20307; 
  size=2>specific PWs. So what</FONT><FONT face=&#23435;&#20307; size=2> you mean may be that 
  all the existing implementations only&nbsp;provide </FONT></DIV>
  <DIV><FONT size=2><FONT face=&#23435;&#20307;>flushing </FONT><FONT face=&#23435;&#20307;>"all" MACs 
  blindly in a VPLS instance, </FONT><FONT face=&#23435;&#20307;>is that true?<SPAN 
  class=289373519-07082008><FONT face=Arial 
  color=#0000ff>&nbsp;</FONT></SPAN></FONT></FONT></DIV>
  <DIV><FONT size=2><FONT face=&#23435;&#20307;><SPAN 
  class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
  <DIV><FONT size=2><FONT face=&#23435;&#20307;><SPAN class=289373519-07082008>WH&gt; There 
  are certain cases where we need to extend them like for PBB where we have 
  C-MAC ties to B-MACs.</SPAN></FONT></FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; color=#800000 size=2><SPAN class=289373519-07082008>Jiang: 
  As far as I know, nowadays even some enterprise-oriented routers can 
  selectively flush MACs.</SPAN></FONT></DIV>
  <DIV><FONT size=2><FONT face=&#23435;&#20307;><SPAN 
  class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>As our anlysis and other sources(such as 
  draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
  </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>impacts all sites in a VPLS instance, even their 
  topology is not </FONT><FONT face=&#23435;&#20307; size=2>affected, the MACs must be 
  re-leaned may be numerous,</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>and the number of frames flooded may be not so 
  small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
  </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>service,</FONT><FONT face=&#23435;&#20307; size=2>and&nbsp;for 
  this&nbsp;reason&nbsp;</FONT><FONT face=&#23435;&#20307; size=2>we proposed some optimized 
  MAC operations such as MAC </FONT><FONT face=&#23435;&#20307; size=2>Switching and MAC 
  notification.</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><SPAN class=289373519-07082008><FONT face=SimSun size=2>WH&gt; The 
  optimizations which are suggested in the mentioned drafts are very simple, 
  whilst your proposal is fairly complex and depends on the topology you propose 
  but might not be efficient in case we change the service 
  topology.</FONT></SPAN></DIV><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=SimSun size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=&#23435;&#20307; size=2></FONT></BLOCKQUOTE>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV><FONT face=&#23435;&#20307; color=#800000 size=2>Jiang: The draft mentioned above also 
  do selective flushing, not blind flushing. And for what I know, this very 
  draft may depend </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; color=#800000 size=2>on the hub-and-spoke network topology 
  while our draft seems to be capable of a broader application scenario such 
  as&nbsp;a CE&nbsp;dual-homed </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; color=#800000 size=2>to PEs by ACs.&nbsp;The formats and 
  contents of the LDP messages are&nbsp;similar in all these drafts or RFC for 
  selective flushing, and as </FONT></DIV>
  <DIV><FONT size=2><FONT face=&#23435;&#20307;><FONT color=#800000><FONT>my analysis goes, 
  complexity&nbsp;(times of lookup and write)</FONT><FONT>&nbsp;in dataplane may 
  be </FONT><FONT>commensurate for all these mechanisms</FONT><FONT>. 
  </FONT></FONT></FONT></FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; 
  size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><FONT 
  face=&#23435;&#20307; size=2></FONT><FONT face=&#23435;&#20307; size=2></FONT><BR><FONT face=&#23435;&#20307; 
  size=2>Best regards,</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>Yuanlong</FONT></DIV>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
    <DIV 
    style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> <A 
    title=wim.henderickx@alcatel-lucent.be 
    href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
    href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, August 07, 2008 1:24 
    PM</DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE: RE:</DIV>
    <DIV><FONT face=&#23435;&#20307; size=2></FONT><BR></DIV>
    <DIV dir=ltr align=left><SPAN class=770051905-07082008>
    <DIV dir=ltr align=left><FONT face=Arial><FONT color=#0000ff><FONT 
    size=2>Yuanlong<SPAN 
    class=924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><SPAN 
    class=770051905-07082008><FONT face=Arial color=#0000ff size=2>My point is 
    that how are you going to define the behavior on each PE when the topology 
    is adapted. I hope you are not going to change the configuration on each PE 
    in the VPLS each time.</FONT></SPAN></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=770051905-07082008><FONT face=Arial 
    color=#0000ff size=2>On top, by doing selective MAC flush you need to go one 
    by one to the items and hence the processing to withdraw the MAC(s) will be 
    longer. When you use the existing mechanisms you can flush without the extra 
    lookups and hence the performance will be faster. FLushing a MAC will result 
    in re-learning which means you do not have traffic impact, just a small 
    flooding until the MAC address is re-learned.</FONT></SPAN></DIV>
    <DIV>&nbsp;</DIV>
    <DIV><SPAN class=770051905-07082008></SPAN><FONT face=Arial><FONT 
    color=#0000ff><FONT size=2>C<SPAN 
    class=770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=+0><FONT color=#0000ff><FONT size=2><SPAN 
    class=770051905-07082008></SPAN></FONT></FONT></FONT><SPAN 
    class=770051905-07082008></SPAN><FONT face=Arial><FONT color=#0000ff><FONT 
    size=2>W<SPAN 
    class=770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
    <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
    <HR tabIndex=-1>
    <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
    [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
    6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
    l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
    <DIV></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=&#23435;&#20307; size=2>Wim:</FONT></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=&#23435;&#20307; size=2>Thanks for you comment. My draft is not a 
    standlone protocol, it can be deployed with other protocol 
    suites.</FONT></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2>For what I know, "PW redundancy" is being 
    standardized in PWE3 WG, which can synchronize the behavior of</FONT></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2>the PEs and PWs in PW switching. So it is 
    not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC 
    switching</FONT></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2>subsequently. </FONT></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=&#23435;&#20307; size=2>The MAC withdrawal process defined in RFC 4762 is 
    not so simple as you think,&nbsp;it maybe not just<FONT size=2>&nbsp;"flush 
    all MACs" </FONT></FONT></DIV>
    <DIV><FONT face=&#23435;&#20307; size=2><FONT size=2>blindly, but can selectively flush 
    MACs, which computation comlexity is on the same level as my 
    draft</FONT><FONT size=2>.</FONT></DIV></FONT>
    <BLOCKQUOTE dir=ltr 
    style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
      <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
      <DIV 
      style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> 
      <A title=wim.henderickx@alcatel-lucent.be 
      href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
      <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
      href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A 
      title=sajassi@cisco.com 
      href="mailto:sajassi@cisco.com">sajassi@cisco.com</A> </DIV>
      <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
      <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Thursday, August 07, 2008 4:18 
      AM</DIV>
      <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE:</DIV>
      <DIV><BR></DIV>
      <DIV dir=ltr align=left>Yuanlong<SPAN 
      class=924080318-06082008>,</SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>The difficulty with your draft is that it requires a 
      PE to do different things depending in which situation the PE is in, and 
      this makes it very cumbersome in case changes happen to topologies or 
      configurations.</FONT></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>My view is that we need a simple MAC flush which we 
      have today in LDP.</FONT></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>Cheers,</FONT></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>WIm</FONT></SPAN></DIV><BR>
      <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
      <HR tabIndex=-1>
      <FONT face=Tahoma size=2><B>From:</B> <A 
      href="mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A> 
      [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang 
      Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> <A 
      href="mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> <A 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> 
      <BR></FONT><BR></DIV>
      <DIV></DIV>
      <DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">Thanks for your comments in the last 
      L2VPN WG meeting, I will try to give you some reasons here.</FONT></DIV>
      <DIV><FONT face="Times New Roman">Following are the&nbsp;two 
      questions:</FONT></DIV>
      <DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">1. What do you think will be faster 
      convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
      <DIV><FONT face="Times New Roman">the learning is done in data plane at 
      wire speed. Why does that prolong convergence time?<BR></FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">-Faster convergence means the time to 
      associate the MACs with the "exact" out port</FONT></DIV>
      <DIV><FONT face="Times New Roman">will be shorter. Though learning is done 
      in data plane at wire speed, the time for a frame to arrive </FONT></DIV>
      <DIV><FONT face="Times New Roman">at the end host and the response time 
      for a host to send a reverse frame&nbsp;accross the PE (on which the 
      </FONT></DIV>
      <DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather 
      long time. So the learning process is not only determined solely by 
      </FONT></DIV>
      <DIV><FONT face="Times New Roman">the wire </FONT><FONT 
      face="Times New Roman">speed, but also on the host processing time and 
      delays on all the intermediate routers.</FONT></DIV>
      <DIV><FONT face="Times New Roman">For this reason, flooding may also last 
      for a rather long time.</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
      <DIV><FONT face="Times New Roman">2. You consider a simple case with 
      simple backup for primary.&nbsp;In cases where primary fails and traffic 
      gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
      that?</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">-I believe the most commonly 
      deployed&nbsp;scenario is&nbsp;one backup PE for a primary 
PE.</FONT></DIV>
      <DIV><FONT face="Times New Roman">In my opinion, for the scenario of 
      multiple backup PEs protecting a primary PE, </FONT></DIV>
      <DIV><FONT face="Times New Roman">only one PE out of them will be elected 
      as the working PE when the primary PE failed, </FONT></DIV>
      <DIV><FONT face="Times New Roman">rather than distribute the traffic over 
      multiple backup PEs,</FONT></DIV>
      <DIV><FONT face="Times New Roman">otherwise, loops may be introduced to 
      the attached customer site.</FONT></DIV>
      <DIV><FONT face="Times New Roman">BTW, in a similar way, for PW redundancy 
      mechanism, only 1 secondary is selected as the working PW (see 
      </FONT></DIV>
      <DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
      in&nbsp;PWE3 WG)</FONT></DIV>
      <DIV><FONT face="Times New Roman">Anyway, if there is any such 
      requirements confirmed, I think it will be OK for us to find some 
      solutions and </FONT></DIV>
      <DIV><FONT face="Times New Roman">add them to&nbsp;the future 
      drafts.&nbsp;</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
      <DIV><FONT 
  face="Times New Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

--Boundary_(ID_tuc0TyZ5PqzB75CZtFEBOw)--


From l2vpn-bounces@ietf.org  Sat Aug  9 23:58:11 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 141413A6BE0;
	Sat,  9 Aug 2008 23:58:11 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 46C573A6BE0
	for <l2vpn@core3.amsl.com>; Sat,  9 Aug 2008 23:58:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.896
X-Spam-Level: **
X-Spam-Status: No, score=2.896 tagged_above=-999 required=5 tests=[AWL=-0.139, 
	BAYES_50=0.001, HELO_EQ_FR=0.35, HTML_FONT_FACE_BAD=0.884,
	HTML_MESSAGE=0.001, J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6,
	J_CHICKENPOX_74=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id UiJj-VskMiKc for <l2vpn@core3.amsl.com>;
	Sat,  9 Aug 2008 23:58:08 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [64.208.49.27])
	by core3.amsl.com (Postfix) with ESMTP id E886C3A6BDD
	for <l2vpn@ietf.org>; Sat,  9 Aug 2008 23:58:07 -0700 (PDT)
Received: from FRVELSBHS02.ad2.ad.alcatel.com (frvelsbhs02.ad2.ad.alcatel.com
	[155.132.6.74])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7A6vfR7027153; 
	Sun, 10 Aug 2008 08:57:43 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS02.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Sun, 10 Aug 2008 08:57:41 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8FAB6.6171B596"
Subject: RE: RE:
Date: Sun, 10 Aug 2008 08:57:44 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302CB8F@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <007301c8f935$ea290980$454d460a@china.huawei.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: RE:
Thread-Index: Acj5Nfs80Rg9H1fiQkW9rVLMvQVW+QANvj+A
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
	<007d01c8f85f$99578350$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C877@FRVELSMBS22.ad2.ad.alcatel.com>
	<007301c8f935$ea290980$454d460a@china.huawei.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Jiang Yuan-long" <yljiang@huawei.com>
X-OriginalArrivalTime: 10 Aug 2008 06:57:41.0486 (UTC)
	FILETIME=[61FA88E0:01C8FAB6]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8FAB6.6171B596
Content-Type: text/plain;
	charset="iso-2022-jp"
Content-Transfer-Encoding: 7bit

Yuanlong, more in-line


________________________________

From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
Sent: vrijdag 8 augustus 2008 11:06
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: RE:


Hi$B!$(JWim$B!$(JPlease see my comments in line.

	----- Original Message ----- 
	From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
	To: Jiang Yuan-long <mailto:yljiang@huawei.com>  
	Cc: l2vpn@ietf.org 
	Sent: Friday, August 08, 2008 3:39 AM
	Subject: RE: RE:
	
	
	
	Yuanlong, see in-line

________________________________

	From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
	Sent: donderdag 7 augustus 2008 9:31
	To: HENDERICKX Wim
	Cc: l2vpn@ietf.org
	Subject: Re: RE:
	
	
	Wim:
	 
	In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group, 
	such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their 
	respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
	Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective 
	PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes 
	working PW).
	The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other 
	configurations need to be changed, can you give me some more clues?
	 
	WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.
	 
	Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails, 
	how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems.  
	 
	Wh> The way it works is you use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt and if the PW status changes you should perform a MAC flush in the full mesh VPLS domain.
	 
	RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except 
	those associated with specific PWs. So what you mean may be that all the existing implementations only provide 
	flushing "all" MACs blindly in a VPLS instance, is that true? 
	 
	WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
	Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs. 
	 
	WH> Isn't the main reason for this the fact that they can't handle large floods of traffic?  
	 
	As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
	impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
	and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
	service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.
	 
	WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
	

	Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend 
	on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed 
	to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as 
	my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms.  
	 
	WH> Right but the main difference is that you only need to indicate the PE or PW for which the MAC flush is performed. In your proposal you need to indicate the full MAC address list which will be very processing intensive if you have a large amount of MAC addresses in the VPLS service. The message will get very big and you will have to sent lots of them due to fragmentation implications. 
	Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe the wrong wording) flush mechanism. The indication will be fast and the re-learning will happen instantly after the MAC is removed from the MAC FIB.
	 
	You proposal will delay the trigger when the amount of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to find he proper MAC and replace it which means delay the removal of the FIB.
	 
	
	Best regards,
	 
	Yuanlong

		----- Original Message ----- 
		From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
		To: Jiang Yuan-long <mailto:yljiang@huawei.com>  
		Cc: l2vpn@ietf.org 
		Sent: Thursday, August 07, 2008 1:24 PM
		Subject: RE: RE:
		
		
		
		Yuanlong,
		 
		My point is that how are you going to define the behavior on each PE when the topology is adapted. I hope you are not going to change the configuration on each PE in the VPLS each time.
		 
		On top, by doing selective MAC flush you need to go one by one to the items and hence the processing to withdraw the MAC(s) will be longer. When you use the existing mechanisms you can flush without the extra lookups and hence the performance will be faster. FLushing a MAC will result in re-learning which means you do not have traffic impact, just a small flooding until the MAC address is re-learned.
		 
		Cheers,
		Wim
		
________________________________

		From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
		Sent: donderdag 7 augustus 2008 6:00
		To: HENDERICKX Wim
		Cc: l2vpn@ietf.org
		Subject: Re: RE:
		
		
		 
		Wim:
		 
		Thanks for you comment. My draft is not a standlone protocol, it can be deployed with other protocol suites.
		For what I know, "PW redundancy" is being standardized in PWE3 WG, which can synchronize the behavior of
		the PEs and PWs in PW switching. So it is not such a difficult task to do PW switching, and do MAC switching
		subsequently. 
		 
		The MAC withdrawal process defined in RFC 4762 is not so simple as you think, it maybe not just "flush all MACs" 
		blindly, but can selectively flush MACs, which computation comlexity is on the same level as my draft.

			----- Original Message ----- 
			From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
			To: Jiang Yuan-long <mailto:yljiang@huawei.com>  ; sajassi@cisco.com 
			Cc: l2vpn@ietf.org 
			Sent: Thursday, August 07, 2008 4:18 AM
			Subject: RE:

			Yuanlong,
			 
			The difficulty with your draft is that it requires a PE to do different things depending in which situation the PE is in, and this makes it very cumbersome in case changes happen to topologies or configurations.
			 
			My view is that we need a simple MAC flush which we have today in LDP.
			 
			Cheers,
			WIm

________________________________

			From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
			Sent: woensdag 6 augustus 2008 11:30
			To: sajassi@cisco.com
			Cc: l2vpn@ietf.org
			Subject: 
			
			
			Hi, Ali:
			 
			Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
			Following are the two questions:
			 
			1. What do you think will be faster convergence? When you flush MAC addresses, 
			the learning is done in data plane at wire speed. Why does that prolong convergence time?
			
			 
			-Faster convergence means the time to associate the MACs with the "exact" out port
			will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
			at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
			MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
			the wire speed, but also on the host processing time and delays on all the intermediate routers.
			For this reason, flooding may also last for a rather long time.
			 
			 
			2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
			 distributed among multiple PE's, how do you take care of that?
			 
			-I believe the most commonly deployed scenario is one backup PE for a primary PE.
			In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
			only one PE out of them will be elected as the working PE when the primary PE failed, 
			rather than distribute the traffic over multiple backup PEs,
			otherwise, loops may be introduced to the attached customer site.
			BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
			draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
			Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
			add them to the future drafts. 
			 
			cheers,
			Yuanlong
			


------_=_NextPart_001_01C8FAB6.6171B596
Content-Type: text/html;
	charset="iso-2022-jp"
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-2022-jp">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV dir=ltr align=left><FONT face=SimSun><FONT color=#0000ff><FONT 
size=2>Yuanlong<SPAN class=289373519-07082008>,&nbsp;<SPAN 
class=012333915-08082008>more</SPAN> 
in-line</SPAN></FONT></FONT></FONT></DIV><FONT face=Arial color=#0000ff 
size=2></FONT><BR>
<DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
<HR tabIndex=-1>
<FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
[mailto:yljiang@huawei.com] <BR><B>Sent:</B> vrijdag 8 augustus 2008 
11:06<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV><FONT face=$BAWBN(B size=2>Hi$B!$(BWim$B!$(BPlease see my comments in line.</FONT></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Friday, August 08, 2008 3:39 AM</DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT><BR></DIV>
  <DIV dir=ltr align=left><FONT face=Arial>
  <DIV><FONT face=SimSun><FONT color=#0000ff><FONT size=2>Yuanlong<SPAN 
  class=289373519-07082008>, see 
  in-line</SPAN></FONT></FONT></FONT></DIV></FONT></DIV><BR>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
  9:31<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
  RE:<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B size=2>In VPLS with dual-homed topology, we 
  must&nbsp;pre-configure the working PE and backup PE to form a protection 
  group,&nbsp;</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>such as </FONT><FONT face=$BAWBN(B size=2>described in 
  ICCP (defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT face=$BAWBN(B 
  size=2>and pre-configure working PWs with their </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>respective backup PWs, </FONT><FONT face=$BAWBN(B 
  size=2>as </FONT><FONT face=$BAWBN(B size=2>described in "PW 
  redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>Anyway, these configurations are prerequesite to 
  provide sevice resiliency for VPLS, and the roles of the respective 
  </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>PE or PW will </FONT><FONT face=$BAWBN(B size=2>change 
  automatically after PE or PW switching (ie. backup PE becomes working PE, 
  backup PW becomes </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>working PW).</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>The mechanism in&nbsp;my draft&nbsp;is based on such 
  a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure what other 
  </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>configurations need to </FONT><FONT face=$BAWBN(B 
  size=2>be changed, can you give me some more clues?</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=SimSun size=2><SPAN class=289373519-07082008>WH&gt; We don;t 
  need any of this for PW resilience in VPLS according to me as PW status 
  together with the existing MAC flushes do the job for these 
  topologies.</SPAN></FONT></DIV>
  <DIV><FONT face=SimSun size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: But how to set 
  the&nbsp;exact&nbsp;PW status?&nbsp;When a CE is dual-homed to PEs by attached 
  circuits, and if the working AC fails, </FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><FONT size=+0>how do the 
  PEs </FONT><FONT size=+0>coordinate their PWs?&nbsp;IMHO, RFC 4762&nbsp;is not 
  perfect for these problems.&nbsp;</FONT><SPAN class=012333915-08082008><FONT 
  face=Arial color=#0000ff 
size=2>&nbsp;</FONT></SPAN></FONT></FONT></FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
  class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><FONT><SPAN class=012333915-08082008><FONT 
  color=#0000ff size=3>Wh&gt; The way it works is you use PW status as defined 
  in 
  the&nbsp;<STRONG>draft</STRONG>-muley-dutta-<B>pw</B>e3-redundancy-bit-02.txt 
  and if the PW status changes you should perform a MAC flush in the full mesh 
  VPLS domain.</FONT></SPAN></FONT></FONT></FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B size=2>RFC 4762 already provides some selective flushing 
  mechansms, such as flushing some specified MACs or all MACs except 
  </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>those associated with </FONT><FONT face=$BAWBN(B 
  size=2>specific PWs. So what</FONT><FONT face=$BAWBN(B size=2> you mean may be that 
  all the existing implementations only&nbsp;provide </FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B>flushing </FONT><FONT face=$BAWBN(B>"all" MACs 
  blindly in a VPLS instance, </FONT><FONT face=$BAWBN(B>is that true?<SPAN 
  class=289373519-07082008><FONT face=Arial 
  color=#0000ff>&nbsp;</FONT></SPAN></FONT></FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN 
  class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN class=289373519-07082008>WH&gt; There 
  are certain cases where we need to extend them like for PBB where we have 
  C-MAC ties to B-MACs.</SPAN></FONT></FONT></DIV>
  <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
  size=2>Jiang: As far as I know, nowadays even some enterprise-oriented routers 
  can selectively flush MACs.<SPAN class=012333915-08082008><FONT face=Arial 
  color=#0000ff>&nbsp;</FONT></SPAN></FONT></SPAN></FONT></DIV>
  <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
  size=2><SPAN class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
  size=2><SPAN class=012333915-08082008><FONT face=Arial 
  color=#0000ff>WH&gt;&nbsp;Isn't the main reason for this the fact that they 
  can't handle large floods of 
  traffic?&nbsp;</FONT>&nbsp;</SPAN></FONT></SPAN></FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN 
  class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B size=2>As our anlysis and other sources(such as 
  draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
  </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>impacts all sites in a VPLS instance, even their 
  topology is not </FONT><FONT face=$BAWBN(B size=2>affected, the MACs must be 
  re-leaned may be numerous,</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>and the number of frames flooded may be not so 
  small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
  </FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2>service,</FONT><FONT face=$BAWBN(B size=2>and&nbsp;for 
  this&nbsp;reason&nbsp;</FONT><FONT face=$BAWBN(B size=2>we proposed some optimized 
  MAC operations such as MAC </FONT><FONT face=$BAWBN(B size=2>Switching and MAC 
  notification.</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
  <DIV><SPAN class=289373519-07082008><FONT face=SimSun size=2>WH&gt; The 
  optimizations which are suggested in the mentioned drafts are very simple, 
  whilst your proposal is fairly complex and depends on the topology you propose 
  but might not be efficient in case we change the service 
  topology.</FONT></SPAN></DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=SimSun size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=$BAWBN(B size=2></FONT></BLOCKQUOTE>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: The draft mentioned above also 
  do selective flushing, not blind flushing. And for what I know, this very 
  draft may depend </FONT></DIV>
  <DIV><FONT face=$BAWBN(B color=#800000 size=2>on the hub-and-spoke network topology 
  while our draft seems to be capable of a broader application scenario such 
  as&nbsp;a CE&nbsp;dual-homed </FONT></DIV>
  <DIV><FONT face=$BAWBN(B color=#800000 size=2>to PEs by ACs.&nbsp;The formats and 
  contents of the LDP messages are&nbsp;similar in all these drafts or RFC for 
  selective flushing, and as </FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><FONT size=+0>my analysis 
  goes, complexity&nbsp;(times of lookup and write)</FONT><FONT size=+0>&nbsp;in 
  dataplane may be </FONT><FONT size=+0>commensurate for all these 
  mechanisms</FONT><FONT size=+0>.&nbsp;</FONT><SPAN 
  class=012333915-08082008><FONT face=Arial color=#0000ff 
  size=2>&nbsp;</FONT></SPAN></FONT></FONT></FONT></DIV>
  <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
  class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
  <DIV><FONT><SPAN class=012333915-08082008><FONT color=#0000ff><FONT face=Arial 
  size=2>WH&gt; Right but the main difference is that you only need to indicate 
  the PE or PW for which the MAC flush is performed.&nbsp;In your proposal you 
  need to indicate the full MAC address list which will be very processing 
  intensive if you have a large amount of MAC addresses in the VPLS service. The 
  message will get very big and you will have to sent lots of them due to 
  fragmentation implications. </FONT></FONT></SPAN></FONT></DIV>
  <DIV><FONT><SPAN class=012333915-08082008><FONT color=#0000ff><FONT face=Arial 
  size=2><FONT face=SimSun color=#000000><FONT 
  color=#0000ff>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides</FONT> </FONT>a 
  simple (blind was maybe the wrong wording)&nbsp;flush mechanism. The 
  indication will be fast and the re-learning will happen instantly after the 
  MAC is removed from the MAC FIB.</FONT></FONT></SPAN></FONT></DIV>
  <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
  class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
  <DIV><FONT face=Arial color=#0000ff size=2><SPAN class=012333915-08082008>You 
  proposal will delay the trigger when the amount of MAC @ grows and on top the 
  receiving PE needs a fast lookup mechanism to find he proper MAC and replace 
  it which means delay the removal of the FIB.</SPAN></FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
  size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
  face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><BR><FONT face=$BAWBN(B 
  size=2>Best regards,</FONT></DIV>
  <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=$BAWBN(B size=2>Yuanlong</FONT></DIV>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
    <DIV 
    style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> <A 
    title=wim.henderickx@alcatel-lucent.be 
    href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
    href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Thursday, August 07, 2008 1:24 
    PM</DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT><BR></DIV>
    <DIV dir=ltr align=left><SPAN class=770051905-07082008>
    <DIV dir=ltr align=left><FONT face=Arial><FONT color=#0000ff><FONT 
    size=2>Yuanlong<SPAN 
    class=924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><SPAN 
    class=770051905-07082008><FONT face=Arial color=#0000ff size=2>My point is 
    that how are you going to define the behavior on each PE when the topology 
    is adapted. I hope you are not going to change the configuration on each PE 
    in the VPLS each time.</FONT></SPAN></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
    color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
    <DIV dir=ltr align=left><SPAN class=770051905-07082008><FONT face=Arial 
    color=#0000ff size=2>On top, by doing selective MAC flush you need to go one 
    by one to the items and hence the processing to withdraw the MAC(s) will be 
    longer. When you use the existing mechanisms you can flush without the extra 
    lookups and hence the performance will be faster. FLushing a MAC will result 
    in re-learning which means you do not have traffic impact, just a small 
    flooding until the MAC address is re-learned.</FONT></SPAN></DIV>
    <DIV>&nbsp;</DIV>
    <DIV><SPAN class=770051905-07082008></SPAN><FONT face=Arial><FONT 
    color=#0000ff><FONT size=2>C<SPAN 
    class=770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=+0><FONT color=#0000ff><FONT size=2><SPAN 
    class=770051905-07082008></SPAN></FONT></FONT></FONT><SPAN 
    class=770051905-07082008></SPAN><FONT face=Arial><FONT color=#0000ff><FONT 
    size=2>W<SPAN 
    class=770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
    <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
    <HR tabIndex=-1>
    <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
    [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
    6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
    l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
    <DIV></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>Thanks for you comment. My draft is not a 
    standlone protocol, it can be deployed with other protocol 
    suites.</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>For what I know, "PW redundancy" is being 
    standardized in PWE3 WG, which can synchronize the behavior of</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>the PEs and PWs in PW switching. So it is 
    not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC 
    switching</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>subsequently. </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>The MAC withdrawal process defined in RFC 4762 is 
    not so simple as you think,&nbsp;it maybe not just<FONT size=2>&nbsp;"flush 
    all MACs" </FONT></FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2><FONT size=2>blindly, but can selectively flush 
    MACs, which computation comlexity is on the same level as my 
    draft</FONT><FONT size=2>.</FONT></DIV></FONT>
    <BLOCKQUOTE dir=ltr 
    style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
      <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
      <DIV 
      style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
      <A title=wim.henderickx@alcatel-lucent.be 
      href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
      href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A 
      title=sajassi@cisco.com 
      href="mailto:sajassi@cisco.com">sajassi@cisco.com</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Thursday, August 07, 2008 4:18 
      AM</DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE:</DIV>
      <DIV><BR></DIV>
      <DIV dir=ltr align=left>Yuanlong<SPAN 
      class=924080318-06082008>,</SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>The difficulty with your draft is that it requires a 
      PE to do different things depending in which situation the PE is in, and 
      this makes it very cumbersome in case changes happen to topologies or 
      configurations.</FONT></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>My view is that we need a simple MAC flush which we 
      have today in LDP.</FONT></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>Cheers,</FONT></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2>WIm</FONT></SPAN></DIV><BR>
      <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
      <HR tabIndex=-1>
      <FONT face=Tahoma size=2><B>From:</B> <A 
      href="mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A> 
      [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang 
      Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> <A 
      href="mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> <A 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> 
      <BR></FONT><BR></DIV>
      <DIV></DIV>
      <DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">Thanks for your comments in the last 
      L2VPN WG meeting, I will try to give you some reasons here.</FONT></DIV>
      <DIV><FONT face="Times New Roman">Following are the&nbsp;two 
      questions:</FONT></DIV>
      <DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">1. What do you think will be faster 
      convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
      <DIV><FONT face="Times New Roman">the learning is done in data plane at 
      wire speed. Why does that prolong convergence time?<BR></FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">-Faster convergence means the time to 
      associate the MACs with the "exact" out port</FONT></DIV>
      <DIV><FONT face="Times New Roman">will be shorter. Though learning is done 
      in data plane at wire speed, the time for a frame to arrive </FONT></DIV>
      <DIV><FONT face="Times New Roman">at the end host and the response time 
      for a host to send a reverse frame&nbsp;accross the PE (on which the 
      </FONT></DIV>
      <DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather 
      long time. So the learning process is not only determined solely by 
      </FONT></DIV>
      <DIV><FONT face="Times New Roman">the wire </FONT><FONT 
      face="Times New Roman">speed, but also on the host processing time and 
      delays on all the intermediate routers.</FONT></DIV>
      <DIV><FONT face="Times New Roman">For this reason, flooding may also last 
      for a rather long time.</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
      <DIV><FONT face="Times New Roman">2. You consider a simple case with 
      simple backup for primary.&nbsp;In cases where primary fails and traffic 
      gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
      that?</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">-I believe the most commonly 
      deployed&nbsp;scenario is&nbsp;one backup PE for a primary 
PE.</FONT></DIV>
      <DIV><FONT face="Times New Roman">In my opinion, for the scenario of 
      multiple backup PEs protecting a primary PE, </FONT></DIV>
      <DIV><FONT face="Times New Roman">only one PE out of them will be elected 
      as the working PE when the primary PE failed, </FONT></DIV>
      <DIV><FONT face="Times New Roman">rather than distribute the traffic over 
      multiple backup PEs,</FONT></DIV>
      <DIV><FONT face="Times New Roman">otherwise, loops may be introduced to 
      the attached customer site.</FONT></DIV>
      <DIV><FONT face="Times New Roman">BTW, in a similar way, for PW redundancy 
      mechanism, only 1 secondary is selected as the working PW (see 
      </FONT></DIV>
      <DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
      in&nbsp;PWE3 WG)</FONT></DIV>
      <DIV><FONT face="Times New Roman">Anyway, if there is any such 
      requirements confirmed, I think it will be OK for us to find some 
      solutions and </FONT></DIV>
      <DIV><FONT face="Times New Roman">add them to&nbsp;the future 
      drafts.&nbsp;</FONT></DIV>
      <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
      <DIV><FONT 
  face="Times New Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C8FAB6.6171B596--


From l2vpn-bounces@ietf.org  Sun Aug 10 19:27:12 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 931953A68F0;
	Sun, 10 Aug 2008 19:27:12 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id D817D3A68CC
	for <l2vpn@core3.amsl.com>; Sun, 10 Aug 2008 19:27:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.138
X-Spam-Level: *
X-Spam-Status: No, score=1.138 tagged_above=-999 required=5 tests=[AWL=1.052, 
	BAYES_00=-2.599, HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001, 
	J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id MxRWT-KmYGFa for <l2vpn@core3.amsl.com>;
	Sun, 10 Aug 2008 19:27:08 -0700 (PDT)
Received: from szxga02-in.huawei.com (szxga02-in.huawei.com [119.145.14.65])
	by core3.amsl.com (Postfix) with ESMTP id 29D6A3A67EA
	for <l2vpn@ietf.org>; Sun, 10 Aug 2008 19:27:08 -0700 (PDT)
Received: from huawei.com (szxga02-in [172.24.2.6])
	by szxga02-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00GD2055AB@szxga02-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 10:27:06 +0800 (CST)
Received: from huawei.com ([172.24.1.18])
	by szxga02-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00DP3055HP@szxga02-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 10:27:05 +0800 (CST)
Received: from j59929 ([10.70.77.69])
	by szxml03-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5F00E1J0548L@szxml03-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 10:27:05 +0800 (CST)
Date: Mon, 11 Aug 2008 10:27:02 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE:
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <006101c8fb59$bd4a6690$454d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_vWbJgKuThrZutKTMcibMBw)"
X-Priority: 3
X-MSMail-priority: Normal
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
	<007d01c8f85f$99578350$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C877@FRVELSMBS22.ad2.ad.alcatel.com>
	<007301c8f935$ea290980$454d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302CB8F@FRVELSMBS22.ad2.ad.alcatel.com>
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_vWbJgKuThrZutKTMcibMBw)
Content-type: text/plain; charset=iso-2022-jp
Content-transfer-encoding: 7BIT

Wim, please see more in-line.

Cheers
Yuanlong
  ----- Original Message ----- 
  From: HENDERICKX Wim 
  To: Jiang Yuan-long 
  Cc: l2vpn@ietf.org 
  Sent: Sunday, August 10, 2008 2:57 PM
  Subject: RE: RE:


  Yuanlong, more in-line



------------------------------------------------------------------------------
  From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
  Sent: vrijdag 8 augustus 2008 11:06
  To: HENDERICKX Wim
  Cc: l2vpn@ietf.org
  Subject: Re: RE:


  Hi$B!$(BWim$B!$(BPlease see my comments in line.
    ----- Original Message ----- 
    From: HENDERICKX Wim 
    To: Jiang Yuan-long 
    Cc: l2vpn@ietf.org 
    Sent: Friday, August 08, 2008 3:39 AM
    Subject: RE: RE:


    Yuanlong, see in-line



----------------------------------------------------------------------------
    From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
    Sent: donderdag 7 augustus 2008 9:31
    To: HENDERICKX Wim
    Cc: l2vpn@ietf.org
    Subject: Re: RE:


    Wim:

    In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group, 
    such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their 
    respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
    Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective 
    PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes 
    working PW).
    The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other 
    configurations need to be changed, can you give me some more clues?

    WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.

    Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails, 
    how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems.  

    Wh> The way it works is you use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt and if the PW status changes you should perform a MAC flush in the full mesh VPLS domain.

    Jiang: As the above mentioned draft indicates, pre-configuration of PWs is necessary for its working. The only difference is whether 
    we should perform MAC flushing or MAC switching when we do PW switching.
    I think blindly flushing, selectively flushing and its optimizations (see Dutta's draft), and the flush-free optimizations as I proposed, 
    should all be optional operations with their different applicability and efficiency. 
    For the case of PW switching in VPLS with redundancy, flush-free optimizations is more favorable to me.

    RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except 
    those associated with specific PWs. So what you mean may be that all the existing implementations only provide 
    flushing "all" MACs blindly in a VPLS instance, is that true? 

    WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
    Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs. 

    WH> Isn't the main reason for this the fact that they can't handle large floods of traffic?  

    Jiang: Selective flushing is an optional operation, not the only choice. If turning down the selecitve MAC flushing
    will improve the performance of a router, surely you can do that.

    As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
    impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
    and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
    service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.

    WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
    Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend 
    on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed 
    to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as 
    my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms.  

    WH> Right but the main difference is that you only need to indicate the PE or PW for which the MAC flush is performed. In your proposal you need to indicate the full MAC address list which will be very processing intensive if you have a large amount of MAC addresses in the VPLS service. The message will get very big and you will have to sent lots of them due to fragmentation implications. 
    Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe the wrong wording) flush mechanism. The indication will be fast and the re-learning will happen instantly after the MAC is removed from the MAC FIB.

    You proposal will delay the trigger when the amount of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to find he proper MAC and replace it which means delay the removal of the FIB.

    Jiang: As indicated in our draft, MAC Switching message can also carry only a null-list of MACs. It's OK to change all the MACs associated with a PW 
    to another PW as egress.


    Best regards,

    Yuanlong
      ----- Original Message ----- 
      From: HENDERICKX Wim 
      To: Jiang Yuan-long 
      Cc: l2vpn@ietf.org 
      Sent: Thursday, August 07, 2008 1:24 PM
      Subject: RE: RE:


      Yuanlong,

      My point is that how are you going to define the behavior on each PE when the topology is adapted. I hope you are not going to change the configuration on each PE in the VPLS each time.

      On top, by doing selective MAC flush you need to go one by one to the items and hence the processing to withdraw the MAC(s) will be longer. When you use the existing mechanisms you can flush without the extra lookups and hence the performance will be faster. FLushing a MAC will result in re-learning which means you do not have traffic impact, just a small flooding until the MAC address is re-learned.

      Cheers,
      Wim


--------------------------------------------------------------------------
      From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
      Sent: donderdag 7 augustus 2008 6:00
      To: HENDERICKX Wim
      Cc: l2vpn@ietf.org
      Subject: Re: RE:



      Wim:

      Thanks for you comment. My draft is not a standlone protocol, it can be deployed with other protocol suites.
      For what I know, "PW redundancy" is being standardized in PWE3 WG, which can synchronize the behavior of
      the PEs and PWs in PW switching. So it is not such a difficult task to do PW switching, and do MAC switching
      subsequently. 

      The MAC withdrawal process defined in RFC 4762 is not so simple as you think, it maybe not just "flush all MACs" 
      blindly, but can selectively flush MACs, which computation comlexity is on the same level as my draft.
        ----- Original Message ----- 
        From: HENDERICKX Wim 
        To: Jiang Yuan-long ; sajassi@cisco.com 
        Cc: l2vpn@ietf.org 
        Sent: Thursday, August 07, 2008 4:18 AM
        Subject: RE:


        Yuanlong,

        The difficulty with your draft is that it requires a PE to do different things depending in which situation the PE is in, and this makes it very cumbersome in case changes happen to topologies or configurations.

        My view is that we need a simple MAC flush which we have today in LDP.

        Cheers,
        WIm



------------------------------------------------------------------------
        From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
        Sent: woensdag 6 augustus 2008 11:30
        To: sajassi@cisco.com
        Cc: l2vpn@ietf.org
        Subject: 


        Hi, Ali:

        Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
        Following are the two questions:

        1. What do you think will be faster convergence? When you flush MAC addresses, 
        the learning is done in data plane at wire speed. Why does that prolong convergence time?


        -Faster convergence means the time to associate the MACs with the "exact" out port
        will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
        at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
        MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
        the wire speed, but also on the host processing time and delays on all the intermediate routers.
        For this reason, flooding may also last for a rather long time.


        2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
         distributed among multiple PE's, how do you take care of that?

        -I believe the most commonly deployed scenario is one backup PE for a primary PE.
        In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
        only one PE out of them will be elected as the working PE when the primary PE failed, 
        rather than distribute the traffic over multiple backup PEs,
        otherwise, loops may be introduced to the attached customer site.
        BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
        draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
        Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
        add them to the future drafts. 

        cheers,
        Yuanlong

--Boundary_(ID_vWbJgKuThrZutKTMcibMBw)
Content-type: text/html; charset=iso-2022-jp
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-2022-jp">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face="Times New Roman" color=#000080>Wim, please see more 
in-line.</FONT></DIV>
<DIV><FONT face="Times New Roman" color=#000080></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman" color=#000080>Cheers</FONT></DIV>
<DIV><FONT face="Times New Roman" color=#000080>Yuanlong</FONT></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Sunday, August 10, 2008 2:57 PM</DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
  <DIV><BR></DIV>
  <DIV dir=ltr align=left><FONT face=SimSun><FONT color=#0000ff><FONT 
  size=2>Yuanlong<SPAN class=289373519-07082008>,&nbsp;<SPAN 
  class=012333915-08082008>more</SPAN> 
  in-line</SPAN></FONT></FONT></FONT></DIV><FONT face=Arial color=#0000ff 
  size=2></FONT><BR>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> vrijdag 8 augustus 2008 
  11:06<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
  RE:<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=$BAWBN(B size=2>Hi$B!$(BWim$B!$(BPlease see my comments in line.</FONT></DIV>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
    <DIV 
    style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> <A 
    title=wim.henderickx@alcatel-lucent.be 
    href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
    href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Friday, August 08, 2008 3:39 AM</DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT><BR></DIV>
    <DIV dir=ltr align=left><FONT face=Arial>
    <DIV><FONT face=SimSun><FONT color=#0000ff><FONT size=2>Yuanlong<SPAN 
    class=289373519-07082008>, see 
    in-line</SPAN></FONT></FONT></FONT></DIV></FONT></DIV><BR>
    <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
    <HR tabIndex=-1>
    <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
    [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
    9:31<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
    RE:<BR></FONT><BR></DIV>
    <DIV></DIV>
    <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>In VPLS with dual-homed topology, we 
    must&nbsp;pre-configure the working PE and backup PE to form a protection 
    group,&nbsp;</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>such as </FONT><FONT face=$BAWBN(B size=2>described in 
    ICCP (defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT face=$BAWBN(B 
    size=2>and pre-configure working PWs with their </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>respective backup PWs, </FONT><FONT face=$BAWBN(B 
    size=2>as </FONT><FONT face=$BAWBN(B size=2>described in "PW 
    redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>Anyway, these configurations are prerequesite to 
    provide sevice resiliency for VPLS, and the roles of the respective 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>PE or PW will </FONT><FONT face=$BAWBN(B size=2>change 
    automatically after PE or PW switching (ie. backup PE becomes working PE, 
    backup PW becomes </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>working PW).</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>The mechanism in&nbsp;my draft&nbsp;is based on 
    such a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure what 
    other </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>configurations need to </FONT><FONT face=$BAWBN(B 
    size=2>be changed, can you give me some more clues?</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=SimSun size=2><SPAN class=289373519-07082008>WH&gt; We don;t 
    need any of this for PW resilience in VPLS according to me as PW status 
    together with the existing MAC flushes do the job for these 
    topologies.</SPAN></FONT></DIV>
    <DIV><FONT face=SimSun size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: But how to set 
    the&nbsp;exact&nbsp;PW status?&nbsp;When a CE is dual-homed to PEs by 
    attached circuits, and if the working AC fails, </FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><FONT size=+0>how do the 
    PEs </FONT><FONT size=+0>coordinate their PWs?&nbsp;IMHO, RFC 4762&nbsp;is 
    not perfect for these problems.&nbsp;</FONT><SPAN 
    class=012333915-08082008><FONT face=Arial color=#0000ff 
    size=2>&nbsp;</FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
    class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT size=+0><SPAN 
    class=012333915-08082008><FONT color=#0000ff size=3>Wh&gt; The way it works 
    is you use PW status as defined in 
    the&nbsp;<STRONG>draft</STRONG>-muley-dutta-<B>pw</B>e3-redundancy-bit-02.txt 
    and if the PW status changes you should perform a MAC flush in the full mesh 
    VPLS domain.</FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#0000ff size=+0><SPAN 
    class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>Jiang: As the above mentioned draft indicates, 
    pre-configuration of PWs is necessary for its working. The only difference 
    is whether </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>we should perform MAC flushing or MAC switching 
    when we do PW switching.</SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>I think blindly flushing, selectively flushing and 
    its optimizations (see Dutta's draft), and the flush-free optimizations as I 
    proposed, </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>should all be optional operations with&nbsp;their 
    different applicability and efficiency. </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>For the case of PW switching in VPLS with 
    redundancy, flush-free optimizations is more favorable to 
    me.</SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>RFC 4762 already provides some selective flushing 
    mechansms, such as flushing some specified MACs or all MACs except 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>those associated with </FONT><FONT face=$BAWBN(B 
    size=2>specific PWs. So what</FONT><FONT face=$BAWBN(B size=2> you mean may be 
    that all the existing implementations only&nbsp;provide </FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B>flushing </FONT><FONT face=$BAWBN(B>"all" MACs 
    blindly in a VPLS instance, </FONT><FONT face=$BAWBN(B>is that true?<SPAN 
    class=289373519-07082008><FONT face=Arial 
    color=#0000ff>&nbsp;</FONT></SPAN></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN 
    class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN class=289373519-07082008>WH&gt; There 
    are certain cases where we need to extend them like for PBB where we have 
    C-MAC ties to B-MACs.</SPAN></FONT></FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    size=2>Jiang: As far as I know, nowadays even some enterprise-oriented 
    routers can selectively flush MACs.<SPAN class=012333915-08082008><FONT 
    face=Arial color=#0000ff>&nbsp;</FONT></SPAN></FONT></SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    size=2><SPAN 
    class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    size=2><SPAN class=012333915-08082008><FONT face=Arial 
    color=#0000ff>WH&gt;&nbsp;Isn't the main reason for this the fact that they 
    can't handle large floods of 
    traffic?&nbsp;</FONT>&nbsp;</SPAN></FONT></SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
    <DIV><SPAN class=289373519-07082008><FONT face="Times New Roman" 
    color=#000080><SPAN class=012333915-08082008><SPAN 
    class=012333915-08082008>Jiang: Selective flushing is an optional operation, 
    not the only choice. If turning down the selecitve MAC 
    flushing</SPAN></SPAN></FONT></SPAN></DIV>
    <DIV><SPAN class=289373519-07082008><FONT face="Times New Roman" 
    color=#000080><SPAN class=012333915-08082008><SPAN 
    class=012333915-08082008>will improve the performance of a router, surely 
    you can do that.</SPAN></SPAN></FONT></SPAN></DIV>
    <DIV><FONT face=$BAWBN(B><FONT color=#000080><SPAN class=289373519-07082008><SPAN 
    class=012333915-08082008></SPAN></SPAN><SPAN 
    class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>As our anlysis and other sources(such as 
    draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>impacts all sites in a VPLS instance, even their 
    topology is not </FONT><FONT face=$BAWBN(B size=2>affected, the MACs must be 
    re-leaned may be numerous,</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>and the number of frames flooded may be not so 
    small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>service,</FONT><FONT face=$BAWBN(B size=2>and&nbsp;for 
    this&nbsp;reason&nbsp;</FONT><FONT face=$BAWBN(B size=2>we proposed some optimized 
    MAC operations such as MAC </FONT><FONT face=$BAWBN(B size=2>Switching and MAC 
    notification.</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><SPAN class=289373519-07082008><FONT face=SimSun size=2>WH&gt; The 
    optimizations which are suggested in the mentioned drafts are very simple, 
    whilst your proposal is fairly complex and depends on the topology you 
    propose but might not be efficient in case we change the service 
    topology.</FONT></SPAN></DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=SimSun size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT></BLOCKQUOTE>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: The draft mentioned above 
    also do selective flushing, not blind flushing. And for what I know, this 
    very draft may depend </FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>on the hub-and-spoke network 
    topology while our draft seems to be capable of a broader application 
    scenario such as&nbsp;a CE&nbsp;dual-homed </FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>to PEs by ACs.&nbsp;The formats and 
    contents of the LDP messages are&nbsp;similar in all these drafts or RFC for 
    selective flushing, and as </FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><FONT size=+0>my 
    analysis goes, complexity&nbsp;(times of lookup and write)</FONT><FONT 
    size=+0>&nbsp;in dataplane may be </FONT><FONT size=+0>commensurate for all 
    these mechanisms</FONT><FONT size=+0>.&nbsp;</FONT><SPAN 
    class=012333915-08082008><FONT face=Arial color=#0000ff 
    size=2>&nbsp;</FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
    class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT size=+0><SPAN class=012333915-08082008><FONT color=#0000ff><FONT 
    face=Arial size=2>WH&gt; Right but the main difference is that you only need 
    to indicate the PE or PW for which the MAC flush is performed.&nbsp;In your 
    proposal you need to indicate the full MAC address list which will be very 
    processing intensive if you have a large amount of MAC addresses in the VPLS 
    service. The message will get very big and you will have to sent lots of 
    them due to fragmentation implications. </FONT></FONT></SPAN></FONT></DIV>
    <DIV><FONT size=+0><SPAN class=012333915-08082008><FONT color=#0000ff><FONT 
    face=Arial size=2><FONT face=SimSun color=#000000><FONT 
    color=#0000ff>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides</FONT> 
    </FONT>a simple (blind was maybe the wrong wording)&nbsp;flush mechanism. 
    The indication will be fast and the re-learning will happen instantly after 
    the MAC is removed from the MAC FIB.</FONT></FONT></SPAN></FONT></DIV>
    <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008>You proposal will delay the trigger when the amount 
    of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to 
    find he proper MAC and replace it which means delay the removal of the 
    FIB.</SPAN></FONT></DIV>
    <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>Jiang:&nbsp;As&nbsp;indicated in our draft, MAC 
    Switching message can also carry only a null-list of MACs. It's OK to change 
    all the MACs associated with a PW </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>to another PW as egress.</SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><BR><FONT face=$BAWBN(B size=2>Best 
    regards,</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>Yuanlong</FONT></DIV>
    <BLOCKQUOTE dir=ltr 
    style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
      <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
      <DIV 
      style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
      <A title=wim.henderickx@alcatel-lucent.be 
      href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
      href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Thursday, August 07, 2008 1:24 
      PM</DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT><BR></DIV>
      <DIV dir=ltr align=left><SPAN class=770051905-07082008>
      <DIV dir=ltr align=left><FONT face=Arial><FONT color=#0000ff><FONT 
      size=2>Yuanlong<SPAN 
      class=924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><SPAN 
      class=770051905-07082008><FONT face=Arial color=#0000ff size=2>My point is 
      that how are you going to define the behavior on each PE when the topology 
      is adapted. I hope you are not going to change the configuration on each 
      PE in the VPLS each time.</FONT></SPAN></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=770051905-07082008><FONT face=Arial 
      color=#0000ff size=2>On top, by doing selective MAC flush you need to go 
      one by one to the items and hence the processing to withdraw the MAC(s) 
      will be longer. When you use the existing mechanisms you can flush without 
      the extra lookups and hence the performance will be faster. FLushing a MAC 
      will result in re-learning which means you do not have traffic impact, 
      just a small flooding until the MAC address is 
      re-learned.</FONT></SPAN></DIV>
      <DIV>&nbsp;</DIV>
      <DIV><SPAN class=770051905-07082008></SPAN><FONT face=Arial><FONT 
      color=#0000ff><FONT size=2>C<SPAN 
      class=770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
      <DIV><FONT size=+0><FONT color=#0000ff><FONT size=2><SPAN 
      class=770051905-07082008></SPAN></FONT></FONT></FONT><SPAN 
      class=770051905-07082008></SPAN><FONT face=Arial><FONT color=#0000ff><FONT 
      size=2>W<SPAN 
      class=770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
      <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
      <HR tabIndex=-1>
      <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
      [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
      6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
      l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
      <DIV></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>Thanks for you comment. My draft is not a 
      standlone protocol, it can be deployed with other protocol 
      suites.</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>For what I know, "PW redundancy" is being 
      standardized in PWE3 WG, which can synchronize the behavior 
of</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>the PEs and PWs in PW switching. So it is 
      not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC 
      switching</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>subsequently. </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>The MAC withdrawal process defined in RFC 4762 
      is not so simple as you think,&nbsp;it maybe not just<FONT 
      size=2>&nbsp;"flush all MACs" </FONT></FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2><FONT size=2>blindly, but can selectively flush 
      MACs, which computation comlexity is on the same level as my 
      draft</FONT><FONT size=2>.</FONT></DIV></FONT>
      <BLOCKQUOTE dir=ltr 
      style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
        <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
        <DIV 
        style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
        <A title=wim.henderickx@alcatel-lucent.be 
        href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
        href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A 
        title=sajassi@cisco.com 
        href="mailto:sajassi@cisco.com">sajassi@cisco.com</A> </DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
        href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Thursday, August 07, 2008 4:18 
        AM</DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE:</DIV>
        <DIV><BR></DIV>
        <DIV dir=ltr align=left>Yuanlong<SPAN 
        class=924080318-06082008>,</SPAN></DIV>
        <DIV dir=ltr align=left><SPAN 
        class=924080318-06082008></SPAN>&nbsp;</DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>The difficulty with your draft is that it requires 
        a PE to do different things depending in which situation the PE is in, 
        and this makes it very cumbersome in case changes happen to topologies 
        or configurations.</FONT></SPAN></DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>My view is that we need a simple MAC flush which we 
        have today in LDP.</FONT></SPAN></DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>Cheers,</FONT></SPAN></DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>WIm</FONT></SPAN></DIV><BR>
        <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
        <HR tabIndex=-1>
        <FONT face=Tahoma size=2><B>From:</B> <A 
        href="mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A> 
        [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang 
        Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> 
        <A href="mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> 
        <A href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> 
        <BR></FONT><BR></DIV>
        <DIV></DIV>
        <DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">Thanks for your comments in the last 
        L2VPN WG meeting, I will try to give you some reasons here.</FONT></DIV>
        <DIV><FONT face="Times New Roman">Following are the&nbsp;two 
        questions:</FONT></DIV>
        <DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">1. What do you think will be faster 
        convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
        <DIV><FONT face="Times New Roman">the learning is done in data plane at 
        wire speed. Why does that prolong convergence time?<BR></FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">-Faster convergence means the time to 
        associate the MACs with the "exact" out port</FONT></DIV>
        <DIV><FONT face="Times New Roman">will be shorter. Though learning is 
        done in data plane at wire speed, the time for a frame to arrive 
        </FONT></DIV>
        <DIV><FONT face="Times New Roman">at the end host and the response time 
        for a host to send a reverse frame&nbsp;accross the PE (on which the 
        </FONT></DIV>
        <DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather 
        long time. So the learning process is not only determined solely by 
        </FONT></DIV>
        <DIV><FONT face="Times New Roman">the wire </FONT><FONT 
        face="Times New Roman">speed, but also on the host processing time and 
        delays on all the intermediate routers.</FONT></DIV>
        <DIV><FONT face="Times New Roman">For this reason, flooding may also 
        last for a rather long time.</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
        <DIV><FONT face="Times New Roman">2. You consider a simple case with 
        simple backup for primary.&nbsp;In cases where primary fails and traffic 
        gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
        that?</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">-I believe the most commonly 
        deployed&nbsp;scenario is&nbsp;one backup PE for a primary 
        PE.</FONT></DIV>
        <DIV><FONT face="Times New Roman">In my opinion, for the scenario of 
        multiple backup PEs protecting a primary PE, </FONT></DIV>
        <DIV><FONT face="Times New Roman">only one PE out of them will be 
        elected as the working PE when the primary PE failed, </FONT></DIV>
        <DIV><FONT face="Times New Roman">rather than distribute the traffic 
        over multiple backup PEs,</FONT></DIV>
        <DIV><FONT face="Times New Roman">otherwise, loops may be introduced to 
        the attached customer site.</FONT></DIV>
        <DIV><FONT face="Times New Roman">BTW, in a similar way, for PW 
        redundancy mechanism, only 1 secondary is selected as the working PW 
        (see </FONT></DIV>
        <DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
        in&nbsp;PWE3 WG)</FONT></DIV>
        <DIV><FONT face="Times New Roman">Anyway, if there is any such 
        requirements confirmed, I think it will be OK for us to find some 
        solutions and </FONT></DIV>
        <DIV><FONT face="Times New Roman">add them to&nbsp;the future 
        drafts.&nbsp;</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
        <DIV><FONT 
    face="Times New Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

--Boundary_(ID_vWbJgKuThrZutKTMcibMBw)--


From l2vpn-bounces@ietf.org  Mon Aug 11 01:12:44 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 817F33A6C7B;
	Mon, 11 Aug 2008 01:12:44 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id A45AD3A6C7B
	for <l2vpn@core3.amsl.com>; Mon, 11 Aug 2008 01:12:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.912
X-Spam-Level: 
X-Spam-Status: No, score=0.912 tagged_above=-999 required=5 tests=[AWL=0.226, 
	BAYES_00=-2.599, HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001, 
	J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6, 
	J_CHICKENPOX_92=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id r-dP1Us9sZBV for <l2vpn@core3.amsl.com>;
	Mon, 11 Aug 2008 01:12:41 -0700 (PDT)
Received: from szxga02-in.huawei.com (szxga02-in.huawei.com [119.145.14.65])
	by core3.amsl.com (Postfix) with ESMTP id BCC783A6C57
	for <l2vpn@ietf.org>; Mon, 11 Aug 2008 01:12:40 -0700 (PDT)
Received: from huawei.com (szxga02-in [172.24.2.6])
	by szxga02-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F003YIFZDIV@szxga02-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 16:09:13 +0800 (CST)
Received: from huawei.com ([172.24.1.18])
	by szxga02-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00E0VFZDET@szxga02-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 16:09:13 +0800 (CST)
Received: from j59929 ([10.70.77.69])
	by szxml03-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5F00G9BFZCPQ@szxml03-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 16:09:13 +0800 (CST)
Date: Mon, 11 Aug 2008 16:09:10 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <001e01c8fb89$890bc380$454d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_VIEoS1S/LtVV2GiE2/Lwhw)"
X-Priority: 3
X-MSMail-priority: Normal
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
	<007d01c8f85f$99578350$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C877@FRVELSMBS22.ad2.ad.alcatel.com>
	<007301c8f935$ea290980$454d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302CB8F@FRVELSMBS22.ad2.ad.alcatel.com>
	<006101c8fb59$bd4a6690$454d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302CBE2@FRVELSMBS22.ad2.ad.alcatel.com>
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_VIEoS1S/LtVV2GiE2/Lwhw)
Content-type: text/plain; charset=iso-2022-jp
Content-transfer-encoding: 7BIT

Hi, Wim:
I am not sure what you mean by "This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03." 
If you mean "MAC switching" behavior, forgive me if I am not aware of it in the above draft.
Otherwise,if you mean the null-list of MACs,  it is RFC 4762 which first introduces such a kind of sub-TLV.

As to the benefits of draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed them in the emails to this WG
mailing list several days ago, I would be more than willing to re-sumarize them here:

1. Faster convergence of topology, maybe more precisely, faster MAC learning compared with MAC flooding.
-Faster convergence means the time to associate the MACs with the "exact" out port
will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
the wire speed, but also on the host processing time and delays on all the intermediate routers.
For this reason, flooding as a result of MAC flooding may last for a rather long time.

With MAC switching, FIB can be established as fast as the MAC flushing operation itself:
In selective flushing as described in RFC 4762 or draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,
the MAC flushing operation may be:
    Given PW1, for each item in the FIB, 
            if the item's outport is not the same as this PW, then flush it.

In MAC switching, as described in draft-jiang-l2vpn-vpls-mac-operation-00, the MAC switching on the FIB table: 
     Given PW1, PW2, for each item in the FIB, 
            if the item's outport is the same as PW1, then change it to PW2.

As the pseudocode shown above, MAC switching is no more complex than the selective MAC flushing.
Actually, with more accurate PW information, it is quite possible that less PWs will be involved in the data plane lookup.

2. Less flooding burst in the core. 
For the customer sites with hundreds or more of MACs, a lot of MAC learning will be needed, and for the customer site 
with few MACs but large bandwidth, although few MACs are flushed, but as the time for the responsive frame to arrive 
at the PEs where MACs are flushed is not trivial, the total frames flooded accross the core may be added to a high volume.

Optimized selecive flooding as described in Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate the MAC learning and 
flooding problem to some degree, but not so completely, and it applies only to a certain hub-and-spoke network topology, 
if CE is dual-homed to the PEs by ACs, it will be out of the capability of this draft. While draft-jiang-l2vpn-vpls-mac-operation-00 
tries to eliminate the MAC leaning and subsequent flooding procedures due to PE or PW switching completely, and in a broader way.

Regards
Yuanlong


Snippet of your last email:

--------------------------------------------------------------------------
      From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
      Sent: donderdag 7 augustus 2008 9:31
      To: HENDERICKX Wim
      Cc: l2vpn@ietf.org
      Subject: Re: RE:


      Wim:

      In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group, 
      such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their 
      respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
      Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective 
      PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes 
      working PW).
      The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other 
      configurations need to be changed, can you give me some more clues?

      WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.

      Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails, 
      how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems.   

      WH> Your draft explains a solution with an MTU-s attached with PW. For this the MTU-s decides the PW status bit. In case of AC attached you need something like ICCP, but this is independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) becomes active.

      Wh> The way it works is you use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt and if the PW status changes you should perform a MAC flush in the full mesh VPLS domain.

      Jiang: As the above mentioned draft indicates, pre-configuration of PWs is necessary for its working. The only difference is whether 
      we should perform MAC flushing or MAC switching when we do PW switching.
      I think blindly flushing, selectively flushing and its optimizations (see Dutta's draft), and the flush-free optimizations as I proposed, 
      should all be optional operations with their different applicability and efficiency. 
      For the case of PW switching in VPLS with redundancy, flush-free optimizations is more favorable to me.

      RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except 
      those associated with specific PWs. So what you mean may be that all the existing implementations only provide 
      flushing "all" MACs blindly in a VPLS instance, is that true? 

      WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
      Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs. 

      WH> Isn't the main reason for this the fact that they can't handle large floods of traffic?  

      Jiang: Selective flushing is an optional operation, not the only choice. If turning down the selecitve MAC flushing
      will improve the performance of a router, surely you can do that.

      As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
      impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
      and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
      service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.

      WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
      Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend 
      on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed 
      to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as 
      my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms.  

      WH> Right but the main difference is that you only need to indicate the PE or PW for which the MAC flush is performed. In your proposal you need to indicate the full MAC address list which will be very processing intensive if you have a large amount of MAC addresses in the VPLS service. The message will get very big and you will have to sent lots of them due to fragmentation implications. 
      Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe the wrong wording) flush mechanism. The indication will be fast and the re-learning will happen instantly after the MAC is removed from the MAC FIB.

      You proposal will delay the trigger when the amount of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to find he proper MAC and replace it which means delay the removal of the FIB.

      Jiang: As indicated in our draft, MAC Switching message can also carry only a null-list of MACs. It's OK to change all the MACs associated with a PW 
      to another PW as egress. 

      WH> This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think your solution is better?


      Best regards,

      Yuanlong

--Boundary_(ID_VIEoS1S/LtVV2GiE2/Lwhw)
Content-type: text/html; charset=iso-2022-jp
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-2022-jp">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face="Times New Roman">Hi, Wim:</FONT></DIV>
<DIV><FONT face="Times New Roman">I am not sure what&nbsp;you mean by "This 
function is already proposed in&nbsp;&nbsp;<SPAN 
class=012333915-08082008>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03." 
</SPAN></FONT></DIV>
<DIV><SPAN class=012333915-08082008><FONT face="Times New Roman">If you mean 
"MAC switching" behavior, forgive me if I am not aware of it in the above 
draft.</FONT></SPAN></DIV>
<DIV><SPAN class=012333915-08082008><FONT face="Times New Roman">Otherwise,if 
you mean the null-list of MACs,&nbsp; it is RFC 4762&nbsp;which first 
introduces&nbsp;such&nbsp;a kind of sub-TLV.</FONT></SPAN></DIV>
<DIV><SPAN class=012333915-08082008><FONT 
face="Times New Roman"></FONT></SPAN>&nbsp;</DIV>
<DIV><SPAN class=012333915-08082008><FONT face="Times New Roman">As to the 
benefits of draft-jiang-l2vpn-vpls-mac-operation-00,&nbsp;although I had 
analyzed them in the emails to this WG</FONT></SPAN></DIV>
<DIV><SPAN class=012333915-08082008><FONT face="Times New Roman">mailing list 
several days ago, I would be more than willing to re-sumarize them 
here:</FONT></SPAN></DIV>
<DIV><SPAN class=012333915-08082008><FONT 
face="Times New Roman"></FONT></SPAN>&nbsp;</DIV>
<DIV><SPAN class=012333915-08082008><FONT face="Times New Roman">1. Faster 
convergence of topology, maybe more precisely, faster MAC learning compared with 
MAC flooding.</FONT></SPAN></DIV>
<DIV><SPAN class=012333915-08082008>
<DIV><FONT face="Times New Roman">-Faster convergence means the time to 
associate the MACs with the "exact" out port</FONT></DIV>
<DIV><FONT face="Times New Roman">will be shorter. Though learning is done in 
data plane at wire speed, the time for a frame to arrive </FONT></DIV>
<DIV><FONT face="Times New Roman">at the end host and the response time for a 
host to send a reverse frame&nbsp;accross the PE (on which the </FONT></DIV>
<DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather long 
time. So the learning process is not only determined solely by </FONT></DIV>
<DIV><FONT face="Times New Roman">the wire speed, but also on the host 
processing time and delays on all the intermediate routers.</FONT></DIV>
<DIV><FONT face="Times New Roman">For this reason, flooding as a result of MAC 
flooding may last for a rather long time.</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">With MAC switching, FIB can be established as 
fast as the MAC flushing&nbsp;operation itself:</FONT></DIV>
<DIV>
<DIV>
<DIV><FONT face="Times New Roman">In selective flushing as described in RFC 4762 
or d<SPAN 
class=012333915-08082008>raft-pdutta-l2vpn-vpls-ldp-mac-opt-03,</SPAN></FONT></DIV>
<DIV><SPAN class=012333915-08082008><FONT face="Times New Roman">the MAC 
flushing&nbsp;operation may be:</FONT></SPAN></DIV>
<DIV><FONT face="Times New Roman"><EM>&nbsp;&nbsp;&nbsp; Given PW1, for 
each&nbsp;item in the FIB, </EM></FONT></DIV>
<DIV><FONT 
face="Times New Roman"><EM>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
if the&nbsp;item's outport&nbsp;is not the same as this PW,&nbsp;then flush 
it.</EM></FONT></DIV></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">In MAC switching, as described 
in&nbsp;draft-jiang-l2vpn-vpls-mac-operation-00, the&nbsp;MAC&nbsp;switching 
on&nbsp;the FIB table: </FONT></DIV>
<DIV><FONT face="Times New Roman"><EM>&nbsp;&nbsp;&nbsp;&nbsp; Given PW1, PW2, 
for each&nbsp;item in the FIB, </EM></FONT></DIV>
<DIV>
<DIV><FONT 
face="Times New Roman"><EM>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
if the&nbsp;item's outport&nbsp;is the same as PW1, then change&nbsp;it 
to&nbsp;PW2.</EM></FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">As the pseudocode shown above, MAC switching 
is no more complex than the selective MAC flushing.</FONT></DIV>
<DIV><FONT face="Times New Roman">Actually, with more accurate PW information, 
it is quite possible that less PWs will be involved in the data plane 
lookup.</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV></DIV>
<DIV><FONT face="Times New Roman">2. Less flooding burst in the core. 
</FONT></DIV>
<DIV><FONT face="Times New Roman">For the customer sites&nbsp;with hundreds or 
more of MACs, a lot of MAC learning will be needed, and&nbsp;for the customer 
site </FONT></DIV>
<DIV><FONT face="Times New Roman">with few MACs but large bandwidth, although 
few MACs are flushed, but as the&nbsp;time for the responsive frame to arrive 
</FONT></DIV>
<DIV><FONT face="Times New Roman">at the PEs where MACs are flushed&nbsp;is not 
trivial, the total frames flooded accross the core may be added to a high 
volume.</FONT></DIV>
<DIV><FONT face="Times New Roman"></FONT><FONT face=$BAWBN(B 
size=2></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">Optimized selecive flooding as described in 
<SPAN class=012333915-08082008>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 </SPAN>can 
alleviate the MAC learning and </FONT></DIV>
<DIV><FONT face="Times New Roman">flooding problem </FONT><FONT 
face="Times New Roman">to some degree,&nbsp;but not so completely, and it 
applies only&nbsp;to a certain hub-and-spoke network topology, </FONT></DIV>
<DIV><FONT face="Times New Roman">if CE is dual-homed&nbsp;to </FONT><FONT 
face="Times New Roman">the PEs by ACs, it will be out of the capability of this 
draft. While draft-jiang-l2vpn-vpls-mac-operation-00 </FONT></DIV>
<DIV><FONT face="Times New Roman">tries to eliminate </FONT><FONT 
face="Times New Roman">the MAC leaning and subsequent </FONT><FONT 
face="Times New Roman">flooding procedures due to PE or PW switching completely, 
and in a broader way.</FONT></DIV>
<DIV>
<DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman">Regards</FONT></DIV>
<DIV><FONT face="Times New Roman">Yuanlong</FONT></DIV></DIV></SPAN></DIV>
<DIV><SPAN class=012333915-08082008><FONT 
face="Times New Roman"></FONT></SPAN>&nbsp;</DIV>
<DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><BR><FONT 
face="Times New Roman">Snippet of your last email:</FONT></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <BLOCKQUOTE dir=ltr 
    style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
      <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
      <HR tabIndex=-1>
      <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
      [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
      9:31<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
      RE:<BR></FONT><BR></DIV>
      <DIV></DIV>
      <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>In VPLS with dual-homed topology, we 
      must&nbsp;pre-configure the working PE and backup PE to form a protection 
      group,&nbsp;</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>such as </FONT><FONT face=$BAWBN(B size=2>described in 
      ICCP (defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT face=$BAWBN(B 
      size=2>and pre-configure working PWs with their </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>respective backup PWs, </FONT><FONT face=$BAWBN(B 
      size=2>as </FONT><FONT face=$BAWBN(B size=2>described in "PW 
      redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>Anyway, these configurations are prerequesite to 
      provide sevice resiliency for VPLS, and the roles of the respective 
      </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>PE or PW will </FONT><FONT face=$BAWBN(B size=2>change 
      automatically after PE or PW switching (ie. backup PE becomes working PE, 
      backup PW becomes </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>working PW).</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>The mechanism in&nbsp;my draft&nbsp;is based on 
      such a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure 
      what other </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>configurations need to </FONT><FONT face=$BAWBN(B 
      size=2>be changed, can you give me some more clues?</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=SimSun size=2><SPAN class=289373519-07082008>WH&gt; We 
      don;t need any of this for PW resilience in VPLS according to me as PW 
      status together with the existing MAC flushes do the job for these 
      topologies.</SPAN></FONT></DIV>
      <DIV><FONT face=SimSun size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: But how to set 
      the&nbsp;exact&nbsp;PW status?&nbsp;When a CE is dual-homed to PEs by 
      attached circuits, and if the working AC fails, </FONT></DIV>
      <DIV><FONT face=$BAWBN(B><FONT size=2><FONT color=#800000>how do the PEs 
      coordinate their PWs?&nbsp;IMHO, RFC 4762&nbsp;is not perfect for these 
      problems.&nbsp;</FONT><FONT color=#0000ff><SPAN 
      class=012333915-08082008><FONT face=Arial>&nbsp;<SPAN 
      class=887291606-11082008>&nbsp;</SPAN></FONT></SPAN></FONT></FONT></FONT></DIV>
      <DIV><FONT face=$BAWBN(B><FONT size=2><FONT color=#0000ff><SPAN 
      class=012333915-08082008><FONT face=Arial><SPAN 
      class=887291606-11082008></SPAN></FONT></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B><FONT size=2><FONT color=#0000ff><SPAN 
      class=012333915-08082008><FONT face=Arial><SPAN 
      class=887291606-11082008>WH&gt;&nbsp;Your draft explains&nbsp;a solution 
      with an MTU-s attached with PW.&nbsp;For this&nbsp;the MTU-s decides the 
      PW status bit. In case of AC attached you need something like ICCP, but 
      this is independent of RFC 4762. Based on the ICCP mechanism one of the 
      AC(s) becomes active.</SPAN></FONT></SPAN></FONT></FONT></FONT></DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
      class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><FONT size=+0><SPAN 
      class=012333915-08082008><FONT color=#0000ff size=3>Wh&gt; The way it 
      works is you use PW status as defined in 
      the&nbsp;<STRONG>draft</STRONG>-muley-dutta-<B>pw</B>e3-redundancy-bit-02.txt 
      and if the PW status changes you should perform a MAC flush in the full 
      mesh VPLS domain.</FONT></SPAN></FONT></FONT></FONT></DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#0000ff size=+0><SPAN 
      class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>Jiang: As the above mentioned draft indicates, 
      pre-configuration of PWs is necessary for its working. The only difference 
      is whether </SPAN></FONT></DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>we should perform MAC flushing or MAC switching 
      when we do PW switching.</SPAN></FONT></DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>I think blindly flushing, selectively flushing 
      and its optimizations (see Dutta's draft), and the flush-free 
      optimizations as I proposed, </SPAN></FONT></DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>should all be optional operations with&nbsp;their 
      different applicability and efficiency. </SPAN></FONT></DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>For the case of PW switching in VPLS with 
      redundancy, flush-free optimizations is more favorable to 
      me.</SPAN></FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>RFC 4762 already provides some selective 
      flushing mechansms, such as flushing some specified MACs or all MACs 
      except </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>those associated with </FONT><FONT face=$BAWBN(B 
      size=2>specific PWs. So what</FONT><FONT face=$BAWBN(B size=2> you mean may be 
      that all the existing implementations only&nbsp;provide </FONT></DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B>flushing </FONT><FONT face=$BAWBN(B>"all" MACs 
      blindly in a VPLS instance, </FONT><FONT face=$BAWBN(B>is that true?<SPAN 
      class=289373519-07082008><FONT face=Arial 
      color=#0000ff>&nbsp;</FONT></SPAN></FONT></FONT></DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN 
      class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN class=289373519-07082008>WH&gt; 
      There are certain cases where we need to extend them like for PBB where we 
      have C-MAC ties to B-MACs.</SPAN></FONT></FONT></DIV>
      <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
      size=2>Jiang: As far as I know, nowadays even some enterprise-oriented 
      routers can selectively flush MACs.<SPAN class=012333915-08082008><FONT 
      face=Arial color=#0000ff>&nbsp;</FONT></SPAN></FONT></SPAN></FONT></DIV>
      <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
      size=2><SPAN 
      class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
      size=2><SPAN class=012333915-08082008><FONT face=Arial 
      color=#0000ff>WH&gt;&nbsp;Isn't the main reason for this the fact that 
      they can't handle large floods of 
      traffic?&nbsp;</FONT>&nbsp;</SPAN></FONT></SPAN></FONT></DIV>
      <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
      face=Arial color=#0000ff size=2><SPAN 
      class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
      <DIV><SPAN class=289373519-07082008><FONT face="Times New Roman" 
      color=#000080><SPAN class=012333915-08082008><SPAN 
      class=012333915-08082008>Jiang: Selective flushing is an optional 
      operation, not the only choice. If turning down the selecitve MAC 
      flushing</SPAN></SPAN></FONT></SPAN></DIV>
      <DIV><SPAN class=289373519-07082008><FONT face="Times New Roman" 
      color=#000080><SPAN class=012333915-08082008><SPAN 
      class=012333915-08082008>will improve the performance of a router, surely 
      you can do that.</SPAN></SPAN></FONT></SPAN></DIV>
      <DIV><FONT face=$BAWBN(B><FONT color=#000080><SPAN 
      class=289373519-07082008><SPAN 
      class=012333915-08082008></SPAN></SPAN><SPAN 
      class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>As our anlysis and other sources(such as 
      draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
      </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>impacts all sites in a VPLS instance, even their 
      topology is not </FONT><FONT face=$BAWBN(B size=2>affected, the MACs must be 
      re-leaned may be numerous,</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>and the number of frames flooded may be not so 
      small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
      </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>service,</FONT><FONT face=$BAWBN(B size=2>and&nbsp;for 
      this&nbsp;reason&nbsp;</FONT><FONT face=$BAWBN(B size=2>we proposed some 
      optimized MAC operations such as MAC </FONT><FONT face=$BAWBN(B size=2>Switching 
      and MAC notification.</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><SPAN class=289373519-07082008><FONT face=SimSun size=2>WH&gt; The 
      optimizations which are suggested in the mentioned drafts are very simple, 
      whilst your proposal is fairly complex and depends on the topology you 
      propose but might not be efficient in case we change the service 
      topology.</FONT></SPAN></DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=SimSun 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT></BLOCKQUOTE>
    <BLOCKQUOTE dir=ltr 
    style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
      <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: The draft mentioned above 
      also do selective flushing, not blind flushing. And for what I know, this 
      very draft may depend </FONT></DIV>
      <DIV><FONT face=$BAWBN(B color=#800000 size=2>on the hub-and-spoke network 
      topology while our draft seems to be capable of a broader application 
      scenario such as&nbsp;a CE&nbsp;dual-homed </FONT></DIV>
      <DIV><FONT face=$BAWBN(B color=#800000 size=2>to PEs by ACs.&nbsp;The formats 
      and contents of the LDP messages are&nbsp;similar in all these drafts or 
      RFC for selective flushing, and as </FONT></DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><FONT size=+0>my 
      analysis goes, complexity&nbsp;(times of lookup and write)</FONT><FONT 
      size=+0>&nbsp;in dataplane may be </FONT><FONT size=+0>commensurate for 
      all these mechanisms</FONT><FONT size=+0>.&nbsp;</FONT><SPAN 
      class=012333915-08082008><FONT face=Arial color=#0000ff 
      size=2>&nbsp;</FONT></SPAN></FONT></FONT></FONT></DIV>
      <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
      class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
      <DIV><FONT size=+0><SPAN class=012333915-08082008><FONT 
      color=#0000ff><FONT face=Arial size=2>WH&gt; Right but the main difference 
      is that you only need to indicate the PE or PW for which the MAC flush is 
      performed.&nbsp;In your proposal you need to indicate the full MAC address 
      list which will be very processing intensive if you have a large amount of 
      MAC addresses in the VPLS service. The message will get very big and you 
      will have to sent lots of them due to fragmentation implications. 
      </FONT></FONT></SPAN></FONT></DIV>
      <DIV><FONT size=+0><SPAN class=012333915-08082008><FONT 
      color=#0000ff><FONT face=Arial size=2><FONT face=SimSun 
      color=#000000><FONT color=#0000ff>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 
      provides</FONT> </FONT>a simple (blind was maybe the wrong 
      wording)&nbsp;flush mechanism. The indication will be fast and the 
      re-learning will happen instantly after the MAC is removed from the MAC 
      FIB.</FONT></FONT></SPAN></FONT></DIV>
      <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
      class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
      <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
      class=012333915-08082008>You proposal will delay the trigger when the 
      amount of MAC @ grows and on top the receiving PE needs a fast lookup 
      mechanism to find he proper MAC and replace it which means delay the 
      removal of the FIB.</SPAN></FONT></DIV>
      <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
      class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>Jiang:&nbsp;As&nbsp;indicated in our draft, MAC 
      Switching message can also carry only a null-list of MACs. It's OK to 
      change all the MACs associated with a PW </SPAN></FONT></DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008>to another PW as egress.<SPAN 
      class=887291606-11082008><FONT face=Arial color=#0000ff 
      size=2>&nbsp;</FONT></SPAN></SPAN></FONT></DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008><SPAN 
      class=887291606-11082008></SPAN></SPAN></FONT>&nbsp;</DIV>
      <DIV><FONT face="Times New Roman" color=#000080><SPAN 
      class=012333915-08082008><SPAN class=887291606-11082008><FONT face=Arial 
      color=#0000ff size=2>WH&gt; This function is already proposed 
      in&nbsp;</FONT>&nbsp;<SPAN class=012333915-08082008><FONT 
      color=#0000ff><FONT face=Arial size=2><FONT face=SimSun 
      color=#000000><FONT color=#0000ff>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. 
      Why do you think your solution is 
      better?</FONT></FONT></FONT></FONT></SPAN></SPAN></SPAN></FONT></DIV>
      <DIV><FONT face=$BAWBN(B></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
      size=2></FONT><BR><FONT face=$BAWBN(B size=2>Best regards,</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>Yuanlong</FONT></DIV>
      <BLOCKQUOTE dir=ltr 
      style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"><FONT 
        face=$BAWBN(B 
size=2></FONT>&nbsp;</BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

--Boundary_(ID_VIEoS1S/LtVV2GiE2/Lwhw)--


From l2vpn-bounces@ietf.org  Mon Aug 11 01:19:10 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 15CDB3A6C88;
	Mon, 11 Aug 2008 01:19:10 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 727B13A6C88
	for <l2vpn@core3.amsl.com>; Mon, 11 Aug 2008 01:19:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.114
X-Spam-Level: *
X-Spam-Status: No, score=1.114 tagged_above=-999 required=5 tests=[AWL=0.962, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35, HTML_MESSAGE=0.001,
	J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6,
	J_CHICKENPOX_92=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id IUReLawK9uBm for <l2vpn@core3.amsl.com>;
	Mon, 11 Aug 2008 01:19:06 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [64.208.49.27])
	by core3.amsl.com (Postfix) with ESMTP id 1238E3A6C71
	for <l2vpn@ietf.org>; Mon, 11 Aug 2008 01:19:05 -0700 (PDT)
Received: from FRVELSBHS02.ad2.ad.alcatel.com (frvelsbhs02.ad2.ad.alcatel.com
	[155.132.6.74])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7B8IY2T008801; 
	Mon, 11 Aug 2008 10:18:35 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS02.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Mon, 11 Aug 2008 10:18:34 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8FB8A.D86B1872"
Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
Date: Mon, 11 Aug 2008 10:18:33 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D7023E7A1C@FRVELSMBS22.ad2.ad.alcatel.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
Thread-Index: Acj7if0ad63S+qb4SFCK+kZ2I8HJfwAANuLX
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: <yljiang@huawei.com>
X-OriginalArrivalTime: 11 Aug 2008 08:18:34.0700 (UTC)
	FILETIME=[D921F8C0:01C8FB8A]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8FB8A.D86B1872
Content-Type: text/plain;
	charset="UTF-8"
Content-Transfer-Encoding: base64
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------_=_NextPart_001_01C8FB8A.D86B1872
Content-Type: text/html;
	charset="UTF-8"
Content-Transfer-Encoding: base64
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------_=_NextPart_001_01C8FB8A.D86B1872--


From l2vpn-bounces@ietf.org  Mon Aug 11 01:52:31 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 5C2A13A69DF;
	Mon, 11 Aug 2008 01:52:31 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 1D0C63A6BED
	for <l2vpn@core3.amsl.com>; Mon, 11 Aug 2008 01:52:30 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.895
X-Spam-Level: **
X-Spam-Status: No, score=2.895 tagged_above=-999 required=5 tests=[AWL=-1.907, 
	BAYES_00=-2.599, FAKE_REPLY_C=2.012, FH_RELAY_NODNS=1.451,
	HELO_MISMATCH_COM=0.553, HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001, 
	J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6,
	J_CHICKENPOX_92=0.6, RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id xxtZNtey-1Da for <l2vpn@core3.amsl.com>;
	Mon, 11 Aug 2008 01:52:28 -0700 (PDT)
Received: from szxga04-in.huawei.com (unknown [119.145.14.67])
	by core3.amsl.com (Postfix) with ESMTP id 9FBBE3A6359
	for <l2vpn@ietf.org>; Mon, 11 Aug 2008 01:52:27 -0700 (PDT)
Received: from huawei.com (szxga04-in [172.24.2.12])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00FXNHUGZA@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 16:49:28 +0800 (CST)
Received: from huawei.com ([172.24.1.24])
	by szxga04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00JEKHUGPV@szxga04-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 16:49:28 +0800 (CST)
Received: from j59929 ([10.70.77.69])
	by szxml04-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5F00I95HUGPS@szxml04-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 16:49:28 +0800 (CST)
Date: Mon, 11 Aug 2008 16:49:26 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <007c01c8fb8f$28bd7860$454d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_eAYkftLrk6ysS22ffCDydA)"
X-Priority: 3
X-MSMail-priority: Normal
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_eAYkftLrk6ysS22ffCDydA)
Content-type: text/plain; charset=utf-8
Content-transfer-encoding: 7BIT

Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00The difference is: 

In draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, selectively flushing MACs by carrying a unique PE-ID.

In draft-jiang-l2vpn-vpls-mac-operation-00, MAC switching by carrying PW info.

Please see my last email for the benefits, if there is anything you don't agree with me, please 
point it out and give your reasons.


  ----- Original Message ----- 
  From: HENDERICKX Wim 
  To: yljiang@huawei.com 
  Cc: l2vpn@ietf.org 
  Sent: Monday, August 11, 2008 4:18 PM
  Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00


  Yanlong, my question is what is the difference between you draft and draft-pdutta-l2vpn-vpls-ldp-mac-opt-03? The function described are the same and I don't see the added value you bring with your draft.

  Cheers,
  Wim
  _________________
  sent from blackberry

  ----- Original Message -----
  From: Jiang Yuan-long <yljiang@huawei.com>
  To: HENDERICKX Wim
  Cc: l2vpn@ietf.org <l2vpn@ietf.org>
  Sent: Mon Aug 11 10:09:10 2008
  Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00

  Hi, Wim:
  I am not sure what you mean by "This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03."
  If you mean "MAC switching" behavior, forgive me if I am not aware of it in the above draft.
  Otherwise,if you mean the null-list of MACs,  it is RFC 4762 which first introduces such a kind of sub-TLV.

  As to the benefits of draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed them in the emails to this WG
  mailing list several days ago, I would be more than willing to re-sumarize them here:

  1. Faster convergence of topology, maybe more precisely, faster MAC learning compared with MAC flooding.
  -Faster convergence means the time to associate the MACs with the "exact" out port
  will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive
  at the end host and the response time for a host to send a reverse frame accross the PE (on which the
  MACs is flushed) will be a rather long time. So the learning process is not only determined solely by
  the wire speed, but also on the host processing time and delays on all the intermediate routers.
  For this reason, flooding as a result of MAC flooding may last for a rather long time.

  With MAC switching, FIB can be established as fast as the MAC flushing operation itself:
  In selective flushing as described in RFC 4762 or draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,
  the MAC flushing operation may be:
      Given PW1, for each item in the FIB,
              if the item's outport is not the same as this PW, then flush it.

  In MAC switching, as described in draft-jiang-l2vpn-vpls-mac-operation-00, the MAC switching on the FIB table:
       Given PW1, PW2, for each item in the FIB,
              if the item's outport is the same as PW1, then change it to PW2.

  As the pseudocode shown above, MAC switching is no more complex than the selective MAC flushing.
  Actually, with more accurate PW information, it is quite possible that less PWs will be involved in the data plane lookup.

  2. Less flooding burst in the core.
  For the customer sites with hundreds or more of MACs, a lot of MAC learning will be needed, and for the customer site
  with few MACs but large bandwidth, although few MACs are flushed, but as the time for the responsive frame to arrive
  at the PEs where MACs are flushed is not trivial, the total frames flooded accross the core may be added to a high volume.

  Optimized selecive flooding as described in Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate the MAC learning and
  flooding problem to some degree, but not so completely, and it applies only to a certain hub-and-spoke network topology,
  if CE is dual-homed to the PEs by ACs, it will be out of the capability of this draft. While draft-jiang-l2vpn-vpls-mac-operation-00
  tries to eliminate the MAC leaning and subsequent flooding procedures due to PE or PW switching completely, and in a broader way.

  Regards
  Yuanlong


  Snippet of your last email:

  ________________________________

                          From: Jiang Yuan-long [mailto:yljiang@huawei.com]
                          Sent: donderdag 7 augustus 2008 9:31
                          To: HENDERICKX Wim
                          Cc: l2vpn@ietf.org
                          Subject: Re: RE:
                         
                         
                          Wim:
                          
                          In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group,
                          such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their
                          respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
                          Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective
                          PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes
                          working PW).
                          The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other
                          configurations need to be changed, can you give me some more clues?
                          
                          WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.
                          
                          Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails,
                          how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems.  
                          
                          WH> Your draft explains a solution with an MTU-s attached with PW. For this the MTU-s decides the PW status bit. In case of AC attached you need something like ICCP, but this is independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) becomes active.
                          
                          Wh> The way it works is you use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt and if the PW status changes you should perform a MAC flush in the full mesh VPLS domain.
                          
                          Jiang: As the above mentioned draft indicates, pre-configuration of PWs is necessary for its working. The only difference is whether
                          we should perform MAC flushing or MAC switching when we do PW switching.
                          I think blindly flushing, selectively flushing and its optimizations (see Dutta's draft), and the flush-free optimizations as I proposed,
                          should all be optional operations with their different applicability and efficiency.
                          For the case of PW switching in VPLS with redundancy, flush-free optimizations is more favorable to me.
                          
                          RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except
                          those associated with specific PWs. So what you mean may be that all the existing implementations only provide
                          flushing "all" MACs blindly in a VPLS instance, is that true?
                          
                          WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
                          Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs.
                          
                          WH> Isn't the main reason for this the fact that they can't handle large floods of traffic? 
                          
                          Jiang: Selective flushing is an optional operation, not the only choice. If turning down the selecitve MAC flushing
                          will improve the performance of a router, surely you can do that.
                          
                          As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely
                          impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
                          and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS
                          service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.
                          
                          WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
                         
                          Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend
                          on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed
                          to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as
                          my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms. 
                          
                          WH> Right but the main difference is that you only need to indicate the PE or PW for which the MAC flush is performed. In your proposal you need to indicate the full MAC address list which will be very processing intensive if you have a large amount of MAC addresses in the VPLS service. The message will get very big and you will have to sent lots of them due to fragmentation implications.
                          Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe the wrong wording) flush mechanism. The indication will be fast and the re-learning will happen instantly after the MAC is removed from the MAC FIB.
                          
                          You proposal will delay the trigger when the amount of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to find he proper MAC and replace it which means delay the removal of the FIB.
                          
                          Jiang: As indicated in our draft, MAC Switching message can also carry only a null-list of MACs. It's OK to change all the MACs associated with a PW
                          to another PW as egress.
                          
                          WH> This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think your solution is better?
                          
                         
                          Best regards,
                          
                          Yuanlong

                                  



--Boundary_(ID_eAYkftLrk6ysS22ffCDydA)
Content-type: text/html; charset=utf-8
Content-transfer-encoding: quoted-printable

=EF=BB=BF<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD><TITLE>Re: RE, and benefits of =
draft-jiang-l2vpn-vpls-mac-operation-00</TITLE>
<META http-equiv=3DContent-Type content=3D"text/html; charset=3Dutf-8">
<META content=3D"MSHTML 6.00.2900.3354" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2>The difference is: =
</FONT></DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2>In =
draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, selectively=20
<EM>flushing</EM> MACs by carrying a unique PE-ID.</FONT></DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2>In =
draft-jiang-l2vpn-vpls-mac-operation-00, MAC=20
<EM>switching</EM> by carrying PW info.</FONT></DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2>Please see my last email =
for the benefits, if there is=20
anything you don't agree with me, please </FONT></DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2>point it out and give your =
reasons.</FONT></DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2></FONT>&nbsp;</DIV>
<DIV><FONT face=3D=E5=AE=8B=E4=BD=93 size=3D2></FONT>&nbsp;</DIV>
<BLOCKQUOTE dir=3Dltr=20
style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; =
BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style=3D"FONT: 9pt =E5=AE=8B=E4=BD=93">----- Original Message =
----- </DIV>
  <DIV style=3D"BACKGROUND: #e4e4e4; FONT: 9pt =E5=AE=8B=E4=BD=93; =
font-color: black"><B>From:</B>=20
  <A title=3Dwim.henderickx@alcatel-lucent.be=20
  href=3D"mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> =
</DIV>
  <DIV style=3D"FONT: 9pt =E5=AE=8B=E4=BD=93"><B>To:</B> <A =
title=3Dyljiang@huawei.com=20
  href=3D"mailto:yljiang@huawei.com">yljiang@huawei.com</A> </DIV>
  <DIV style=3D"FONT: 9pt =E5=AE=8B=E4=BD=93"><B>Cc:</B> <A =
title=3Dl2vpn@ietf.org=20
  href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style=3D"FONT: 9pt =E5=AE=8B=E4=BD=93"><B>Sent:</B> Monday, =
August 11, 2008 4:18 PM</DIV>
  <DIV style=3D"FONT: 9pt =E5=AE=8B=E4=BD=93"><B>Subject:</B> Re: RE, =
and benefits of=20
  draft-jiang-l2vpn-vpls-mac-operation-00</DIV>
  <DIV><BR></DIV><!-- Converted from text/plain format -->
  <P><FONT size=3D2>Yanlong, my question is what is the difference =
between you=20
  draft and draft-pdutta-l2vpn-vpls-ldp-mac-opt-03? The function =
described are=20
  the same and I don't see the added value you bring with your=20
  draft.<BR><BR>Cheers,<BR>Wim<BR>_________________<BR>sent from=20
  blackberry<BR><BR>----- Original Message -----<BR>From: Jiang =
Yuan-long &lt;<A=20
  href=3D"mailto:yljiang@huawei.com">yljiang@huawei.com</A>&gt;<BR>To: =
HENDERICKX=20
  Wim<BR>Cc: <A href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> &lt;<A =

  href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A>&gt;<BR>Sent: Mon Aug =
11=20
  10:09:10 2008<BR>Subject: Re: RE, and benefits of=20
  draft-jiang-l2vpn-vpls-mac-operation-00<BR><BR>Hi, Wim:<BR>I am not =
sure what=20
  you mean by "This function is already proposed in&nbsp;=20
  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03."<BR>If you mean "MAC =
switching"=20
  behavior, forgive me if I am not aware of it in the above=20
  draft.<BR>Otherwise,if you mean the null-list of MACs,&nbsp; it is RFC =
4762=20
  which first introduces such a kind of sub-TLV.<BR><BR>As to the =
benefits of=20
  draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed them =
in the=20
  emails to this WG<BR>mailing list several days ago, I would be more =
than=20
  willing to re-sumarize them here:<BR><BR>1. Faster convergence of =
topology,=20
  maybe more precisely, faster MAC learning compared with MAC=20
  flooding.<BR>-Faster convergence means the time to associate the MACs =
with the=20
  "exact" out port<BR>will be shorter. Though learning is done in data =
plane at=20
  wire speed, the time for a frame to arrive<BR>at the end host and the =
response=20
  time for a host to send a reverse frame accross the PE (on which =
the<BR>MACs=20
  is flushed) will be a rather long time. So the learning process is not =
only=20
  determined solely by<BR>the wire speed, but also on the host =
processing time=20
  and delays on all the intermediate routers.<BR>For this reason, =
flooding as a=20
  result of MAC flooding may last for a rather long time.<BR><BR>With =
MAC=20
  switching, FIB can be established as fast as the MAC flushing =
operation=20
  itself:<BR>In selective flushing as described in RFC 4762 or=20
  draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,<BR>the MAC flushing operation =
may=20
  be:<BR>&nbsp;&nbsp;&nbsp; Given PW1, for each item in the=20
  =
FIB,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
; if=20
  the item's outport is not the same as this PW, then flush =
it.<BR><BR>In MAC=20
  switching, as described in draft-jiang-l2vpn-vpls-mac-operation-00, =
the MAC=20
  switching on the FIB table:<BR>&nbsp;&nbsp;&nbsp;&nbsp; Given PW1, =
PW2, for=20
  each item in the=20
  =
FIB,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
; if=20
  the item's outport is the same as PW1, then change it to =
PW2.<BR><BR>As the=20
  pseudocode shown above, MAC switching is no more complex than the =
selective=20
  MAC flushing.<BR>Actually, with more accurate PW information, it is =
quite=20
  possible that less PWs will be involved in the data plane =
lookup.<BR><BR>2.=20
  Less flooding burst in the core.<BR>For the customer sites with =
hundreds or=20
  more of MACs, a lot of MAC learning will be needed, and for the =
customer=20
  site<BR>with few MACs but large bandwidth, although few MACs are =
flushed, but=20
  as the time for the responsive frame to arrive<BR>at the PEs where =
MACs are=20
  flushed is not trivial, the total frames flooded accross the core may =
be added=20
  to a high volume.<BR><BR>Optimized selecive flooding as described in=20
  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate the MAC learning=20
  and<BR>flooding problem to some degree, but not so completely, and it =
applies=20
  only to a certain hub-and-spoke network topology,<BR>if CE is =
dual-homed to=20
  the PEs by ACs, it will be out of the capability of this draft. While=20
  draft-jiang-l2vpn-vpls-mac-operation-00<BR>tries to eliminate the MAC =
leaning=20
  and subsequent flooding procedures due to PE or PW switching =
completely, and=20
  in a broader way.<BR><BR>Regards<BR>Yuanlong<BR><BR><BR>Snippet of =
your last=20
  =
email:<BR><BR>________________________________<BR><BR>&nbsp;&nbsp;&nbsp;&=
nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; From: Jiang Yuan-long [<A=20
  =
href=3D"mailto:yljiang@huawei.com">mailto:yljiang@huawei.com</A>]<BR>&nbs=
p;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Sent: donderdag 7 augustus =
2008=20
  9:31<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; To: HENDERICKX=20
  Wim<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Cc:=20
  l2vpn@ietf.org<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Subject: Re:=20
  RE:<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  Wim:<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; In VPLS with dual-homed =
topology,=20
  we must pre-configure the working PE and backup PE to form a =
protection=20
  group,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; such as described in ICCP =
(defined=20
  in draft-martini-pwe3-iccp-00), and pre-configure working PWs with=20
  their<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; respective backup PWs, as =
described=20
  in "PW=20
  =
redundancy"(draft-ietf-pwe3-redundancy-bit-00).<BR>&nbsp;&nbsp;&nbsp;&nbs=
p;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Anyway, these =
configurations are=20
  prerequesite to provide sevice resiliency for VPLS, and the roles of =
the=20
  respective<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; PE or PW will change =
automatically=20
  after PE or PW switching (ie. backup PE becomes working PE, backup PW=20
  becomes<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; working=20
  PW).<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; The mechanism in my draft =
is based=20
  on such a protection relationship of PEs and PWs, I am not sure what=20
  other<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; configurations need to be =
changed,=20
  can you give me some more =
clues?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; We don;t need any of =
this=20
  for PW resilience in VPLS according to me as PW status together with =
the=20
  existing MAC flushes do the job for these=20
  topologies.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: But how to set the =
exact PW=20
  status? When a CE is dual-homed to PEs by attached circuits, and if =
the=20
  working AC fails,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; how do the PEs coordinate =
their=20
  PWs? IMHO, RFC 4762 is not perfect for these=20
  problems.&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Your draft explains =
a=20
  solution with an MTU-s attached with PW. For this the MTU-s decides =
the PW=20
  status bit. In case of AC attached you need something like ICCP, but =
this is=20
  independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) =
becomes=20
  active.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Wh&gt; The way it works is =
you use=20
  PW status as defined in the =
draft-muley-dutta-pwe3-redundancy-bit-02.txt and=20
  if the PW status changes you should perform a MAC flush in the full =
mesh VPLS=20
  domain.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As the above =
mentioned draft=20
  indicates, pre-configuration of PWs is necessary for its working. The =
only=20
  difference is whether<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; we should perform MAC =
flushing or=20
  MAC switching when we do PW=20
  switching.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; I think blindly flushing,=20
  selectively flushing and its optimizations (see Dutta's draft), and =
the=20
  flush-free optimizations as I=20
  proposed,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; should all be optional =
operations=20
  with their different applicability and=20
  efficiency.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; For the case of PW =
switching in=20
  VPLS with redundancy, flush-free optimizations is more favorable to=20
  me.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; RFC 4762 already provides =
some=20
  selective flushing mechansms, such as flushing some specified MACs or =
all MACs=20
  except<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; those associated with =
specific PWs.=20
  So what you mean may be that all the existing implementations only=20
  provide<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; flushing "all" MACs blindly =
in a=20
  VPLS instance, is that =
true?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; There are certain =
cases=20
  where we need to extend them like for PBB where we have C-MAC ties to=20
  B-MACs.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As far as I know, =
nowadays=20
  even some enterprise-oriented routers can selectively flush=20
  MACs.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Isn't the main =
reason for=20
  this the fact that they can't handle large floods of=20
  traffic?&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: Selective flushing =
is an=20
  optional operation, not the only choice. If turning down the selecitve =
MAC=20
  flushing<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; will improve the =
performance of a=20
  router, surely you can do =
that.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; As our anlysis and other=20
  sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly =

  flushing adversely<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; impacts all sites in a VPLS =

  instance, even their topology is not affected, the MACs must be =
re-leaned may=20
  be numerous,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; and the number of frames =
flooded=20
  may be not so small. Thus blindly flushing is unacceptable to a =
carrier grade=20
  VPLS<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; service,and for this reason =
we=20
  proposed some optimized MAC operations such as MAC Switching and MAC=20
  notification.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; The optimizations =
which are=20
  suggested in the mentioned drafts are very simple, whilst your =
proposal is=20
  fairly complex and depends on the topology you propose but might not =
be=20
  efficient in case we change the service=20
  topology.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: The draft mentioned =
above=20
  also do selective flushing, not blind flushing. And for what I know, =
this very=20
  draft may depend<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; on the hub-and-spoke =
network=20
  topology while our draft seems to be capable of a broader application =
scenario=20
  such as a CE dual-homed<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to PEs by ACs. The formats =
and=20
  contents of the LDP messages are similar in all these drafts or RFC =
for=20
  selective flushing, and =
as<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; my analysis goes, =
complexity (times=20
  of lookup and write) in dataplane may be commensurate for all these=20
  mechanisms.&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Right but the main=20
  difference is that you only need to indicate the PE or PW for which =
the MAC=20
  flush is performed. In your proposal you need to indicate the full MAC =
address=20
  list which will be very processing intensive if you have a large =
amount of MAC=20
  addresses in the VPLS service. The message will get very big and you =
will have=20
  to sent lots of them due to fragmentation=20
  implications.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was =
maybe the=20
  wrong wording) flush mechanism. The indication will be fast and the=20
  re-learning will happen instantly after the MAC is removed from the =
MAC=20
  FIB.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; You proposal will delay the =
trigger=20
  when the amount of MAC @ grows and on top the receiving PE needs a =
fast lookup=20
  mechanism to find he proper MAC and replace it which means delay the =
removal=20
  of the FIB.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As indicated in our =
draft,=20
  MAC Switching message can also carry only a null-list of MACs. It's OK =
to=20
  change all the MACs associated with a=20
  PW<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to another PW as=20
  egress.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; This function is =
already=20
  proposed in&nbsp; Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you =
think=20
  your solution is better?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; =

  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Best=20
  regards,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  Yuanlong<BR><BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
  =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR><BR></FONT></P></BLOC=
KQUOTE></BODY></HTML>

--Boundary_(ID_eAYkftLrk6ysS22ffCDydA)--


From l2vpn-bounces@ietf.org  Mon Aug 11 02:34:20 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 01E3F3A6E36;
	Mon, 11 Aug 2008 02:34:20 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id BB0D328C12D
	for <l2vpn@core3.amsl.com>; Mon, 11 Aug 2008 02:34:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.365
X-Spam-Level: **
X-Spam-Status: No, score=2.365 tagged_above=-999 required=5 tests=[AWL=-0.424, 
	BAYES_00=-2.599, FH_RELAY_NODNS=1.451, HELO_MISMATCH_COM=0.553, 
	HTML_FONT_FACE_BAD=0.884, HTML_MESSAGE=0.001,
	J_CHICKENPOX_31=0.6, 
	J_CHICKENPOX_73=0.6, J_CHICKENPOX_74=0.6, J_CHICKENPOX_92=0.6,
	RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id EDZ8dzouflBp for <l2vpn@core3.amsl.com>;
	Mon, 11 Aug 2008 02:34:16 -0700 (PDT)
Received: from szxga01-in.huawei.com (unknown [119.145.14.64])
	by core3.amsl.com (Postfix) with ESMTP id 8CC0E3A6E36
	for <l2vpn@ietf.org>; Mon, 11 Aug 2008 02:34:15 -0700 (PDT)
Received: from huawei.com (szxga01-in [172.24.2.3])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00333JWUFI@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 17:34:06 +0800 (CST)
Received: from huawei.com ([172.24.1.18])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K5F00NPPJWU2H@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 17:34:06 +0800 (CST)
Received: from j59929 ([10.70.77.69])
	by szxml03-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K5F00GSEJWTPQ@szxml03-in.huawei.com> for
	l2vpn@ietf.org; Mon, 11 Aug 2008 17:34:06 +0800 (CST)
Date: Mon, 11 Aug 2008 17:34:03 +0800
From: Jiang Yuan-long <yljiang@huawei.com>
Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
To: HENDERICKX Wim <wim.henderickx@alcatel-lucent.be>
Message-id: <000f01c8fb95$649d04d0$454d460a@china.huawei.com>
MIME-version: 1.0
X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2900.3198
X-Mailer: Microsoft Outlook Express 6.00.2900.3138
Content-type: multipart/alternative;
	boundary="Boundary_(ID_pNyHqK7gsEqF5smTrL7JWw)"
X-Priority: 3
X-MSMail-priority: Normal
References: <B128F666D4C8BD4FBF56CEAFB2DB66D7023E7A1C@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8fb8e$af9c6810$454d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D7030CEED0@FRVELSMBS22.ad2.ad.alcatel.com>
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

--Boundary_(ID_pNyHqK7gsEqF5smTrL7JWw)
Content-type: text/plain; charset=iso-8859-1
Content-transfer-encoding: 7BIT

Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00Surely switching is another way of optimization, and no need to flush.

  ----- Original Message ----- 
  From: HENDERICKX Wim 
  To: Jiang Yuan-long 
  Sent: Monday, August 11, 2008 4:55 PM
  Subject: RE: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00


  Indeed, the difference is in what you signal a PW or a PE-ID. My view is that draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 covers everything we need to optimize the flush behavior.

  Do you agree on that or what is missing according to you?



------------------------------------------------------------------------------
  From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
  Sent: maandag 11 augustus 2008 10:46
  To: HENDERICKX Wim
  Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00


  The difference is: 

  In draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, selectively flushing MACs by carrying a unique PE-ID.

  In draft-jiang-l2vpn-vpls-mac-operation-00, MAC switching by carrying PW info.

  Please see my last email for the benefits, if there is anything you don't agree with me, please 
  point it out and give your reasons.


    ----- Original Message ----- 
    From: HENDERICKX Wim 
    To: yljiang@huawei.com 
    Cc: l2vpn@ietf.org 
    Sent: Monday, August 11, 2008 4:18 PM
    Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00


    Yanlong, my question is what is the difference between you draft and draft-pdutta-l2vpn-vpls-ldp-mac-opt-03? The function described are the same and I don't see the added value you bring with your draft.

    Cheers,
    Wim
    _________________
    sent from blackberry

    ----- Original Message -----
    From: Jiang Yuan-long <yljiang@huawei.com>
    To: HENDERICKX Wim
    Cc: l2vpn@ietf.org <l2vpn@ietf.org>
    Sent: Mon Aug 11 10:09:10 2008
    Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00

    Hi, Wim:
    I am not sure what you mean by "This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03."
    If you mean "MAC switching" behavior, forgive me if I am not aware of it in the above draft.
    Otherwise,if you mean the null-list of MACs,  it is RFC 4762 which first introduces such a kind of sub-TLV.

    As to the benefits of draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed them in the emails to this WG
    mailing list several days ago, I would be more than willing to re-sumarize them here:

    1. Faster convergence of topology, maybe more precisely, faster MAC learning compared with MAC flooding.
    -Faster convergence means the time to associate the MACs with the "exact" out port
    will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive
    at the end host and the response time for a host to send a reverse frame accross the PE (on which the
    MACs is flushed) will be a rather long time. So the learning process is not only determined solely by
    the wire speed, but also on the host processing time and delays on all the intermediate routers.
    For this reason, flooding as a result of MAC flooding may last for a rather long time.

    With MAC switching, FIB can be established as fast as the MAC flushing operation itself:
    In selective flushing as described in RFC 4762 or draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,
    the MAC flushing operation may be:
        Given PW1, for each item in the FIB,
                if the item's outport is not the same as this PW, then flush it.

    In MAC switching, as described in draft-jiang-l2vpn-vpls-mac-operation-00, the MAC switching on the FIB table:
         Given PW1, PW2, for each item in the FIB,
                if the item's outport is the same as PW1, then change it to PW2.

    As the pseudocode shown above, MAC switching is no more complex than the selective MAC flushing.
    Actually, with more accurate PW information, it is quite possible that less PWs will be involved in the data plane lookup.

    2. Less flooding burst in the core.
    For the customer sites with hundreds or more of MACs, a lot of MAC learning will be needed, and for the customer site
    with few MACs but large bandwidth, although few MACs are flushed, but as the time for the responsive frame to arrive
    at the PEs where MACs are flushed is not trivial, the total frames flooded accross the core may be added to a high volume.

    Optimized selecive flooding as described in Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate the MAC learning and
    flooding problem to some degree, but not so completely, and it applies only to a certain hub-and-spoke network topology,
    if CE is dual-homed to the PEs by ACs, it will be out of the capability of this draft. While draft-jiang-l2vpn-vpls-mac-operation-00
    tries to eliminate the MAC leaning and subsequent flooding procedures due to PE or PW switching completely, and in a broader way.

    Regards
    Yuanlong


    Snippet of your last email:

    ________________________________

                            From: Jiang Yuan-long [mailto:yljiang@huawei.com]
                            Sent: donderdag 7 augustus 2008 9:31
                            To: HENDERICKX Wim
                            Cc: l2vpn@ietf.org
                            Subject: Re: RE:
                           
                           
                            Wim:
                            
                            In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group,
                            such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their
                            respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
                            Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective
                            PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes
                            working PW).
                            The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other
                            configurations need to be changed, can you give me some more clues?
                            
                            WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.
                            
                            Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails,
                            how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems.  
                            
                            WH> Your draft explains a solution with an MTU-s attached with PW. For this the MTU-s decides the PW status bit. In case of AC attached you need something like ICCP, but this is independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) becomes active.
                            
                            Wh> The way it works is you use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt and if the PW status changes you should perform a MAC flush in the full mesh VPLS domain.
                            
                            Jiang: As the above mentioned draft indicates, pre-configuration of PWs is necessary for its working. The only difference is whether
                            we should perform MAC flushing or MAC switching when we do PW switching.
                            I think blindly flushing, selectively flushing and its optimizations (see Dutta's draft), and the flush-free optimizations as I proposed,
                            should all be optional operations with their different applicability and efficiency.
                            For the case of PW switching in VPLS with redundancy, flush-free optimizations is more favorable to me.
                            
                            RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except
                            those associated with specific PWs. So what you mean may be that all the existing implementations only provide
                            flushing "all" MACs blindly in a VPLS instance, is that true?
                            
                            WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
                            Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs.
                            
                            WH> Isn't the main reason for this the fact that they can't handle large floods of traffic? 
                            
                            Jiang: Selective flushing is an optional operation, not the only choice. If turning down the selecitve MAC flushing
                            will improve the performance of a router, surely you can do that.
                            
                            As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely
                            impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
                            and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS
                            service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.
                            
                            WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
                           
                            Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend
                            on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed
                            to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as
                            my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms. 
                            
                            WH> Right but the main difference is that you only need to indicate the PE or PW for which the MAC flush is performed. In your proposal you need to indicate the full MAC address list which will be very processing intensive if you have a large amount of MAC addresses in the VPLS service. The message will get very big and you will have to sent lots of them due to fragmentation implications.
                            Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe the wrong wording) flush mechanism. The indication will be fast and the re-learning will happen instantly after the MAC is removed from the MAC FIB.
                            
                            You proposal will delay the trigger when the amount of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to find he proper MAC and replace it which means delay the removal of the FIB.
                            
                            Jiang: As indicated in our draft, MAC Switching message can also carry only a null-list of MACs. It's OK to change all the MACs associated with a PW
                            to another PW as egress.
                            
                            WH> This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think your solution is better?
                            
                           
                            Best regards,
                            
                            Yuanlong

                                    



--Boundary_(ID_pNyHqK7gsEqF5smTrL7JWw)
Content-type: text/html; charset=iso-8859-1
Content-transfer-encoding: 7BIT

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD><TITLE>Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00</TITLE>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV><FONT face=&#23435;&#20307; size=2>Surely&nbsp;switching is another way of optimization, 
and no need to flush.</FONT><BR></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Monday, August 11, 2008 4:55 PM</DIV>
  <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE: RE, and benefits of 
  draft-jiang-l2vpn-vpls-mac-operation-00</DIV>
  <DIV><BR></DIV>
  <DIV dir=ltr align=left><SPAN class=131155308-11082008><FONT size=2><FONT 
  face=Arial color=#0000ff>Indeed, the difference is in what you signal a PW or 
  a PE-ID. My view is that draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 covers 
  everything we need to optimize the flush behavior.</FONT></FONT></SPAN></DIV>
  <DIV dir=ltr align=left><SPAN class=131155308-11082008><FONT face=Arial 
  color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=ltr align=left><SPAN class=131155308-11082008><FONT face=Arial 
  color=#0000ff size=2>Do you agree on that or what is missing according to 
  you?</FONT></SPAN></DIV><BR>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> maandag 11 augustus 2008 
  10:46<BR><B>To:</B> HENDERICKX Wim<BR><B>Subject:</B> Re: RE, and benefits of 
  draft-jiang-l2vpn-vpls-mac-operation-00<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>The difference is: </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>In draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, 
  selectively <EM>flushing</EM> MACs by carrying a unique PE-ID.</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>In draft-jiang-l2vpn-vpls-mac-operation-00, MAC 
  <EM>switching</EM> by carrying PW info.</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>Please see my last email for the benefits, if there 
  is anything you don't agree with me, please </FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2>point it out and give your reasons.</FONT></DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <DIV><FONT face=&#23435;&#20307; size=2></FONT>&nbsp;</DIV>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style="FONT: 9pt &#23435;&#20307;">----- Original Message ----- </DIV>
    <DIV 
    style="BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; font-color: black"><B>From:</B> <A 
    title=wim.henderickx@alcatel-lucent.be 
    href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>To:</B> <A title=yljiang@huawei.com 
    href="mailto:yljiang@huawei.com">yljiang@huawei.com</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A title=l2vpn@ietf.org 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Monday, August 11, 2008 4:18 PM</DIV>
    <DIV style="FONT: 9pt &#23435;&#20307;"><B>Subject:</B> Re: RE, and benefits of 
    draft-jiang-l2vpn-vpls-mac-operation-00</DIV>
    <DIV><BR></DIV><!-- Converted from text/plain format -->
    <P><FONT size=2>Yanlong, my question is what is the difference between you 
    draft and draft-pdutta-l2vpn-vpls-ldp-mac-opt-03? The function described are 
    the same and I don't see the added value you bring with your 
    draft.<BR><BR>Cheers,<BR>Wim<BR>_________________<BR>sent from 
    blackberry<BR><BR>----- Original Message -----<BR>From: Jiang Yuan-long 
    &lt;<A href="mailto:yljiang@huawei.com">yljiang@huawei.com</A>&gt;<BR>To: 
    HENDERICKX Wim<BR>Cc: <A href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> 
    &lt;<A href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A>&gt;<BR>Sent: Mon Aug 
    11 10:09:10 2008<BR>Subject: Re: RE, and benefits of 
    draft-jiang-l2vpn-vpls-mac-operation-00<BR><BR>Hi, Wim:<BR>I am not sure 
    what you mean by "This function is already proposed in&nbsp; 
    Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03."<BR>If you mean "MAC switching" 
    behavior, forgive me if I am not aware of it in the above 
    draft.<BR>Otherwise,if you mean the null-list of MACs,&nbsp; it is RFC 4762 
    which first introduces such a kind of sub-TLV.<BR><BR>As to the benefits of 
    draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed them in the 
    emails to this WG<BR>mailing list several days ago, I would be more than 
    willing to re-sumarize them here:<BR><BR>1. Faster convergence of topology, 
    maybe more precisely, faster MAC learning compared with MAC 
    flooding.<BR>-Faster convergence means the time to associate the MACs with 
    the "exact" out port<BR>will be shorter. Though learning is done in data 
    plane at wire speed, the time for a frame to arrive<BR>at the end host and 
    the response time for a host to send a reverse frame accross the PE (on 
    which the<BR>MACs is flushed) will be a rather long time. So the learning 
    process is not only determined solely by<BR>the wire speed, but also on the 
    host processing time and delays on all the intermediate routers.<BR>For this 
    reason, flooding as a result of MAC flooding may last for a rather long 
    time.<BR><BR>With MAC switching, FIB can be established as fast as the MAC 
    flushing operation itself:<BR>In selective flushing as described in RFC 4762 
    or draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,<BR>the MAC flushing operation may 
    be:<BR>&nbsp;&nbsp;&nbsp; Given PW1, for each item in the 
    FIB,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    if the item's outport is not the same as this PW, then flush it.<BR><BR>In 
    MAC switching, as described in draft-jiang-l2vpn-vpls-mac-operation-00, the 
    MAC switching on the FIB table:<BR>&nbsp;&nbsp;&nbsp;&nbsp; Given PW1, PW2, 
    for each item in the 
    FIB,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    if the item's outport is the same as PW1, then change it to PW2.<BR><BR>As 
    the pseudocode shown above, MAC switching is no more complex than the 
    selective MAC flushing.<BR>Actually, with more accurate PW information, it 
    is quite possible that less PWs will be involved in the data plane 
    lookup.<BR><BR>2. Less flooding burst in the core.<BR>For the customer sites 
    with hundreds or more of MACs, a lot of MAC learning will be needed, and for 
    the customer site<BR>with few MACs but large bandwidth, although few MACs 
    are flushed, but as the time for the responsive frame to arrive<BR>at the 
    PEs where MACs are flushed is not trivial, the total frames flooded accross 
    the core may be added to a high volume.<BR><BR>Optimized selecive flooding 
    as described in Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate the MAC 
    learning and<BR>flooding problem to some degree, but not so completely, and 
    it applies only to a certain hub-and-spoke network topology,<BR>if CE is 
    dual-homed to the PEs by ACs, it will be out of the capability of this 
    draft. While draft-jiang-l2vpn-vpls-mac-operation-00<BR>tries to eliminate 
    the MAC leaning and subsequent flooding procedures due to PE or PW switching 
    completely, and in a broader 
    way.<BR><BR>Regards<BR>Yuanlong<BR><BR><BR>Snippet of your last 
    email:<BR><BR>________________________________<BR><BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; From: Jiang Yuan-long [<A 
    href="mailto:yljiang@huawei.com">mailto:yljiang@huawei.com</A>]<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Sent: donderdag 7 augustus 2008 
    9:31<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; To: HENDERICKX 
    Wim<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Cc: 
    l2vpn@ietf.org<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Subject: Re: 
    RE:<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    Wim:<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; In VPLS with dual-homed topology, 
    we must pre-configure the working PE and backup PE to form a protection 
    group,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; such as described in ICCP 
    (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with 
    their<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; respective backup PWs, as 
    described in "PW 
    redundancy"(draft-ietf-pwe3-redundancy-bit-00).<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Anyway, these configurations are 
    prerequesite to provide sevice resiliency for VPLS, and the roles of the 
    respective<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; PE or PW will change 
    automatically after PE or PW switching (ie. backup PE becomes working PE, 
    backup PW becomes<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; working 
    PW).<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; The mechanism in my draft is 
    based on such a protection relationship of PEs and PWs, I am not sure what 
    other<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; configurations need to be 
    changed, can you give me some more 
    clues?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; We don;t need any of this 
    for PW resilience in VPLS according to me as PW status together with the 
    existing MAC flushes do the job for these 
    topologies.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: But how to set the exact 
    PW status? When a CE is dual-homed to PEs by attached circuits, and if the 
    working AC fails,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; how do the PEs coordinate their 
    PWs? IMHO, RFC 4762 is not perfect for these 
    problems.&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Your draft explains a 
    solution with an MTU-s attached with PW. For this the MTU-s decides the PW 
    status bit. In case of AC attached you need something like ICCP, but this is 
    independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) 
    becomes active.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Wh&gt; The way it works is you 
    use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt 
    and if the PW status changes you should perform a MAC flush in the full mesh 
    VPLS domain.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As the above mentioned 
    draft indicates, pre-configuration of PWs is necessary for its working. The 
    only difference is whether<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; we should perform MAC flushing or 
    MAC switching when we do PW 
    switching.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; I think blindly flushing, 
    selectively flushing and its optimizations (see Dutta's draft), and the 
    flush-free optimizations as I 
    proposed,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; should all be optional operations 
    with their different applicability and 
    efficiency.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; For the case of PW switching in 
    VPLS with redundancy, flush-free optimizations is more favorable to 
    me.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; RFC 4762 already provides some 
    selective flushing mechansms, such as flushing some specified MACs or all 
    MACs except<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; those associated with specific 
    PWs. So what you mean may be that all the existing implementations only 
    provide<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; flushing "all" MACs blindly in a 
    VPLS instance, is that true?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; There are certain cases 
    where we need to extend them like for PBB where we have C-MAC ties to 
    B-MACs.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As far as I know, nowadays 
    even some enterprise-oriented routers can selectively flush 
    MACs.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Isn't the main reason for 
    this the fact that they can't handle large floods of 
    traffic?&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: Selective flushing is an 
    optional operation, not the only choice. If turning down the selecitve MAC 
    flushing<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; will improve the performance of a 
    router, surely you can do 
    that.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; As our anlysis and other 
    sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly 
    flushing adversely<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; impacts all sites in a VPLS 
    instance, even their topology is not affected, the MACs must be re-leaned 
    may be numerous,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; and the number of frames flooded 
    may be not so small. Thus blindly flushing is unacceptable to a carrier 
    grade VPLS<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; service,and for this reason we 
    proposed some optimized MAC operations such as MAC Switching and MAC 
    notification.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; The optimizations which 
    are suggested in the mentioned drafts are very simple, whilst your proposal 
    is fairly complex and depends on the topology you propose but might not be 
    efficient in case we change the service 
    topology.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: The draft mentioned above 
    also do selective flushing, not blind flushing. And for what I know, this 
    very draft may depend<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; on the hub-and-spoke network 
    topology while our draft seems to be capable of a broader application 
    scenario such as a CE 
    dual-homed<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to PEs by ACs. The formats and 
    contents of the LDP messages are similar in all these drafts or RFC for 
    selective flushing, and as<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; my analysis goes, complexity 
    (times of lookup and write) in dataplane may be commensurate for all these 
    mechanisms.&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Right but the main 
    difference is that you only need to indicate the PE or PW for which the MAC 
    flush is performed. In your proposal you need to indicate the full MAC 
    address list which will be very processing intensive if you have a large 
    amount of MAC addresses in the VPLS service. The message will get very big 
    and you will have to sent lots of them due to fragmentation 
    implications.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe 
    the wrong wording) flush mechanism. The indication will be fast and the 
    re-learning will happen instantly after the MAC is removed from the MAC 
    FIB.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; You proposal will delay the 
    trigger when the amount of MAC @ grows and on top the receiving PE needs a 
    fast lookup mechanism to find he proper MAC and replace it which means delay 
    the removal of the FIB.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As indicated in our draft, 
    MAC Switching message can also carry only a null-list of MACs. It's OK to 
    change all the MACs associated with a 
    PW<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to another PW as 
    egress.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; This function is already 
    proposed in&nbsp; Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think 
    your solution is better?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Best 
    regards,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    Yuanlong<BR><BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; 
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR><BR></FONT></P></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

--Boundary_(ID_pNyHqK7gsEqF5smTrL7JWw)--


From l2vpn-bounces@ietf.org  Mon Aug 11 05:15:39 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 469943A6A4D;
	Mon, 11 Aug 2008 05:15:39 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 659693A68FE
	for <l2vpn@core3.amsl.com>; Mon, 11 Aug 2008 05:15:38 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.316
X-Spam-Level: *
X-Spam-Status: No, score=1.316 tagged_above=-999 required=5 tests=[AWL=0.280, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35, HTML_FONT_FACE_BAD=0.884,
	HTML_MESSAGE=0.001, J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6,
	J_CHICKENPOX_74=0.6, J_CHICKENPOX_92=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id s217K+jCe+2f for <l2vpn@core3.amsl.com>;
	Mon, 11 Aug 2008 05:15:36 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [62.23.212.27])
	by core3.amsl.com (Postfix) with ESMTP id 8E9713A6922
	for <l2vpn@ietf.org>; Mon, 11 Aug 2008 05:15:35 -0700 (PDT)
Received: from FRVELSBHS05.ad2.ad.alcatel.com (frvelsbhs05.ad2.ad.alcatel.com
	[155.132.6.77])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7BCFOLj014027; 
	Mon, 11 Aug 2008 14:15:24 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS05.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Mon, 11 Aug 2008 14:15:24 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8FBAB.EE3FF1C6"
Subject: RE: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
Date: Mon, 11 Aug 2008 14:15:25 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D7030CEFAB@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <000f01c8fb95$649d04d0$454d460a@china.huawei.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00
Thread-Index: Acj7lf9lnUQvobgYSc2M9J73U72CtAAFZ2eA
References: <B128F666D4C8BD4FBF56CEAFB2DB66D7023E7A1C@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8fb8e$af9c6810$454d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D7030CEED0@FRVELSMBS22.ad2.ad.alcatel.com>
	<000f01c8fb95$649d04d0$454d460a@china.huawei.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Jiang Yuan-long" <yljiang@huawei.com>
X-OriginalArrivalTime: 11 Aug 2008 12:15:24.0356 (UTC)
	FILETIME=[EEBF6840:01C8FBAB]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8FBAB.EE3FF1C6
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Do you have proof about the optimization we get with MAC switching
versus MAC flushing?

________________________________

From: Jiang Yuan-long [mailto:yljiang@huawei.com]=20
Sent: maandag 11 augustus 2008 11:34
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: RE, and benefits of draft-jiang-l2vpn-vpls-mac-operation-00


Surely switching is another way of optimization, and no need to flush.


	----- Original Message -----=20
	From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be> =20
	To: Jiang Yuan-long <mailto:yljiang@huawei.com> =20
	Sent: Monday, August 11, 2008 4:55 PM
	Subject: RE: RE, and benefits of
draft-jiang-l2vpn-vpls-mac-operation-00

	Indeed, the difference is in what you signal a PW or a PE-ID. My
view is that draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 covers everything we
need to optimize the flush behavior.
	=20
	Do you agree on that or what is missing according to you?

________________________________

	From: Jiang Yuan-long [mailto:yljiang@huawei.com]=20
	Sent: maandag 11 augustus 2008 10:46
	To: HENDERICKX Wim
	Subject: Re: RE, and benefits of
draft-jiang-l2vpn-vpls-mac-operation-00
=09
=09
	The difference is:=20
	=20
	In draft-pdutta-l2vpn-vpls-ldp-mac-opt-03, selectively flushing
MACs by carrying a unique PE-ID.
	=20
	In draft-jiang-l2vpn-vpls-mac-operation-00, MAC switching by
carrying PW info.
	=20
	Please see my last email for the benefits, if there is anything
you don't agree with me, please=20
	point it out and give your reasons.
	=20
	=20

		----- Original Message -----=20
		From: HENDERICKX Wim
<mailto:wim.henderickx@alcatel-lucent.be> =20
		To: yljiang@huawei.com=20
		Cc: l2vpn@ietf.org=20
		Sent: Monday, August 11, 2008 4:18 PM
		Subject: Re: RE, and benefits of
draft-jiang-l2vpn-vpls-mac-operation-00


		Yanlong, my question is what is the difference between
you draft and draft-pdutta-l2vpn-vpls-ldp-mac-opt-03? The function
described are the same and I don't see the added value you bring with
your draft.
	=09
		Cheers,
		Wim
		_________________
		sent from blackberry
	=09
		----- Original Message -----
		From: Jiang Yuan-long <yljiang@huawei.com>
		To: HENDERICKX Wim
		Cc: l2vpn@ietf.org <l2vpn@ietf.org>
		Sent: Mon Aug 11 10:09:10 2008
		Subject: Re: RE, and benefits of
draft-jiang-l2vpn-vpls-mac-operation-00
	=09
		Hi, Wim:
		I am not sure what you mean by "This function is already
proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03."
		If you mean "MAC switching" behavior, forgive me if I am
not aware of it in the above draft.
		Otherwise,if you mean the null-list of MACs,  it is RFC
4762 which first introduces such a kind of sub-TLV.
	=09
		As to the benefits of
draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed them in
the emails to this WG
		mailing list several days ago, I would be more than
willing to re-sumarize them here:
	=09
		1. Faster convergence of topology, maybe more precisely,
faster MAC learning compared with MAC flooding.
		-Faster convergence means the time to associate the MACs
with the "exact" out port
		will be shorter. Though learning is done in data plane
at wire speed, the time for a frame to arrive
		at the end host and the response time for a host to send
a reverse frame accross the PE (on which the
		MACs is flushed) will be a rather long time. So the
learning process is not only determined solely by
		the wire speed, but also on the host processing time and
delays on all the intermediate routers.
		For this reason, flooding as a result of MAC flooding
may last for a rather long time.
	=09
		With MAC switching, FIB can be established as fast as
the MAC flushing operation itself:
		In selective flushing as described in RFC 4762 or
draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,
		the MAC flushing operation may be:
		    Given PW1, for each item in the FIB,
		            if the item's outport is not the same as
this PW, then flush it.
	=09
		In MAC switching, as described in
draft-jiang-l2vpn-vpls-mac-operation-00, the MAC switching on the FIB
table:
		     Given PW1, PW2, for each item in the FIB,
		            if the item's outport is the same as PW1,
then change it to PW2.
	=09
		As the pseudocode shown above, MAC switching is no more
complex than the selective MAC flushing.
		Actually, with more accurate PW information, it is quite
possible that less PWs will be involved in the data plane lookup.
	=09
		2. Less flooding burst in the core.
		For the customer sites with hundreds or more of MACs, a
lot of MAC learning will be needed, and for the customer site
		with few MACs but large bandwidth, although few MACs are
flushed, but as the time for the responsive frame to arrive
		at the PEs where MACs are flushed is not trivial, the
total frames flooded accross the core may be added to a high volume.
	=09
		Optimized selecive flooding as described in
Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate the MAC learning
and
		flooding problem to some degree, but not so completely,
and it applies only to a certain hub-and-spoke network topology,
		if CE is dual-homed to the PEs by ACs, it will be out of
the capability of this draft. While
draft-jiang-l2vpn-vpls-mac-operation-00
		tries to eliminate the MAC leaning and subsequent
flooding procedures due to PE or PW switching completely, and in a
broader way.
	=09
		Regards
		Yuanlong
	=09
	=09
		Snippet of your last email:
	=09
		________________________________
	=09
		                        From: Jiang Yuan-long
[mailto:yljiang@huawei.com]
		                        Sent: donderdag 7 augustus 2008
9:31
		                        To: HENDERICKX Wim
		                        Cc: l2vpn@ietf.org
		                        Subject: Re: RE:
		                      =20
		                      =20
		                        Wim:
		                       =20
		                        In VPLS with dual-homed
topology, we must pre-configure the working PE and backup PE to form a
protection group,
		                        such as described in ICCP
(defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs
with their
		                        respective backup PWs, as
described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
		                        Anyway, these configurations are
prerequesite to provide sevice resiliency for VPLS, and the roles of the
respective
		                        PE or PW will change
automatically after PE or PW switching (ie. backup PE becomes working
PE, backup PW becomes
		                        working PW).
		                        The mechanism in my draft is
based on such a protection relationship of PEs and PWs, I am not sure
what other
		                        configurations need to be
changed, can you give me some more clues?
		                       =20
		                        WH> We don;t need any of this
for PW resilience in VPLS according to me as PW status together with the
existing MAC flushes do the job for these topologies.
		                       =20
		                        Jiang: But how to set the exact
PW status? When a CE is dual-homed to PEs by attached circuits, and if
the working AC fails,
		                        how do the PEs coordinate their
PWs? IMHO, RFC 4762 is not perfect for these problems. =20
		                       =20
		                        WH> Your draft explains a
solution with an MTU-s attached with PW. For this the MTU-s decides the
PW status bit. In case of AC attached you need something like ICCP, but
this is independent of RFC 4762. Based on the ICCP mechanism one of the
AC(s) becomes active.
		                       =20
		                        Wh> The way it works is you use
PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt
and if the PW status changes you should perform a MAC flush in the full
mesh VPLS domain.
		                       =20
		                        Jiang: As the above mentioned
draft indicates, pre-configuration of PWs is necessary for its working.
The only difference is whether
		                        we should perform MAC flushing
or MAC switching when we do PW switching.
		                        I think blindly flushing,
selectively flushing and its optimizations (see Dutta's draft), and the
flush-free optimizations as I proposed,
		                        should all be optional
operations with their different applicability and efficiency.
		                        For the case of PW switching in
VPLS with redundancy, flush-free optimizations is more favorable to me.
		                       =20
		                        RFC 4762 already provides some
selective flushing mechansms, such as flushing some specified MACs or
all MACs except
		                        those associated with specific
PWs. So what you mean may be that all the existing implementations only
provide
		                        flushing "all" MACs blindly in a
VPLS instance, is that true?
		                       =20
		                        WH> There are certain cases
where we need to extend them like for PBB where we have C-MAC ties to
B-MACs.
		                        Jiang: As far as I know,
nowadays even some enterprise-oriented routers can selectively flush
MACs.
		                       =20
		                        WH> Isn't the main reason for
this the fact that they can't handle large floods of traffic?=20
		                       =20
		                        Jiang: Selective flushing is an
optional operation, not the only choice. If turning down the selecitve
MAC flushing
		                        will improve the performance of
a router, surely you can do that.
		                       =20
		                        As our anlysis and other
sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly
flushing adversely
		                        impacts all sites in a VPLS
instance, even their topology is not affected, the MACs must be
re-leaned may be numerous,
		                        and the number of frames flooded
may be not so small. Thus blindly flushing is unacceptable to a carrier
grade VPLS
		                        service,and for this reason we
proposed some optimized MAC operations such as MAC Switching and MAC
notification.
		                       =20
		                        WH> The optimizations which are
suggested in the mentioned drafts are very simple, whilst your proposal
is fairly complex and depends on the topology you propose but might not
be efficient in case we change the service topology.
		                      =20
		                        Jiang: The draft mentioned above
also do selective flushing, not blind flushing. And for what I know,
this very draft may depend
		                        on the hub-and-spoke network
topology while our draft seems to be capable of a broader application
scenario such as a CE dual-homed
		                        to PEs by ACs. The formats and
contents of the LDP messages are similar in all these drafts or RFC for
selective flushing, and as
		                        my analysis goes, complexity
(times of lookup and write) in dataplane may be commensurate for all
these mechanisms.=20
		                       =20
		                        WH> Right but the main
difference is that you only need to indicate the PE or PW for which the
MAC flush is performed. In your proposal you need to indicate the full
MAC address list which will be very processing intensive if you have a
large amount of MAC addresses in the VPLS service. The message will get
very big and you will have to sent lots of them due to fragmentation
implications.
=09
Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was
maybe the wrong wording) flush mechanism. The indication will be fast
and the re-learning will happen instantly after the MAC is removed from
the MAC FIB.
		                       =20
		                        You proposal will delay the
trigger when the amount of MAC @ grows and on top the receiving PE needs
a fast lookup mechanism to find he proper MAC and replace it which means
delay the removal of the FIB.
		                       =20
		                        Jiang: As indicated in our
draft, MAC Switching message can also carry only a null-list of MACs.
It's OK to change all the MACs associated with a PW
		                        to another PW as egress.
		                       =20
		                        WH> This function is already
proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think
your solution is better?
		                       =20
		                      =20
		                        Best regards,
		                       =20
		                        Yuanlong
	=09
		                               =20
	=09
	=09


------_=_NextPart_001_01C8FBAB.EE3FF1C6
Content-Type: text/html;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD><TITLE>Re: RE, and benefits of =
draft-jiang-l2vpn-vpls-mac-operation-00</TITLE>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Dus-ascii">
<META content=3D"MSHTML 6.00.2900.3354" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV dir=3Dltr align=3Dleft><SPAN class=3D315071312-11082008><FONT =
face=3DArial=20
color=3D#0000ff size=3D2>Do you have proof about the optimization we get =
with MAC=20
switching versus MAC flushing?</FONT></SPAN></DIV><BR>
<DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
<HR tabIndex=3D-1>
<FONT face=3DTahoma size=3D2><B>From:</B> Jiang Yuan-long=20
[mailto:yljiang@huawei.com] <BR><B>Sent:</B> maandag 11 augustus 2008=20
11:34<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B>=20
l2vpn@ietf.org<BR><B>Subject:</B> Re: RE, and benefits of=20
draft-jiang-l2vpn-vpls-mac-operation-00<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV><FONT face=3D&#23435;&#20307; size=3D2>Surely&nbsp;switching is =
another way of optimization,=20
and no need to flush.</FONT><BR></DIV>
<BLOCKQUOTE dir=3Dltr=20
style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; =
BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style=3D"FONT: 9pt &#23435;&#20307;">----- Original Message ----- =
</DIV>
  <DIV style=3D"BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; =
font-color: black"><B>From:</B>=20
  <A title=3Dwim.henderickx@alcatel-lucent.be=20
  href=3D"mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> =
</DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>To:</B> <A =
title=3Dyljiang@huawei.com=20
  href=3D"mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Monday, August =
11, 2008 4:55 PM</DIV>
  <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Subject:</B> RE: RE, and =
benefits of=20
  draft-jiang-l2vpn-vpls-mac-operation-00</DIV>
  <DIV><BR></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D131155308-11082008><FONT =
size=3D2><FONT=20
  face=3DArial color=3D#0000ff>Indeed, the difference is in what you =
signal a PW or=20
  a PE-ID. My view is that draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 covers =

  everything we need to optimize the flush =
behavior.</FONT></FONT></SPAN></DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D131155308-11082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2></FONT></SPAN>&nbsp;</DIV>
  <DIV dir=3Dltr align=3Dleft><SPAN class=3D131155308-11082008><FONT =
face=3DArial=20
  color=3D#0000ff size=3D2>Do you agree on that or what is missing =
according to=20
  you?</FONT></SPAN></DIV><BR>
  <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft>
  <HR tabIndex=3D-1>
  <FONT face=3DTahoma size=3D2><B>From:</B> Jiang Yuan-long=20
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> maandag 11 augustus 2008=20
  10:46<BR><B>To:</B> HENDERICKX Wim<BR><B>Subject:</B> Re: RE, and =
benefits of=20
  draft-jiang-l2vpn-vpls-mac-operation-00<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>The difference is: =
</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>In =
draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,=20
  selectively <EM>flushing</EM> MACs by carrying a unique =
PE-ID.</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>In =
draft-jiang-l2vpn-vpls-mac-operation-00, MAC=20
  <EM>switching</EM> by carrying PW info.</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>Please see my last email =
for the benefits, if there=20
  is anything you don't agree with me, please </FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2>point it out and give your =
reasons.</FONT></DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <DIV><FONT face=3D&#23435;&#20307; size=3D2></FONT>&nbsp;</DIV>
  <BLOCKQUOTE dir=3Dltr=20
  style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; =
BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style=3D"FONT: 9pt &#23435;&#20307;">----- Original Message =
----- </DIV>
    <DIV=20
    style=3D"BACKGROUND: #e4e4e4; FONT: 9pt &#23435;&#20307;; =
font-color: black"><B>From:</B> <A=20
    title=3Dwim.henderickx@alcatel-lucent.be=20
    href=3D"mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> =
</DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>To:</B> <A =
title=3Dyljiang@huawei.com=20
    href=3D"mailto:yljiang@huawei.com">yljiang@huawei.com</A> </DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Cc:</B> <A =
title=3Dl2vpn@ietf.org=20
    href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Sent:</B> Monday, =
August 11, 2008 4:18 PM</DIV>
    <DIV style=3D"FONT: 9pt &#23435;&#20307;"><B>Subject:</B> Re: RE, =
and benefits of=20
    draft-jiang-l2vpn-vpls-mac-operation-00</DIV>
    <DIV><BR></DIV><!-- Converted from text/plain format -->
    <P><FONT size=3D2>Yanlong, my question is what is the difference =
between you=20
    draft and draft-pdutta-l2vpn-vpls-ldp-mac-opt-03? The function =
described are=20
    the same and I don't see the added value you bring with your=20
    draft.<BR><BR>Cheers,<BR>Wim<BR>_________________<BR>sent from=20
    blackberry<BR><BR>----- Original Message -----<BR>From: Jiang =
Yuan-long=20
    &lt;<A =
href=3D"mailto:yljiang@huawei.com">yljiang@huawei.com</A>&gt;<BR>To:=20
    HENDERICKX Wim<BR>Cc: <A =
href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A>=20
    &lt;<A =
href=3D"mailto:l2vpn@ietf.org">l2vpn@ietf.org</A>&gt;<BR>Sent: Mon Aug=20
    11 10:09:10 2008<BR>Subject: Re: RE, and benefits of=20
    draft-jiang-l2vpn-vpls-mac-operation-00<BR><BR>Hi, Wim:<BR>I am not =
sure=20
    what you mean by "This function is already proposed in&nbsp;=20
    Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03."<BR>If you mean "MAC =
switching"=20
    behavior, forgive me if I am not aware of it in the above=20
    draft.<BR>Otherwise,if you mean the null-list of MACs,&nbsp; it is =
RFC 4762=20
    which first introduces such a kind of sub-TLV.<BR><BR>As to the =
benefits of=20
    draft-jiang-l2vpn-vpls-mac-operation-00, although I had analyzed =
them in the=20
    emails to this WG<BR>mailing list several days ago, I would be more =
than=20
    willing to re-sumarize them here:<BR><BR>1. Faster convergence of =
topology,=20
    maybe more precisely, faster MAC learning compared with MAC=20
    flooding.<BR>-Faster convergence means the time to associate the =
MACs with=20
    the "exact" out port<BR>will be shorter. Though learning is done in =
data=20
    plane at wire speed, the time for a frame to arrive<BR>at the end =
host and=20
    the response time for a host to send a reverse frame accross the PE =
(on=20
    which the<BR>MACs is flushed) will be a rather long time. So the =
learning=20
    process is not only determined solely by<BR>the wire speed, but also =
on the=20
    host processing time and delays on all the intermediate =
routers.<BR>For this=20
    reason, flooding as a result of MAC flooding may last for a rather =
long=20
    time.<BR><BR>With MAC switching, FIB can be established as fast as =
the MAC=20
    flushing operation itself:<BR>In selective flushing as described in =
RFC 4762=20
    or draft-pdutta-l2vpn-vpls-ldp-mac-opt-03,<BR>the MAC flushing =
operation may=20
    be:<BR>&nbsp;&nbsp;&nbsp; Given PW1, for each item in the=20
    =
FIB,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;=20
    if the item's outport is not the same as this PW, then flush =
it.<BR><BR>In=20
    MAC switching, as described in =
draft-jiang-l2vpn-vpls-mac-operation-00, the=20
    MAC switching on the FIB table:<BR>&nbsp;&nbsp;&nbsp;&nbsp; Given =
PW1, PW2,=20
    for each item in the=20
    =
FIB,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp=
;=20
    if the item's outport is the same as PW1, then change it to =
PW2.<BR><BR>As=20
    the pseudocode shown above, MAC switching is no more complex than =
the=20
    selective MAC flushing.<BR>Actually, with more accurate PW =
information, it=20
    is quite possible that less PWs will be involved in the data plane=20
    lookup.<BR><BR>2. Less flooding burst in the core.<BR>For the =
customer sites=20
    with hundreds or more of MACs, a lot of MAC learning will be needed, =
and for=20
    the customer site<BR>with few MACs but large bandwidth, although few =
MACs=20
    are flushed, but as the time for the responsive frame to =
arrive<BR>at the=20
    PEs where MACs are flushed is not trivial, the total frames flooded =
accross=20
    the core may be added to a high volume.<BR><BR>Optimized selecive =
flooding=20
    as described in Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 can alleviate =
the MAC=20
    learning and<BR>flooding problem to some degree, but not so =
completely, and=20
    it applies only to a certain hub-and-spoke network topology,<BR>if =
CE is=20
    dual-homed to the PEs by ACs, it will be out of the capability of =
this=20
    draft. While draft-jiang-l2vpn-vpls-mac-operation-00<BR>tries to =
eliminate=20
    the MAC leaning and subsequent flooding procedures due to PE or PW =
switching=20
    completely, and in a broader=20
    way.<BR><BR>Regards<BR>Yuanlong<BR><BR><BR>Snippet of your last=20
    =
email:<BR><BR>________________________________<BR><BR>&nbsp;&nbsp;&nbsp;&=
nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; From: Jiang Yuan-long [<A =

    =
href=3D"mailto:yljiang@huawei.com">mailto:yljiang@huawei.com</A>]<BR>&nbs=
p;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Sent: donderdag 7 =
augustus 2008=20
    9:31<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; To: HENDERICKX=20
    Wim<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Cc:=20
    l2vpn@ietf.org<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Subject: Re:=20
    RE:<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    Wim:<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; In VPLS with dual-homed =
topology,=20
    we must pre-configure the working PE and backup PE to form a =
protection=20
    group,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; such as described in ICCP =

    (defined in draft-martini-pwe3-iccp-00), and pre-configure working =
PWs with=20
    their<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; respective backup PWs, as =

    described in "PW=20
    =
redundancy"(draft-ietf-pwe3-redundancy-bit-00).<BR>&nbsp;&nbsp;&nbsp;&nbs=
p;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Anyway, these =
configurations are=20
    prerequesite to provide sevice resiliency for VPLS, and the roles of =
the=20
    respective<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; PE or PW will change=20
    automatically after PE or PW switching (ie. backup PE becomes =
working PE,=20
    backup PW becomes<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; working=20
    PW).<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; The mechanism in my draft =
is=20
    based on such a protection relationship of PEs and PWs, I am not =
sure what=20
    other<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; configurations need to be =

    changed, can you give me some more=20
    clues?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; We don;t need any =
of this=20
    for PW resilience in VPLS according to me as PW status together with =
the=20
    existing MAC flushes do the job for these=20
    topologies.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: But how to set the =
exact=20
    PW status? When a CE is dual-homed to PEs by attached circuits, and =
if the=20
    working AC fails,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; how do the PEs coordinate =
their=20
    PWs? IMHO, RFC 4762 is not perfect for these=20
    problems.&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Your draft =
explains a=20
    solution with an MTU-s attached with PW. For this the MTU-s decides =
the PW=20
    status bit. In case of AC attached you need something like ICCP, but =
this is=20
    independent of RFC 4762. Based on the ICCP mechanism one of the =
AC(s)=20
    becomes active.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Wh&gt; The way it works =
is you=20
    use PW status as defined in the =
draft-muley-dutta-pwe3-redundancy-bit-02.txt=20
    and if the PW status changes you should perform a MAC flush in the =
full mesh=20
    VPLS domain.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As the above =
mentioned=20
    draft indicates, pre-configuration of PWs is necessary for its =
working. The=20
    only difference is =
whether<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; we should perform MAC =
flushing or=20
    MAC switching when we do PW=20
    switching.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; I think blindly flushing, =

    selectively flushing and its optimizations (see Dutta's draft), and =
the=20
    flush-free optimizations as I=20
    proposed,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; should all be optional =
operations=20
    with their different applicability and=20
    efficiency.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; For the case of PW =
switching in=20
    VPLS with redundancy, flush-free optimizations is more favorable to=20
    me.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; RFC 4762 already provides =
some=20
    selective flushing mechansms, such as flushing some specified MACs =
or all=20
    MACs except<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; those associated with =
specific=20
    PWs. So what you mean may be that all the existing implementations =
only=20
    provide<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; flushing "all" MACs =
blindly in a=20
    VPLS instance, is that =
true?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; There are certain =
cases=20
    where we need to extend them like for PBB where we have C-MAC ties =
to=20
    B-MACs.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As far as I know, =
nowadays=20
    even some enterprise-oriented routers can selectively flush=20
    MACs.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Isn't the main =
reason for=20
    this the fact that they can't handle large floods of=20
    traffic?&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: Selective flushing =
is an=20
    optional operation, not the only choice. If turning down the =
selecitve MAC=20
    flushing<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; will improve the =
performance of a=20
    router, surely you can do=20
    that.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; As our anlysis and other=20
    sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, =
blindly=20
    flushing adversely<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; impacts all sites in a =
VPLS=20
    instance, even their topology is not affected, the MACs must be =
re-leaned=20
    may be numerous,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; and the number of frames =
flooded=20
    may be not so small. Thus blindly flushing is unacceptable to a =
carrier=20
    grade VPLS<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; service,and for this =
reason we=20
    proposed some optimized MAC operations such as MAC Switching and MAC =

    notification.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; The optimizations =
which=20
    are suggested in the mentioned drafts are very simple, whilst your =
proposal=20
    is fairly complex and depends on the topology you propose but might =
not be=20
    efficient in case we change the service=20
    topology.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: The draft =
mentioned above=20
    also do selective flushing, not blind flushing. And for what I know, =
this=20
    very draft may depend<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; on the hub-and-spoke =
network=20
    topology while our draft seems to be capable of a broader =
application=20
    scenario such as a CE=20
    dual-homed<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to PEs by ACs. The =
formats and=20
    contents of the LDP messages are similar in all these drafts or RFC =
for=20
    selective flushing, and =
as<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; my analysis goes, =
complexity=20
    (times of lookup and write) in dataplane may be commensurate for all =
these=20
    mechanisms.&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; Right but the main =

    difference is that you only need to indicate the PE or PW for which =
the MAC=20
    flush is performed. In your proposal you need to indicate the full =
MAC=20
    address list which will be very processing intensive if you have a =
large=20
    amount of MAC addresses in the VPLS service. The message will get =
very big=20
    and you will have to sent lots of them due to fragmentation=20
    implications.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was =
maybe=20
    the wrong wording) flush mechanism. The indication will be fast and =
the=20
    re-learning will happen instantly after the MAC is removed from the =
MAC=20
    FIB.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; You proposal will delay =
the=20
    trigger when the amount of MAC @ grows and on top the receiving PE =
needs a=20
    fast lookup mechanism to find he proper MAC and replace it which =
means delay=20
    the removal of the =
FIB.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Jiang: As indicated in =
our draft,=20
    MAC Switching message can also carry only a null-list of MACs. It's =
OK to=20
    change all the MACs associated with a=20
    PW<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; to another PW as=20
    egress.<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; WH&gt; This function is =
already=20
    proposed in&nbsp; Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you =
think=20
    your solution is =
better?<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Best=20
    regards,<BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR>&nbsp;&nbsp;&nbsp;&nb=
sp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    Yuanlong<BR><BR>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=20
    =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;<BR><BR></FONT></P></BLOC=
KQUOTE></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C8FBAB.EE3FF1C6--


From l2vpn-bounces@ietf.org  Mon Aug 11 07:06:20 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id EC2203A6C95;
	Mon, 11 Aug 2008 07:06:19 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 2B7FF3A6768
	for <l2vpn@core3.amsl.com>; Sun, 10 Aug 2008 23:21:40 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.666
X-Spam-Level: *
X-Spam-Status: No, score=1.666 tagged_above=-999 required=5 tests=[AWL=1.230, 
	BAYES_00=-2.599, HELO_EQ_FR=0.35, HTML_FONT_FACE_BAD=0.884,
	HTML_MESSAGE=0.001, J_CHICKENPOX_31=0.6, J_CHICKENPOX_73=0.6,
	J_CHICKENPOX_74=0.6]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id L2+ZCcLhIntd for <l2vpn@core3.amsl.com>;
	Sun, 10 Aug 2008 23:21:38 -0700 (PDT)
Received: from smail5.alcatel.fr (smail5.alcatel.fr [64.208.49.27])
	by core3.amsl.com (Postfix) with ESMTP id B2AFE3A67EA
	for <l2vpn@ietf.org>; Sun, 10 Aug 2008 23:21:34 -0700 (PDT)
Received: from FRVELSBHS05.ad2.ad.alcatel.com (frvelsbhs05.ad2.ad.alcatel.com
	[155.132.6.77])
	by smail5.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7B6LPxq022824; 
	Mon, 11 Aug 2008 08:21:25 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.51]) by
	FRVELSBHS05.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Mon, 11 Aug 2008 08:21:25 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----_=_NextPart_001_01C8FB7A.7B0EE58A"
Subject: RE: RE:
Date: Mon, 11 Aug 2008 08:21:27 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70302CBE2@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <006101c8fb59$bd4a6690$454d460a@china.huawei.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: RE:
Thread-Index: Acj7WcWlPAlpHS3nTeqU7G0yEuUP1AAIAVzw
References: <001c01c8f7a6$fa221ab0$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C5FC@FRVELSMBS22.ad2.ad.alcatel.com>
	<006c01c8f842$1347c620$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C624@FRVELSMBS22.ad2.ad.alcatel.com>
	<007d01c8f85f$99578350$5d4d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C877@FRVELSMBS22.ad2.ad.alcatel.com>
	<007301c8f935$ea290980$454d460a@china.huawei.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302CB8F@FRVELSMBS22.ad2.ad.alcatel.com>
	<006101c8fb59$bd4a6690$454d460a@china.huawei.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Jiang Yuan-long" <yljiang@huawei.com>
X-OriginalArrivalTime: 11 Aug 2008 06:21:25.0593 (UTC)
	FILETIME=[7B755C90:01C8FB7A]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.13
X-Mailman-Approved-At: Mon, 11 Aug 2008 07:06:19 -0700
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

This is a multi-part message in MIME format.

------_=_NextPart_001_01C8FB7A.7B0EE58A
Content-Type: text/plain;
	charset="iso-2022-jp"
Content-Transfer-Encoding: 7bit

Yuanlong, more in-line.

________________________________

From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
Sent: maandag 11 augustus 2008 4:27
To: HENDERICKX Wim
Cc: l2vpn@ietf.org
Subject: Re: RE:


Wim, please see more in-line.
 
Cheers
Yuanlong

	----- Original Message ----- 
	From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
	To: Jiang Yuan-long <mailto:yljiang@huawei.com>  
	Cc: l2vpn@ietf.org 
	Sent: Sunday, August 10, 2008 2:57 PM
	Subject: RE: RE:

	Yuanlong, more in-line
	
	
________________________________

	From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
	Sent: vrijdag 8 augustus 2008 11:06
	To: HENDERICKX Wim
	Cc: l2vpn@ietf.org
	Subject: Re: RE:
	
	
	Hi$B!$(JWim$B!$(JPlease see my comments in line.

		----- Original Message ----- 
		From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
		To: Jiang Yuan-long <mailto:yljiang@huawei.com>  
		Cc: l2vpn@ietf.org 
		Sent: Friday, August 08, 2008 3:39 AM
		Subject: RE: RE:
		
		
		
		Yuanlong, see in-line

________________________________

		From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
		Sent: donderdag 7 augustus 2008 9:31
		To: HENDERICKX Wim
		Cc: l2vpn@ietf.org
		Subject: Re: RE:
		
		
		Wim:
		 
		In VPLS with dual-homed topology, we must pre-configure the working PE and backup PE to form a protection group, 
		such as described in ICCP (defined in draft-martini-pwe3-iccp-00), and pre-configure working PWs with their 
		respective backup PWs, as described in "PW redundancy"(draft-ietf-pwe3-redundancy-bit-00).
		Anyway, these configurations are prerequesite to provide sevice resiliency for VPLS, and the roles of the respective 
		PE or PW will change automatically after PE or PW switching (ie. backup PE becomes working PE, backup PW becomes 
		working PW).
		The mechanism in my draft is based on such a protection relationship of PEs and PWs, I am not sure what other 
		configurations need to be changed, can you give me some more clues?
		 
		WH> We don;t need any of this for PW resilience in VPLS according to me as PW status together with the existing MAC flushes do the job for these topologies.
		 
		Jiang: But how to set the exact PW status? When a CE is dual-homed to PEs by attached circuits, and if the working AC fails, 
		how do the PEs coordinate their PWs? IMHO, RFC 4762 is not perfect for these problems.   
		 
		WH> Your draft explains a solution with an MTU-s attached with PW. For this the MTU-s decides the PW status bit. In case of AC attached you need something like ICCP, but this is independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) becomes active.
		 
		Wh> The way it works is you use PW status as defined in the draft-muley-dutta-pwe3-redundancy-bit-02.txt and if the PW status changes you should perform a MAC flush in the full mesh VPLS domain.
		 
		Jiang: As the above mentioned draft indicates, pre-configuration of PWs is necessary for its working. The only difference is whether 
		we should perform MAC flushing or MAC switching when we do PW switching.
		I think blindly flushing, selectively flushing and its optimizations (see Dutta's draft), and the flush-free optimizations as I proposed, 
		should all be optional operations with their different applicability and efficiency. 
		For the case of PW switching in VPLS with redundancy, flush-free optimizations is more favorable to me.
		 
		RFC 4762 already provides some selective flushing mechansms, such as flushing some specified MACs or all MACs except 
		those associated with specific PWs. So what you mean may be that all the existing implementations only provide 
		flushing "all" MACs blindly in a VPLS instance, is that true? 
		 
		WH> There are certain cases where we need to extend them like for PBB where we have C-MAC ties to B-MACs.
		Jiang: As far as I know, nowadays even some enterprise-oriented routers can selectively flush MACs. 
		 
		WH> Isn't the main reason for this the fact that they can't handle large floods of traffic?  
		 
		Jiang: Selective flushing is an optional operation, not the only choice. If turning down the selecitve MAC flushing
		will improve the performance of a router, surely you can do that.
		 
		As our anlysis and other sources(such as draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
		impacts all sites in a VPLS instance, even their topology is not affected, the MACs must be re-leaned may be numerous,
		and the number of frames flooded may be not so small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
		service,and for this reason we proposed some optimized MAC operations such as MAC Switching and MAC notification.
		 
		WH> The optimizations which are suggested in the mentioned drafts are very simple, whilst your proposal is fairly complex and depends on the topology you propose but might not be efficient in case we change the service topology.
		

		Jiang: The draft mentioned above also do selective flushing, not blind flushing. And for what I know, this very draft may depend 
		on the hub-and-spoke network topology while our draft seems to be capable of a broader application scenario such as a CE dual-homed 
		to PEs by ACs. The formats and contents of the LDP messages are similar in all these drafts or RFC for selective flushing, and as 
		my analysis goes, complexity (times of lookup and write) in dataplane may be commensurate for all these mechanisms.  
		 
		WH> Right but the main difference is that you only need to indicate the PE or PW for which the MAC flush is performed. In your proposal you need to indicate the full MAC address list which will be very processing intensive if you have a large amount of MAC addresses in the VPLS service. The message will get very big and you will have to sent lots of them due to fragmentation implications. 
		Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides a simple (blind was maybe the wrong wording) flush mechanism. The indication will be fast and the re-learning will happen instantly after the MAC is removed from the MAC FIB.
		 
		You proposal will delay the trigger when the amount of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to find he proper MAC and replace it which means delay the removal of the FIB.
		 
		Jiang: As indicated in our draft, MAC Switching message can also carry only a null-list of MACs. It's OK to change all the MACs associated with a PW 
		to another PW as egress. 
		 
		WH> This function is already proposed in  Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think your solution is better?
		 
		
		Best regards,
		 
		Yuanlong

			----- Original Message ----- 
			From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
			To: Jiang Yuan-long <mailto:yljiang@huawei.com>  
			Cc: l2vpn@ietf.org 
			Sent: Thursday, August 07, 2008 1:24 PM
			Subject: RE: RE:
			
			
			
			Yuanlong,
			 
			My point is that how are you going to define the behavior on each PE when the topology is adapted. I hope you are not going to change the configuration on each PE in the VPLS each time.
			 
			On top, by doing selective MAC flush you need to go one by one to the items and hence the processing to withdraw the MAC(s) will be longer. When you use the existing mechanisms you can flush without the extra lookups and hence the performance will be faster. FLushing a MAC will result in re-learning which means you do not have traffic impact, just a small flooding until the MAC address is re-learned.
			 
			Cheers,
			Wim
			
________________________________

			From: Jiang Yuan-long [mailto:yljiang@huawei.com] 
			Sent: donderdag 7 augustus 2008 6:00
			To: HENDERICKX Wim
			Cc: l2vpn@ietf.org
			Subject: Re: RE:
			
			
			 
			Wim:
			 
			Thanks for you comment. My draft is not a standlone protocol, it can be deployed with other protocol suites.
			For what I know, "PW redundancy" is being standardized in PWE3 WG, which can synchronize the behavior of
			the PEs and PWs in PW switching. So it is not such a difficult task to do PW switching, and do MAC switching
			subsequently. 
			 
			The MAC withdrawal process defined in RFC 4762 is not so simple as you think, it maybe not just "flush all MACs" 
			blindly, but can selectively flush MACs, which computation comlexity is on the same level as my draft.

				----- Original Message ----- 
				From: HENDERICKX Wim <mailto:wim.henderickx@alcatel-lucent.be>  
				To: Jiang Yuan-long <mailto:yljiang@huawei.com>  ; sajassi@cisco.com 
				Cc: l2vpn@ietf.org 
				Sent: Thursday, August 07, 2008 4:18 AM
				Subject: RE:

				Yuanlong,
				 
				The difficulty with your draft is that it requires a PE to do different things depending in which situation the PE is in, and this makes it very cumbersome in case changes happen to topologies or configurations.
				 
				My view is that we need a simple MAC flush which we have today in LDP.
				 
				Cheers,
				WIm

________________________________

				From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf Of Jiang Yuan-long
				Sent: woensdag 6 augustus 2008 11:30
				To: sajassi@cisco.com
				Cc: l2vpn@ietf.org
				Subject: 
				
				
				Hi, Ali:
				 
				Thanks for your comments in the last L2VPN WG meeting, I will try to give you some reasons here.
				Following are the two questions:
				 
				1. What do you think will be faster convergence? When you flush MAC addresses, 
				the learning is done in data plane at wire speed. Why does that prolong convergence time?
				
				 
				-Faster convergence means the time to associate the MACs with the "exact" out port
				will be shorter. Though learning is done in data plane at wire speed, the time for a frame to arrive 
				at the end host and the response time for a host to send a reverse frame accross the PE (on which the 
				MACs is flushed) will be a rather long time. So the learning process is not only determined solely by 
				the wire speed, but also on the host processing time and delays on all the intermediate routers.
				For this reason, flooding may also last for a rather long time.
				 
				 
				2. You consider a simple case with simple backup for primary. In cases where primary fails and traffic gets
				 distributed among multiple PE's, how do you take care of that?
				 
				-I believe the most commonly deployed scenario is one backup PE for a primary PE.
				In my opinion, for the scenario of multiple backup PEs protecting a primary PE, 
				only one PE out of them will be elected as the working PE when the primary PE failed, 
				rather than distribute the traffic over multiple backup PEs,
				otherwise, loops may be introduced to the attached customer site.
				BTW, in a similar way, for PW redundancy mechanism, only 1 secondary is selected as the working PW (see 
				draft-ietf-pwe3-redundancy-bit-00.txt in PWE3 WG)
				Anyway, if there is any such requirements confirmed, I think it will be OK for us to find some solutions and 
				add them to the future drafts. 
				 
				cheers,
				Yuanlong
				


------_=_NextPart_001_01C8FB7A.7B0EE58A
Content-Type: text/html;
	charset="iso-2022-jp"
Content-Transfer-Encoding: 7bit

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-2022-jp">
<META content="MSHTML 6.00.2900.3354" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV dir=ltr align=left><FONT face=Arial><FONT color=#000080>Yuanlong<SPAN 
class=887291606-11082008>, more in-line.</SPAN></FONT></FONT></DIV><BR>
<DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
<HR tabIndex=-1>
<FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
[mailto:yljiang@huawei.com] <BR><B>Sent:</B> maandag 11 augustus 2008 
4:27<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV><FONT face="Times New Roman" color=#000080>Wim, please see more 
in-line.</FONT></DIV>
<DIV><FONT face="Times New Roman" color=#000080></FONT>&nbsp;</DIV>
<DIV><FONT face="Times New Roman" color=#000080>Cheers</FONT></DIV>
<DIV><FONT face="Times New Roman" color=#000080>Yuanlong</FONT></DIV>
<BLOCKQUOTE dir=ltr 
style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
  <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
  <DIV style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
  <A title=wim.henderickx@alcatel-lucent.be 
  href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
  href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Sunday, August 10, 2008 2:57 PM</DIV>
  <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
  <DIV><BR></DIV>
  <DIV dir=ltr align=left><FONT face=SimSun><FONT color=#0000ff><FONT 
  size=2>Yuanlong<SPAN class=289373519-07082008>,&nbsp;<SPAN 
  class=012333915-08082008>more</SPAN> 
  in-line</SPAN></FONT></FONT></FONT></DIV><FONT face=Arial color=#0000ff 
  size=2></FONT><BR>
  <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
  <HR tabIndex=-1>
  <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
  [mailto:yljiang@huawei.com] <BR><B>Sent:</B> vrijdag 8 augustus 2008 
  11:06<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
  href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
  RE:<BR></FONT><BR></DIV>
  <DIV></DIV>
  <DIV><FONT face=$BAWBN(B size=2>Hi$B!$(BWim$B!$(BPlease see my comments in line.</FONT></DIV>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
    <DIV 
    style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> <A 
    title=wim.henderickx@alcatel-lucent.be 
    href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
    href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Friday, August 08, 2008 3:39 AM</DIV>
    <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT><BR></DIV>
    <DIV dir=ltr align=left><FONT face=Arial>
    <DIV><FONT face=SimSun><FONT color=#0000ff><FONT size=2>Yuanlong<SPAN 
    class=289373519-07082008>, see 
    in-line</SPAN></FONT></FONT></FONT></DIV></FONT></DIV><BR>
    <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
    <HR tabIndex=-1>
    <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
    [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
    9:31<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> <A 
    href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> Re: 
    RE:<BR></FONT><BR></DIV>
    <DIV></DIV>
    <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>In VPLS with dual-homed topology, we 
    must&nbsp;pre-configure the working PE and backup PE to form a protection 
    group,&nbsp;</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>such as </FONT><FONT face=$BAWBN(B size=2>described in 
    ICCP (defined in draft-martini-pwe3-iccp-00),</FONT>&nbsp;<FONT face=$BAWBN(B 
    size=2>and pre-configure working PWs with their </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>respective backup PWs, </FONT><FONT face=$BAWBN(B 
    size=2>as </FONT><FONT face=$BAWBN(B size=2>described in "PW 
    redundancy"(draft-ietf-pwe3-redundancy-bit-00).</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>Anyway, these configurations are prerequesite to 
    provide sevice resiliency for VPLS, and the roles of the respective 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>PE or PW will </FONT><FONT face=$BAWBN(B size=2>change 
    automatically after PE or PW switching (ie. backup PE becomes working PE, 
    backup PW becomes </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>working PW).</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>The mechanism in&nbsp;my draft&nbsp;is based on 
    such a&nbsp;protection relationship of PEs and PWs,&nbsp;I am not sure what 
    other </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>configurations need to </FONT><FONT face=$BAWBN(B 
    size=2>be changed, can you give me some more clues?</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=SimSun size=2><SPAN class=289373519-07082008>WH&gt; We don;t 
    need any of this for PW resilience in VPLS according to me as PW status 
    together with the existing MAC flushes do the job for these 
    topologies.</SPAN></FONT></DIV>
    <DIV><FONT face=SimSun size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: But how to set 
    the&nbsp;exact&nbsp;PW status?&nbsp;When a CE is dual-homed to PEs by 
    attached circuits, and if the working AC fails, </FONT></DIV>
    <DIV><FONT face=$BAWBN(B><FONT size=2><FONT color=#800000>how do the PEs 
    coordinate their PWs?&nbsp;IMHO, RFC 4762&nbsp;is not perfect for these 
    problems.&nbsp;</FONT><FONT color=#0000ff><SPAN 
    class=012333915-08082008><FONT face=Arial>&nbsp;<SPAN 
    class=887291606-11082008>&nbsp;</SPAN></FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT face=$BAWBN(B><FONT size=2><FONT color=#0000ff><SPAN 
    class=012333915-08082008><FONT face=Arial><SPAN 
    class=887291606-11082008></SPAN></FONT></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B><FONT size=2><FONT color=#0000ff><SPAN 
    class=012333915-08082008><FONT face=Arial><SPAN 
    class=887291606-11082008>WH&gt;&nbsp;Your draft explains&nbsp;a solution 
    with an MTU-s attached with PW.&nbsp;For this&nbsp;the MTU-s decides the PW 
    status bit. In case of AC attached you need something like ICCP, but this is 
    independent of RFC 4762. Based on the ICCP mechanism one of the AC(s) 
    becomes active.</SPAN></FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
    class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT size=+0><SPAN 
    class=012333915-08082008><FONT color=#0000ff size=3>Wh&gt; The way it works 
    is you use PW status as defined in 
    the&nbsp;<STRONG>draft</STRONG>-muley-dutta-<B>pw</B>e3-redundancy-bit-02.txt 
    and if the PW status changes you should perform a MAC flush in the full mesh 
    VPLS domain.</FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#0000ff size=+0><SPAN 
    class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>Jiang: As the above mentioned draft indicates, 
    pre-configuration of PWs is necessary for its working. The only difference 
    is whether </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>we should perform MAC flushing or MAC switching 
    when we do PW switching.</SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>I think blindly flushing, selectively flushing and 
    its optimizations (see Dutta's draft), and the flush-free optimizations as I 
    proposed, </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>should all be optional operations with&nbsp;their 
    different applicability and efficiency. </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>For the case of PW switching in VPLS with 
    redundancy, flush-free optimizations is more favorable to 
    me.</SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>RFC 4762 already provides some selective flushing 
    mechansms, such as flushing some specified MACs or all MACs except 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>those associated with </FONT><FONT face=$BAWBN(B 
    size=2>specific PWs. So what</FONT><FONT face=$BAWBN(B size=2> you mean may be 
    that all the existing implementations only&nbsp;provide </FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B>flushing </FONT><FONT face=$BAWBN(B>"all" MACs 
    blindly in a VPLS instance, </FONT><FONT face=$BAWBN(B>is that true?<SPAN 
    class=289373519-07082008><FONT face=Arial 
    color=#0000ff>&nbsp;</FONT></SPAN></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN 
    class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><SPAN class=289373519-07082008>WH&gt; There 
    are certain cases where we need to extend them like for PBB where we have 
    C-MAC ties to B-MACs.</SPAN></FONT></FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    size=2>Jiang: As far as I know, nowadays even some enterprise-oriented 
    routers can selectively flush MACs.<SPAN class=012333915-08082008><FONT 
    face=Arial color=#0000ff>&nbsp;</FONT></SPAN></FONT></SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    size=2><SPAN 
    class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    size=2><SPAN class=012333915-08082008><FONT face=Arial 
    color=#0000ff>WH&gt;&nbsp;Isn't the main reason for this the fact that they 
    can't handle large floods of 
    traffic?&nbsp;</FONT>&nbsp;</SPAN></FONT></SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000><SPAN class=289373519-07082008><FONT 
    face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008></SPAN></FONT></SPAN></FONT>&nbsp;</DIV>
    <DIV><SPAN class=289373519-07082008><FONT face="Times New Roman" 
    color=#000080><SPAN class=012333915-08082008><SPAN 
    class=012333915-08082008>Jiang: Selective flushing is an optional operation, 
    not the only choice. If turning down the selecitve MAC 
    flushing</SPAN></SPAN></FONT></SPAN></DIV>
    <DIV><SPAN class=289373519-07082008><FONT face="Times New Roman" 
    color=#000080><SPAN class=012333915-08082008><SPAN 
    class=012333915-08082008>will improve the performance of a router, surely 
    you can do that.</SPAN></SPAN></FONT></SPAN></DIV>
    <DIV><FONT face=$BAWBN(B><FONT color=#000080><SPAN class=289373519-07082008><SPAN 
    class=012333915-08082008></SPAN></SPAN><SPAN 
    class=289373519-07082008></SPAN></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>As our anlysis and other sources(such as 
    draft-pdutta-l2vpn-vpls-ldp-mac-opt-03) shows, blindly flushing adversely 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>impacts all sites in a VPLS instance, even their 
    topology is not </FONT><FONT face=$BAWBN(B size=2>affected, the MACs must be 
    re-leaned may be numerous,</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>and the number of frames flooded may be not so 
    small. Thus blindly flushing is unacceptable to a carrier grade VPLS 
    </FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2>service,</FONT><FONT face=$BAWBN(B size=2>and&nbsp;for 
    this&nbsp;reason&nbsp;</FONT><FONT face=$BAWBN(B size=2>we proposed some optimized 
    MAC operations such as MAC </FONT><FONT face=$BAWBN(B size=2>Switching and MAC 
    notification.</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><SPAN class=289373519-07082008><FONT face=SimSun size=2>WH&gt; The 
    optimizations which are suggested in the mentioned drafts are very simple, 
    whilst your proposal is fairly complex and depends on the topology you 
    propose but might not be efficient in case we change the service 
    topology.</FONT></SPAN></DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=SimSun size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT></BLOCKQUOTE>
  <BLOCKQUOTE dir=ltr 
  style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>Jiang: The draft mentioned above 
    also do selective flushing, not blind flushing. And for what I know, this 
    very draft may depend </FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>on the hub-and-spoke network 
    topology while our draft seems to be capable of a broader application 
    scenario such as&nbsp;a CE&nbsp;dual-homed </FONT></DIV>
    <DIV><FONT face=$BAWBN(B color=#800000 size=2>to PEs by ACs.&nbsp;The formats and 
    contents of the LDP messages are&nbsp;similar in all these drafts or RFC for 
    selective flushing, and as </FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><FONT size=+0>my 
    analysis goes, complexity&nbsp;(times of lookup and write)</FONT><FONT 
    size=+0>&nbsp;in dataplane may be </FONT><FONT size=+0>commensurate for all 
    these mechanisms</FONT><FONT size=+0>.&nbsp;</FONT><SPAN 
    class=012333915-08082008><FONT face=Arial color=#0000ff 
    size=2>&nbsp;</FONT></SPAN></FONT></FONT></FONT></DIV>
    <DIV><FONT size=2><FONT face=$BAWBN(B><FONT color=#800000><SPAN 
    class=012333915-08082008></SPAN></FONT></FONT></FONT>&nbsp;</DIV>
    <DIV><FONT size=+0><SPAN class=012333915-08082008><FONT color=#0000ff><FONT 
    face=Arial size=2>WH&gt; Right but the main difference is that you only need 
    to indicate the PE or PW for which the MAC flush is performed.&nbsp;In your 
    proposal you need to indicate the full MAC address list which will be very 
    processing intensive if you have a large amount of MAC addresses in the VPLS 
    service. The message will get very big and you will have to sent lots of 
    them due to fragmentation implications. </FONT></FONT></SPAN></FONT></DIV>
    <DIV><FONT size=+0><SPAN class=012333915-08082008><FONT color=#0000ff><FONT 
    face=Arial size=2><FONT face=SimSun color=#000000><FONT 
    color=#0000ff>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03 provides</FONT> 
    </FONT>a simple (blind was maybe the wrong wording)&nbsp;flush mechanism. 
    The indication will be fast and the re-learning will happen instantly after 
    the MAC is removed from the MAC FIB.</FONT></FONT></SPAN></FONT></DIV>
    <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008>You proposal will delay the trigger when the amount 
    of MAC @ grows and on top the receiving PE needs a fast lookup mechanism to 
    find he proper MAC and replace it which means delay the removal of the 
    FIB.</SPAN></FONT></DIV>
    <DIV><FONT face=Arial color=#0000ff size=2><SPAN 
    class=012333915-08082008></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>Jiang:&nbsp;As&nbsp;indicated in our draft, MAC 
    Switching message can also carry only a null-list of MACs. It's OK to change 
    all the MACs associated with a PW </SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008>to another PW as egress.<SPAN 
    class=887291606-11082008><FONT face=Arial color=#0000ff 
    size=2>&nbsp;</FONT></SPAN></SPAN></FONT></DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008><SPAN 
    class=887291606-11082008></SPAN></SPAN></FONT>&nbsp;</DIV>
    <DIV><FONT face="Times New Roman" color=#000080><SPAN 
    class=012333915-08082008><SPAN class=887291606-11082008><FONT face=Arial 
    color=#0000ff size=2>WH&gt; This function is already proposed 
    in&nbsp;</FONT>&nbsp;<SPAN class=012333915-08082008><FONT 
    color=#0000ff><FONT face=Arial size=2><FONT face=SimSun color=#000000><FONT 
    color=#0000ff>Draft-pdutta-l2vpn-vpls-ldp-mac-opt-03. Why do you think your 
    solution is 
    better?</FONT></FONT></FONT></FONT></SPAN></SPAN></SPAN></FONT></DIV>
    <DIV><FONT face=$BAWBN(B></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT 
    face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B size=2></FONT><FONT face=$BAWBN(B 
    size=2></FONT><FONT face=$BAWBN(B size=2></FONT><BR><FONT face=$BAWBN(B size=2>Best 
    regards,</FONT></DIV>
    <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
    <DIV><FONT face=$BAWBN(B size=2>Yuanlong</FONT></DIV>
    <BLOCKQUOTE dir=ltr 
    style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
      <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
      <DIV 
      style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
      <A title=wim.henderickx@alcatel-lucent.be 
      href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
      href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
      href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Thursday, August 07, 2008 1:24 
      PM</DIV>
      <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE: RE:</DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT><BR></DIV>
      <DIV dir=ltr align=left><SPAN class=770051905-07082008>
      <DIV dir=ltr align=left><FONT face=Arial><FONT color=#0000ff><FONT 
      size=2>Yuanlong<SPAN 
      class=924080318-06082008>,</SPAN></FONT></FONT></FONT></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><SPAN 
      class=770051905-07082008><FONT face=Arial color=#0000ff size=2>My point is 
      that how are you going to define the behavior on each PE when the topology 
      is adapted. I hope you are not going to change the configuration on each 
      PE in the VPLS each time.</FONT></SPAN></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
      color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV></SPAN></DIV>
      <DIV dir=ltr align=left><SPAN class=770051905-07082008><FONT face=Arial 
      color=#0000ff size=2>On top, by doing selective MAC flush you need to go 
      one by one to the items and hence the processing to withdraw the MAC(s) 
      will be longer. When you use the existing mechanisms you can flush without 
      the extra lookups and hence the performance will be faster. FLushing a MAC 
      will result in re-learning which means you do not have traffic impact, 
      just a small flooding until the MAC address is 
      re-learned.</FONT></SPAN></DIV>
      <DIV>&nbsp;</DIV>
      <DIV><SPAN class=770051905-07082008></SPAN><FONT face=Arial><FONT 
      color=#0000ff><FONT size=2>C<SPAN 
      class=770051905-07082008>heers,</SPAN></FONT></FONT></FONT></DIV>
      <DIV><FONT size=+0><FONT color=#0000ff><FONT size=2><SPAN 
      class=770051905-07082008></SPAN></FONT></FONT></FONT><SPAN 
      class=770051905-07082008></SPAN><FONT face=Arial><FONT color=#0000ff><FONT 
      size=2>W<SPAN 
      class=770051905-07082008>im</SPAN></FONT></FONT></FONT><BR></DIV>
      <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
      <HR tabIndex=-1>
      <FONT face=Tahoma size=2><B>From:</B> Jiang Yuan-long 
      [mailto:yljiang@huawei.com] <BR><B>Sent:</B> donderdag 7 augustus 2008 
      6:00<BR><B>To:</B> HENDERICKX Wim<BR><B>Cc:</B> 
      l2vpn@ietf.org<BR><B>Subject:</B> Re: RE:<BR></FONT><BR></DIV>
      <DIV></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>Wim:</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>Thanks for you comment. My draft is not a 
      standlone protocol, it can be deployed with other protocol 
      suites.</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>For what I know, "PW redundancy" is being 
      standardized in PWE3 WG, which can synchronize the behavior 
of</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>the PEs and PWs in PW switching. So it is 
      not&nbsp;such a&nbsp;difficult task to do PW switching, and do MAC 
      switching</FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2>subsequently. </FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2></FONT>&nbsp;</DIV>
      <DIV><FONT face=$BAWBN(B size=2>The MAC withdrawal process defined in RFC 4762 
      is not so simple as you think,&nbsp;it maybe not just<FONT 
      size=2>&nbsp;"flush all MACs" </FONT></FONT></DIV>
      <DIV><FONT face=$BAWBN(B size=2><FONT size=2>blindly, but can selectively flush 
      MACs, which computation comlexity is on the same level as my 
      draft</FONT><FONT size=2>.</FONT></DIV></FONT>
      <BLOCKQUOTE dir=ltr 
      style="PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px">
        <DIV style="FONT: 9pt $BAWBN(B">----- Original Message ----- </DIV>
        <DIV 
        style="BACKGROUND: #e4e4e4; FONT: 9pt $BAWBN(B; font-color: black"><B>From:</B> 
        <A title=wim.henderickx@alcatel-lucent.be 
        href="mailto:wim.henderickx@alcatel-lucent.be">HENDERICKX Wim</A> </DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>To:</B> <A title=yljiang@huawei.com 
        href="mailto:yljiang@huawei.com">Jiang Yuan-long</A> ; <A 
        title=sajassi@cisco.com 
        href="mailto:sajassi@cisco.com">sajassi@cisco.com</A> </DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>Cc:</B> <A title=l2vpn@ietf.org 
        href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A> </DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>Sent:</B> Thursday, August 07, 2008 4:18 
        AM</DIV>
        <DIV style="FONT: 9pt $BAWBN(B"><B>Subject:</B> RE:</DIV>
        <DIV><BR></DIV>
        <DIV dir=ltr align=left>Yuanlong<SPAN 
        class=924080318-06082008>,</SPAN></DIV>
        <DIV dir=ltr align=left><SPAN 
        class=924080318-06082008></SPAN>&nbsp;</DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>The difficulty with your draft is that it requires 
        a PE to do different things depending in which situation the PE is in, 
        and this makes it very cumbersome in case changes happen to topologies 
        or configurations.</FONT></SPAN></DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>My view is that we need a simple MAC flush which we 
        have today in LDP.</FONT></SPAN></DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2></FONT></SPAN>&nbsp;</DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>Cheers,</FONT></SPAN></DIV>
        <DIV dir=ltr align=left><SPAN class=924080318-06082008><FONT face=Arial 
        color=#0000ff size=2>WIm</FONT></SPAN></DIV><BR>
        <DIV class=OutlookMessageHeader lang=en-us dir=ltr align=left>
        <HR tabIndex=-1>
        <FONT face=Tahoma size=2><B>From:</B> <A 
        href="mailto:l2vpn-bounces@ietf.org">l2vpn-bounces@ietf.org</A> 
        [mailto:l2vpn-bounces@ietf.org] <B>On Behalf Of </B>Jiang 
        Yuan-long<BR><B>Sent:</B> woensdag 6 augustus 2008 11:30<BR><B>To:</B> 
        <A href="mailto:sajassi@cisco.com">sajassi@cisco.com</A><BR><B>Cc:</B> 
        <A href="mailto:l2vpn@ietf.org">l2vpn@ietf.org</A><BR><B>Subject:</B> 
        <BR></FONT><BR></DIV>
        <DIV></DIV>
        <DIV><FONT face="Times New Roman">Hi, Ali:</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">Thanks for your comments in the last 
        L2VPN WG meeting, I will try to give you some reasons here.</FONT></DIV>
        <DIV><FONT face="Times New Roman">Following are the&nbsp;two 
        questions:</FONT></DIV>
        <DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">1. What do you think will be faster 
        convergence?&nbsp;When you flush MAC addresses, </FONT></DIV>
        <DIV><FONT face="Times New Roman">the learning is done in data plane at 
        wire speed. Why does that prolong convergence time?<BR></FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">-Faster convergence means the time to 
        associate the MACs with the "exact" out port</FONT></DIV>
        <DIV><FONT face="Times New Roman">will be shorter. Though learning is 
        done in data plane at wire speed, the time for a frame to arrive 
        </FONT></DIV>
        <DIV><FONT face="Times New Roman">at the end host and the response time 
        for a host to send a reverse frame&nbsp;accross the PE (on which the 
        </FONT></DIV>
        <DIV><FONT face="Times New Roman">MACs&nbsp;is flushed) will be a rather 
        long time. So the learning process is not only determined solely by 
        </FONT></DIV>
        <DIV><FONT face="Times New Roman">the wire </FONT><FONT 
        face="Times New Roman">speed, but also on the host processing time and 
        delays on all the intermediate routers.</FONT></DIV>
        <DIV><FONT face="Times New Roman">For this reason, flooding may also 
        last for a rather long time.</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV></DIV>
        <DIV><FONT face="Times New Roman">2. You consider a simple case with 
        simple backup for primary.&nbsp;In cases where primary fails and traffic 
        gets<BR>&nbsp;distributed among multiple PE's, how do you take care of 
        that?</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">-I believe the most commonly 
        deployed&nbsp;scenario is&nbsp;one backup PE for a primary 
        PE.</FONT></DIV>
        <DIV><FONT face="Times New Roman">In my opinion, for the scenario of 
        multiple backup PEs protecting a primary PE, </FONT></DIV>
        <DIV><FONT face="Times New Roman">only one PE out of them will be 
        elected as the working PE when the primary PE failed, </FONT></DIV>
        <DIV><FONT face="Times New Roman">rather than distribute the traffic 
        over multiple backup PEs,</FONT></DIV>
        <DIV><FONT face="Times New Roman">otherwise, loops may be introduced to 
        the attached customer site.</FONT></DIV>
        <DIV><FONT face="Times New Roman">BTW, in a similar way, for PW 
        redundancy mechanism, only 1 secondary is selected as the working PW 
        (see </FONT></DIV>
        <DIV><FONT face="Times New Roman">draft-ietf-pwe3-redundancy-bit-00.txt 
        in&nbsp;PWE3 WG)</FONT></DIV>
        <DIV><FONT face="Times New Roman">Anyway, if there is any such 
        requirements confirmed, I think it will be OK for us to find some 
        solutions and </FONT></DIV>
        <DIV><FONT face="Times New Roman">add them to&nbsp;the future 
        drafts.&nbsp;</FONT></DIV>
        <DIV><FONT face="Times New Roman"></FONT>&nbsp;</DIV>
        <DIV><FONT face="Times New Roman">cheers,</FONT></DIV>
        <DIV><FONT 
    face="Times New Roman">Yuanlong<BR></FONT></DIV></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML>

------_=_NextPart_001_01C8FB7A.7B0EE58A--


From l2vpn-bounces@ietf.org  Tue Aug 26 20:55:11 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id AEACE3A6AAD;
	Tue, 26 Aug 2008 20:55:11 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id E68AC3A6AAD
	for <l2vpn@core3.amsl.com>; Tue, 26 Aug 2008 20:55:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.705
X-Spam-Level: **
X-Spam-Status: No, score=2.705 tagged_above=-999 required=5
	tests=[BAYES_50=0.001, FH_RELAY_NODNS=1.451, HELO_MISMATCH_COM=0.553,
	J_CHICKENPOX_13=0.6, RDNS_NONE=0.1]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id FDBr3t7TbF07 for <l2vpn@core3.amsl.com>;
	Tue, 26 Aug 2008 20:55:10 -0700 (PDT)
Received: from szxga01-in.huawei.com (unknown [119.145.14.64])
	by core3.amsl.com (Postfix) with ESMTP id 200E63A69FB
	for <l2vpn@ietf.org>; Tue, 26 Aug 2008 20:55:10 -0700 (PDT)
Received: from huawei.com (szxga01-in [172.24.2.3])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTP id <0K6800ETYQROQY@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Wed, 27 Aug 2008 11:52:36 +0800 (CST)
Received: from c50408 ([10.111.12.79])
	by szxga01-in.huawei.com (iPlanet Messaging Server 5.2 HotFix 2.14
	(built Aug
	8 2006)) with ESMTPA id <0K6800CVGQRNU2@szxga01-in.huawei.com> for
	l2vpn@ietf.org; Wed, 27 Aug 2008 11:52:36 +0800 (CST)
Date: Wed, 27 Aug 2008 11:52:25 +0800
From: Cao Wei <caowayne@huawei.com>
Subject: Question about VPMS
To: "l2vpn@ietf.org" <l2vpn@ietf.org>, "y.kamite" <y.kamite@ntt.com>
Message-id: <200808271152247241769@huawei.com>
MIME-version: 1.0
X-Mailer: Foxmail 6, 10, 201, 20 [cn]
Content-type: text/plain; charset=gb2312
Content-transfer-encoding: 7BIT
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Yuji,
  I am interested in VPMS. IMHO, support multicast traffic in traditional VPLS
might require a lot of filtering work, this may compromise the transparency 
of VPLS forwarding in SP network.
 
  But I have a few questions about VPMS, I would like to make it clear:

1. How often do you think the activation/deactivation of CE/AC will happen?
   Supporting dynamicity is useful, but if you want a very flexible solution
   for dynamic joining/leaving, it may cause the solution be complex (
   especially when considering dual multicast sources, multi-homing, 
   protection, etc). I hope debate about MVPN in L3VPN WG don't happen in 
   L2VPN again. :)

2. Compare to VPLS, there is no MAC learning. And the roles for the PE (sender
   or reciver ) can not be changed in one VPMS. So if in VPLS similar environment,
   why not use P2MP LSP directly? It is simple.

Regards,
Cao Wei   				
--------------
Cao Wei
2008-08-27



From l2vpn-bounces@ietf.org  Fri Aug 29 10:45:51 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 065CD28C111;
	Fri, 29 Aug 2008 10:45:51 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 122EE3A6A59
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 10:45:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id QYcDODLvi9XP for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 10:45:44 -0700 (PDT)
Received: from sj-iport-3.cisco.com (sj-iport-3.cisco.com [171.71.176.72])
	by core3.amsl.com (Postfix) with ESMTP id 2CE6C28C10D
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 10:45:42 -0700 (PDT)
X-IronPort-AV: E=Sophos;i="4.32,293,1217808000"; d="scan'208";a="98196619"
Received: from sj-dkim-3.cisco.com ([171.71.179.195])
	by sj-iport-3.cisco.com with ESMTP; 29 Aug 2008 17:45:11 +0000
Received: from sj-core-4.cisco.com (sj-core-4.cisco.com [171.68.223.138])
	by sj-dkim-3.cisco.com (8.12.11/8.12.11) with ESMTP id m7THjB5d018158; 
	Fri, 29 Aug 2008 10:45:11 -0700
Received: from xbh-sjc-221.amer.cisco.com (xbh-sjc-221.cisco.com
	[128.107.191.63])
	by sj-core-4.cisco.com (8.13.8/8.13.8) with ESMTP id m7THjBXD011289;
	Fri, 29 Aug 2008 17:45:11 GMT
Received: from xmb-sjc-239.amer.cisco.com ([128.107.191.105]) by
	xbh-sjc-221.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 29 Aug 2008 10:45:11 -0700
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Date: Fri, 29 Aug 2008 10:45:00 -0700
Message-ID: <483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
In-Reply-To: <B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2cfusbpP5APx1QhmSP06sW6dyZwATRPJQBM9UCOA=
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
From: "Ali Sajassi (sajassi)" <sajassi@cisco.com>
To: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>,
	"Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 17:45:11.0664 (UTC)
	FILETIME=[FC567F00:01C909FE]
DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; l=986; t=1220031911; x=1220895911;
	c=relaxed/simple; s=sjdkim3002;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=sajassi@cisco.com;
	z=From:=20=22Ali=20Sajassi=20(sajassi)=22=20<sajassi@cisco.c
	om> |Subject:=20Concern=20regarding=20=20draft-kompella-l2vpn-v
	pls-multihoming-01 |Sender:=20;
	bh=l5Pht7n6GnJasHv0zaj8uj+tIcx9KoC2R7CjQdE/zFg=;
	b=TMQtCnqIsTcg6lKBl6YuCbdgsmy1yWnlbD8CJb0rHtK5DHM8M2czZoZhiz
	udYMCPna99+jzXt7OGdaIeHoKGLSPTLvud9UUpeyRaOfMd8VyBm+pHLJBCeM
	KgxSlX4TJh;
Authentication-Results: sj-dkim-3; header.From=sajassi@cisco.com; dkim=pass (
	sig from cisco.com/sjdkim3002 verified; ); 
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

=20
Bhupesh, Wim:

I was just going through this draft and I have the following concern
regarding the dual-home mechanism described in this draft:

Both BGP Path selection and VPLS Path selection described in section 3.4
are based on the corresponding VSIs for the same VPLS on different PE
devices (AD =3D <RD, VE-ID, VBO, VBS, LB; DOM, BNH, CF, VP, LP>). In =
other
words, the BGP advertisement allows to pick one VSI for a given VPLS
instance among multiple PE devices that are dual/multi-homed. However,
redundancy needs to be provided at the port level and not VSI level for
dual/multi-homing of CE devices !!

In other words, if a CE1-PE1 link fails, you don't want to switch the
VSI from PE1 to PE2 and impact all other CE devices that are dual-homed
to PE1/PE2 for the same VPLS instance. Besides, the current H-VPLS
allows for dual-homing of the CE/u-PE devices that are load balanced
among dual/multi-homed PE devices for the same VPLS instance !!

-Ali



From l2vpn-bounces@ietf.org  Fri Aug 29 11:20:00 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 850253A6A72;
	Fri, 29 Aug 2008 11:20:00 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 526403A6A72
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 11:20:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.249
X-Spam-Level: 
X-Spam-Status: No, score=-2.249 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, HELO_EQ_FR=0.35]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id dnldhXvOMOLy for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 11:19:59 -0700 (PDT)
Received: from smail6.alcatel.fr (colt-na5.alcatel.fr [62.23.212.5])
	by core3.amsl.com (Postfix) with ESMTP id 4F8D73A6A59
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 11:19:58 -0700 (PDT)
Received: from FRVELSBHS02.ad2.ad.alcatel.com (frvelsbhs02.ad2.ad.alcatel.com
	[155.132.6.74])
	by smail6.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7TIJwMT014626; 
	Fri, 29 Aug 2008 20:19:58 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.54]) by
	FRVELSBHS02.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Fri, 29 Aug 2008 20:19:58 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Date: Fri, 29 Aug 2008 20:20:29 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2cfusbpP5APx1QhmSP06sW6dyZwATRPJQBM9UCOAAAdS9MA==
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Ali Sajassi (sajassi)" <sajassi@cisco.com>,
	"Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 18:19:58.0451 (UTC)
	FILETIME=[D828EC30:01C90A03]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.84
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Ali,

This is the same concern I raised in the emails before and will be
difficult to solve if you let the remote PE(s) make a decision for
multi-homing.

Cheers,
Wim=20

-----Original Message-----
From: Ali Sajassi (sajassi) [mailto:sajassi@cisco.com]=20
Sent: vrijdag 29 augustus 2008 19:45
To: HENDERICKX Wim; Bhupesh Kothari
Cc: l2vpn@ietf.org
Subject: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01

=20
Bhupesh, Wim:

I was just going through this draft and I have the following concern
regarding the dual-home mechanism described in this draft:

Both BGP Path selection and VPLS Path selection described in section 3.4
are based on the corresponding VSIs for the same VPLS on different PE
devices (AD =3D <RD, VE-ID, VBO, VBS, LB; DOM, BNH, CF, VP, LP>). In =
other
words, the BGP advertisement allows to pick one VSI for a given VPLS
instance among multiple PE devices that are dual/multi-homed. However,
redundancy needs to be provided at the port level and not VSI level for
dual/multi-homing of CE devices !!

In other words, if a CE1-PE1 link fails, you don't want to switch the
VSI from PE1 to PE2 and impact all other CE devices that are dual-homed
to PE1/PE2 for the same VPLS instance. Besides, the current H-VPLS
allows for dual-homing of the CE/u-PE devices that are load balanced
among dual/multi-homed PE devices for the same VPLS instance !!

-Ali



From l2vpn-bounces@ietf.org  Fri Aug 29 11:31:08 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id DC8D53A6A37;
	Fri, 29 Aug 2008 11:31:08 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 8EBF33A69CA
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 11:31:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id aWaLxle75ZAS for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 11:31:06 -0700 (PDT)
Received: from sj-iport-6.cisco.com (sj-iport-6.cisco.com [171.71.176.117])
	by core3.amsl.com (Postfix) with ESMTP id 5EC9B3A6959
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 11:31:06 -0700 (PDT)
X-IronPort-AV: E=Sophos;i="4.32,293,1217808000"; d="scan'208";a="149235080"
Received: from sj-dkim-3.cisco.com ([171.71.179.195])
	by sj-iport-6.cisco.com with ESMTP; 29 Aug 2008 18:31:10 +0000
Received: from sj-core-5.cisco.com (sj-core-5.cisco.com [171.71.177.238])
	by sj-dkim-3.cisco.com (8.12.11/8.12.11) with ESMTP id m7TIVAoG000591; 
	Fri, 29 Aug 2008 11:31:10 -0700
Received: from xbh-sjc-231.amer.cisco.com (xbh-sjc-231.cisco.com
	[128.107.191.100])
	by sj-core-5.cisco.com (8.13.8/8.13.8) with ESMTP id m7TIVAIv013421;
	Fri, 29 Aug 2008 18:31:10 GMT
Received: from xmb-sjc-239.amer.cisco.com ([128.107.191.105]) by
	xbh-sjc-231.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 29 Aug 2008 11:31:10 -0700
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Date: Fri, 29 Aug 2008 11:30:58 -0700
Message-ID: <483F121CD2155B43B56DB71228DBAF150220969B@xmb-sjc-239.amer.cisco.com>
In-Reply-To: <B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2cfusbpP5APx1QhmSP06sW6dyZwATRPJQBM9UCOAAAdS9MAAAZAtg
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
From: "Ali Sajassi (sajassi)" <sajassi@cisco.com>
To: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>,
	"Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 18:31:10.0198 (UTC)
	FILETIME=[688D6160:01C90A05]
DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; l=1971; t=1220034670;
	x=1220898670; c=relaxed/simple; s=sjdkim3002;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=sajassi@cisco.com;
	z=From:=20=22Ali=20Sajassi=20(sajassi)=22=20<sajassi@cisco.c
	om> |Subject:=20RE=3A=20Concern=20regarding=20=20draft-kompella
	-l2vpn-vpls-multihoming-01 |Sender:=20;
	bh=i3c3Wcq++MLcGwMqExaQQ81IXmyw1QSN58BoCqsarfc=;
	b=eyyPwSO1IQ2x3ntwDmOK8hUmf9TYfjAaaXlg596fLC/hKc4T1kG5amd/vZ
	svA8zoRcLsPcRw2f8tkk9pRPWYaTaJjw0V9RbUeqAeDsQVyeWSPIDEspcfFa
	5DY80E2Udh;
Authentication-Results: sj-dkim-3; header.From=sajassi@cisco.com; dkim=pass (
	sig from cisco.com/sjdkim3002 verified; ); 
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org



Wim,

So, has this issue been acknowledged and if so, then what is the remedy
?

Cheers,
Ali=20

> -----Original Message-----
> From: HENDERICKX Wim [mailto:wim.henderickx@alcatel-lucent.be]=20
> Sent: Friday, August 29, 2008 11:20 AM
> To: Ali Sajassi (sajassi); Bhupesh Kothari
> Cc: l2vpn@ietf.org
> Subject: RE: Concern regarding=20
> draft-kompella-l2vpn-vpls-multihoming-01
>=20
> Ali,
>=20
> This is the same concern I raised in the emails before and=20
> will be difficult to solve if you let the remote PE(s) make a=20
> decision for multi-homing.
>=20
> Cheers,
> Wim=20
>=20
> -----Original Message-----
> From: Ali Sajassi (sajassi) [mailto:sajassi@cisco.com]
> Sent: vrijdag 29 augustus 2008 19:45
> To: HENDERICKX Wim; Bhupesh Kothari
> Cc: l2vpn@ietf.org
> Subject: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
>=20
> =20
> Bhupesh, Wim:
>=20
> I was just going through this draft and I have the following=20
> concern regarding the dual-home mechanism described in this draft:
>=20
> Both BGP Path selection and VPLS Path selection described in=20
> section 3.4 are based on the corresponding VSIs for the same=20
> VPLS on different PE devices (AD =3D <RD, VE-ID, VBO, VBS, LB;=20
> DOM, BNH, CF, VP, LP>). In other words, the BGP advertisement=20
> allows to pick one VSI for a given VPLS instance among=20
> multiple PE devices that are dual/multi-homed. However,=20
> redundancy needs to be provided at the port level and not VSI=20
> level for dual/multi-homing of CE devices !!
>=20
> In other words, if a CE1-PE1 link fails, you don't want to=20
> switch the VSI from PE1 to PE2 and impact all other CE=20
> devices that are dual-homed to PE1/PE2 for the same VPLS=20
> instance. Besides, the current H-VPLS allows for dual-homing=20
> of the CE/u-PE devices that are load balanced among=20
> dual/multi-homed PE devices for the same VPLS instance !!
>=20
> -Ali
>=20
>=20


From l2vpn-bounces@ietf.org  Fri Aug 29 11:32:55 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 9CF6428C11C;
	Fri, 29 Aug 2008 11:32:55 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id A283F28C11C
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 11:32:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.249
X-Spam-Level: 
X-Spam-Status: No, score=-2.249 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, HELO_EQ_FR=0.35]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id hkfiqgclLUYK for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 11:32:53 -0700 (PDT)
Received: from smail6.alcatel.fr (gc-na5.alcatel.fr [64.208.49.5])
	by core3.amsl.com (Postfix) with ESMTP id 887B03A6AA0
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 11:32:53 -0700 (PDT)
Received: from FRVELSBHS02.ad2.ad.alcatel.com (frvelsbhs02.ad2.ad.alcatel.com
	[155.132.6.74])
	by smail6.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7TIWrJd015415; 
	Fri, 29 Aug 2008 20:32:53 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.54]) by
	FRVELSBHS02.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Fri, 29 Aug 2008 20:32:53 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Date: Fri, 29 Aug 2008 20:33:24 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D703265844@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <483F121CD2155B43B56DB71228DBAF150220969B@xmb-sjc-239.amer.cisco.com>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding  draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: Acj2cfusbpP5APx1QhmSP06sW6dyZwATRPJQBM9UCOAAAdS9MAAAZAtgAAAVcyA=
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF150220969B@xmb-sjc-239.amer.cisco.com>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Ali Sajassi (sajassi)" <sajassi@cisco.com>,
	"Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 18:32:53.0815 (UTC)
	FILETIME=[A6501470:01C90A05]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.84
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Ali,

It has been acknowledge, but the remedy has not been provided so far
afaik.

Cheers,
Wim=20

-----Original Message-----
From: Ali Sajassi (sajassi) [mailto:sajassi@cisco.com]=20
Sent: vrijdag 29 augustus 2008 20:31
To: HENDERICKX Wim; Bhupesh Kothari
Cc: l2vpn@ietf.org
Subject: RE: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01



Wim,

So, has this issue been acknowledged and if so, then what is the remedy
?

Cheers,
Ali=20

> -----Original Message-----
> From: HENDERICKX Wim [mailto:wim.henderickx@alcatel-lucent.be]
> Sent: Friday, August 29, 2008 11:20 AM
> To: Ali Sajassi (sajassi); Bhupesh Kothari
> Cc: l2vpn@ietf.org
> Subject: RE: Concern regarding
> draft-kompella-l2vpn-vpls-multihoming-01
>=20
> Ali,
>=20
> This is the same concern I raised in the emails before and will be=20
> difficult to solve if you let the remote PE(s) make a decision for=20
> multi-homing.
>=20
> Cheers,
> Wim
>=20
> -----Original Message-----
> From: Ali Sajassi (sajassi) [mailto:sajassi@cisco.com]
> Sent: vrijdag 29 augustus 2008 19:45
> To: HENDERICKX Wim; Bhupesh Kothari
> Cc: l2vpn@ietf.org
> Subject: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
>=20
> =20
> Bhupesh, Wim:
>=20
> I was just going through this draft and I have the following concern=20
> regarding the dual-home mechanism described in this draft:
>=20
> Both BGP Path selection and VPLS Path selection described in section=20
> 3.4 are based on the corresponding VSIs for the same VPLS on different

> PE devices (AD =3D <RD, VE-ID, VBO, VBS, LB; DOM, BNH, CF, VP, LP>). =
In=20
> other words, the BGP advertisement allows to pick one VSI for a given=20
> VPLS instance among multiple PE devices that are dual/multi-homed.=20
> However, redundancy needs to be provided at the port level and not VSI

> level for dual/multi-homing of CE devices !!
>=20
> In other words, if a CE1-PE1 link fails, you don't want to switch the=20
> VSI from PE1 to PE2 and impact all other CE devices that are=20
> dual-homed to PE1/PE2 for the same VPLS instance. Besides, the current

> H-VPLS allows for dual-homing of the CE/u-PE devices that are load=20
> balanced among dual/multi-homed PE devices for the same VPLS instance=20
> !!
>=20
> -Ali
>=20
>=20


From l2vpn-bounces@ietf.org  Fri Aug 29 12:03:11 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id C32B03A685A;
	Fri, 29 Aug 2008 12:03:11 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id AC46E3A685A
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 12:03:10 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id 4YIqBt57u3x2 for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 12:03:09 -0700 (PDT)
Received: from chip3og57.obsmtp.com (chip3og57.obsmtp.com [64.18.14.179])
	by core3.amsl.com (Postfix) with ESMTP id 5CCC63A67AB
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 12:03:03 -0700 (PDT)
Received: from source ([66.129.228.6]) by chip3ob57.postini.com ([64.18.6.12])
	with SMTP; Fri, 29 Aug 2008 12:02:49 PDT
Received: from p-emlb01-sac.jnpr.net ([66.129.254.46]) by p-emsmtp01.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Fri, 29 Aug 2008 12:02:10 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb01-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Fri, 29 Aug 2008 12:02:10 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Fri, 29 Aug 2008 11:55:38 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m7TItcu06090;
	Fri, 29 Aug 2008 11:55:38 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m7TItcSw008190;
	Fri, 29 Aug 2008 11:55:38 -0700
To: "Ali Sajassi (sajassi)" <sajassi@cisco.com>
Subject: Re: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
Date: Fri, 29 Aug 2008 11:55:38 -0700
Message-ID: <8189.1220036138@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 18:55:38.0851 (UTC)
	FILETIME=[D3F00B30:01C90A08]
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Ali,


Ali Sajassi (sajassi) <sajassi@cisco.com> wrote:

>  
> Bhupesh, Wim:
> 
> I was just going through this draft and I have the following concern
> regarding the dual-home mechanism described in this draft:
> 
> Both BGP Path selection and VPLS Path selection described in section 3.4
> are based on the corresponding VSIs for the same VPLS on different PE
> devices (AD = <RD, VE-ID, VBO, VBS, LB; DOM, BNH, CF, VP, LP>). In other
> words, the BGP advertisement allows to pick one VSI for a given VPLS
> instance among multiple PE devices that are dual/multi-homed. However,
> redundancy needs to be provided at the port level and not VSI level for
> dual/multi-homing of CE devices !!

Not correct.  If you see other email thread on this topic on the
mailing list, I described scenarios that have multiple sites (or VE
IDs).  So, a PE can be a DF for a particular VE ID and non-DF for
another VE ID for the same VSI.  So, redundancy is not at the VSI
level; it is at the VE ID level.  A VE ID can map to a port or a set
of port depending on the topology.

> 
> In other words, if a CE1-PE1 link fails, you don't want to switch the
> VSI from PE1 to PE2 and impact all other CE devices that are dual-homed
> to PE1/PE2 for the same VPLS instance. 

Correct and thats not the case in BGP VPLS. 


> Besides, the current H-VPLS allows for dual-homing of the CE/u-PE
> devices that are load balanced among dual/multi-homed PE devices for
> the same VPLS instance !!


For a particular VLAN, a CE multi-homed to two PEs would require STP
or some other protocol in LDP VPLS to prevent loop as LDP VPLS PEs are
unware of the multi-homing.


> 
> -Ali
> 

Thanks,
Bhupesh


From l2vpn-bounces@ietf.org  Fri Aug 29 12:14:17 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 207AB3A6AAA;
	Fri, 29 Aug 2008 12:14:17 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 95B773A6A37
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 12:14:15 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.599
X-Spam-Level: 
X-Spam-Status: No, score=-2.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id zp0FgqUO8qw5 for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 12:14:14 -0700 (PDT)
Received: from audl951.usa.alcatel.com (audl951.usa.alcatel.com
	[143.209.238.161])
	by core3.amsl.com (Postfix) with ESMTP id A0C493A6A94
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 12:14:13 -0700 (PDT)
Received: from usdalsbhs01.ad3.ad.alcatel.com (usdalsbhs01.usa.alcatel.com
	[172.22.216.19])
	by audl951.usa.alcatel.com (ALCANET) with ESMTP id m7TJE8iG002173;
	Fri, 29 Aug 2008 14:14:11 -0500
Received: from USDALSMBS03.ad3.ad.alcatel.com ([172.22.216.11]) by
	usdalsbhs01.ad3.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Fri, 29 Aug 2008 14:14:09 -0500
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="US-ASCII"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
Date: Fri, 29 Aug 2008 14:14:08 -0500
Message-ID: <4A5028372622294A99B8FFF6BD06EB7B04926ABC@USDALSMBS03.ad3.ad.alcatel.com>
In-Reply-To: <8189.1220036138@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: AckKCefqyId4Pvq1TL6cFK75cg5awwAAI8TA
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com><483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
	<8189.1220036138@bhupesh-f8.jnpr.net>
From: "BALUS Florin" <Florin.Balus@alcatel-lucent.com>
To: "Bhupesh Kothari" <bhupesh@juniper.net>,
	"Ali Sajassi \(sajassi\)" <sajassi@cisco.com>
X-OriginalArrivalTime: 29 Aug 2008 19:14:09.0970 (UTC)
	FILETIME=[6A375920:01C90A0B]
X-Scanned-By: MIMEDefang 2.64 on 143.209.238.34
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

See in-line..

> -----Original Message-----
> From: l2vpn-bounces@ietf.org [mailto:l2vpn-bounces@ietf.org] On Behalf
> Of Bhupesh Kothari
> Sent: Friday, August 29, 2008 11:56 AM
> To: Ali Sajassi (sajassi)
> Cc: l2vpn@ietf.org
> Subject: Re: Concern regarding
draft-kompella-l2vpn-vpls-multihoming-01
>=20
> Ali,
>=20
>=20
> Ali Sajassi (sajassi) <sajassi@cisco.com> wrote:
>=20
> >
> > Bhupesh, Wim:
> >
> > I was just going through this draft and I have the following concern
> > regarding the dual-home mechanism described in this draft:
> >
> > Both BGP Path selection and VPLS Path selection described in section
> 3.4
> > are based on the corresponding VSIs for the same VPLS on different
PE
> > devices (AD =3D <RD, VE-ID, VBO, VBS, LB; DOM, BNH, CF, VP, LP>). In
> other
> > words, the BGP advertisement allows to pick one VSI for a given VPLS
> > instance among multiple PE devices that are dual/multi-homed.
> However,
> > redundancy needs to be provided at the port level and not VSI level
> for
> > dual/multi-homing of CE devices !!
>=20
> Not correct.  If you see other email thread on this topic on the
> mailing list, I described scenarios that have multiple sites (or VE
> IDs).  So, a PE can be a DF for a particular VE ID and non-DF for
> another VE ID for the same VSI.  So, redundancy is not at the VSI
> level; it is at the VE ID level.  A VE ID can map to a port or a set
> of port depending on the topology.
>=20
[[FB]] For multiple Dual-Homed CE sites & associated VE IDs located on
two PEs (PE1 and PE2) any failure of a link to the lowest Dual-Homed VE
ID (say VE ID 1) will actually generate infrastructure changes on both
PE1 and PE2 - i.e. on PE1 all the PW infrastructure with the next VE ID
(say VE2) takes over, and on PE2 the new lowest ID (i.e. VE ID 1)
imposes his own new PW infrastructure.

> >
> > In other words, if a CE1-PE1 link fails, you don't want to switch
the
> > VSI from PE1 to PE2 and impact all other CE devices that are dual-
> homed
> > to PE1/PE2 for the same VPLS instance.
>=20
> Correct and thats not the case in BGP VPLS.

[[FB]] Actually I do think there are changes that will affect the VSI
infrastructure on both PEs - see above.

>=20
>=20
> > Besides, the current H-VPLS allows for dual-homing of the CE/u-PE
> > devices that are load balanced among dual/multi-homed PE devices for
> > the same VPLS instance !!
>=20
>=20
> For a particular VLAN, a CE multi-homed to two PEs would require STP
> or some other protocol in LDP VPLS to prevent loop as LDP VPLS PEs are
> unware of the multi-homing.
>=20
>=20
> >
> > -Ali
> >
>=20
> Thanks,
> Bhupesh


From l2vpn-bounces@ietf.org  Fri Aug 29 12:47:41 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id B75523A686B;
	Fri, 29 Aug 2008 12:47:41 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id B02BE3A67F2
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 12:47:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id tO62grp+n24X for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 12:47:38 -0700 (PDT)
Received: from chip3og55.obsmtp.com (chip3og55.obsmtp.com [64.18.14.175])
	by core3.amsl.com (Postfix) with ESMTP id 5D1253A686B
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 12:47:36 -0700 (PDT)
Received: from source ([66.129.228.6]) by chip3ob55.postini.com ([64.18.6.12])
	with SMTP; Fri, 29 Aug 2008 12:47:36 PDT
Received: from p-emlb02-sac.jnpr.net ([66.129.254.47]) by p-emsmtp01.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Fri, 29 Aug 2008 12:02:10 -0700
Received: from emailsmtp56.jnpr.net ([172.24.60.77]) by p-emlb02-sac.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Fri, 29 Aug 2008 12:02:10 -0700
Received: from magenta.juniper.net ([172.17.27.123]) by emailsmtp56.jnpr.net
	with Microsoft SMTPSVC(6.0.3790.3959); 
	Fri, 29 Aug 2008 12:01:33 -0700
Received: from bhupesh-f8.jnpr.net (bhupesh-f8.jnpr.net [172.24.24.195])
	by magenta.juniper.net (8.11.3/8.11.3) with ESMTP id m7TJ1Xu09422;
	Fri, 29 Aug 2008 12:01:33 -0700 (PDT)
	(envelope-from bhupesh@juniper.net)
Received: from bhupesh-f8.jnpr.net (localhost6.localdomain6 [127.0.0.1])
	by bhupesh-f8.jnpr.net (8.14.2/8.14.2) with ESMTP id m7TJ1WsX008634;
	Fri, 29 Aug 2008 12:01:33 -0700
To: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
Subject: Re: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
In-reply-to: <B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
X-Mailer: MH-E 8.0.3; nmh 1.2-20070115cvs; GNU Emacs 22.1.1
Date: Fri, 29 Aug 2008 12:01:32 -0700
Message-ID: <8633.1220036492@bhupesh-f8.jnpr.net>
From: Bhupesh Kothari <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 19:01:33.0358 (UTC)
	FILETIME=[A73D84E0:01C90A09]
Cc: l2vpn@ietf.org, "Ali Sajassi \(sajassi\)" <sajassi@cisco.com>
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Hi Wim,

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:

> Ali,
> 
> This is the same concern I raised in the emails before and will be
> difficult to solve if you let the remote PE(s) make a decision for
> multi-homing.

I thought I addressed all you MAC flush concerns. :-) 

As I said, MAC flush capability in BGP based VPLS will be added to
address the concern you pointed out (with multiple VE IDs) and other
cases where MAC flush is needed.

Can you be specific on what is your open concern. 

> 
> Cheers,
> Wim 

Thanks
Bhupesh


From l2vpn-bounces@ietf.org  Fri Aug 29 13:44:20 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 973A53A6AD8;
	Fri, 29 Aug 2008 13:44:20 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 221903A6AD8
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 13:44:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.599
X-Spam-Level: 
X-Spam-Status: No, score=-6.599 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, RCVD_IN_DNSWL_MED=-4]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id OS7GFisZg0cf for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 13:44:18 -0700 (PDT)
Received: from sj-iport-6.cisco.com (sj-iport-6.cisco.com [171.71.176.117])
	by core3.amsl.com (Postfix) with ESMTP id 15AB53A6893
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 13:44:18 -0700 (PDT)
X-IronPort-AV: E=Sophos;i="4.32,294,1217808000"; d="scan'208";a="149313067"
Received: from sj-dkim-1.cisco.com ([171.71.179.21])
	by sj-iport-6.cisco.com with ESMTP; 29 Aug 2008 20:44:22 +0000
Received: from sj-core-5.cisco.com (sj-core-5.cisco.com [171.71.177.238])
	by sj-dkim-1.cisco.com (8.12.11/8.12.11) with ESMTP id m7TKiLCr027759; 
	Fri, 29 Aug 2008 13:44:21 -0700
Received: from xbh-sjc-211.amer.cisco.com (xbh-sjc-211.cisco.com
	[171.70.151.144])
	by sj-core-5.cisco.com (8.13.8/8.13.8) with ESMTP id m7TKiLEi026386;
	Fri, 29 Aug 2008 20:44:21 GMT
Received: from xmb-sjc-239.amer.cisco.com ([128.107.191.105]) by
	xbh-sjc-211.amer.cisco.com with Microsoft SMTPSVC(6.0.3790.1830); 
	Fri, 29 Aug 2008 13:44:21 -0700
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
Date: Fri, 29 Aug 2008 13:44:09 -0700
Message-ID: <483F121CD2155B43B56DB71228DBAF1502209716@xmb-sjc-239.amer.cisco.com>
In-Reply-To: <8189.1220036138@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: AckKCeOJVHZGWkkPS96yPl1DIs2kGgADGX2g
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
	<8189.1220036138@bhupesh-f8.jnpr.net>
From: "Ali Sajassi (sajassi)" <sajassi@cisco.com>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 29 Aug 2008 20:44:21.0746 (UTC)
	FILETIME=[03E2D520:01C90A18]
DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; l=2651; t=1220042661;
	x=1220906661; c=relaxed/simple; s=sjdkim1004;
	h=Content-Type:From:Subject:Content-Transfer-Encoding:MIME-Version;
	d=cisco.com; i=sajassi@cisco.com;
	z=From:=20=22Ali=20Sajassi=20(sajassi)=22=20<sajassi@cisco.c
	om> |Subject:=20RE=3A=20Concern=20regarding=20draft-kompella-l2
	vpn-vpls-multihoming-01 |Sender:=20;
	bh=vGpeMQnzi5oR6uzymp47TRFMZ9SeOJb/IzwAr8L5LnM=;
	b=jaRX4S8VCI6MsB9V9nKBiuizn8178PWcYuk/RGYVXRvmpKrUWQA8JWqzmA
	+xP5dyOqIe2Rej0ddh4gvwW9HywmYE+H23x+bjkZiIhaL2+QyU3gSPkcwX/I
	UR38Uzdz2vfGneX+PmQVsVRhPmO66zQw5YbhI4MSiVJ2fCaKPP6z8=;
Authentication-Results: sj-dkim-1; header.From=sajassi@cisco.com; dkim=pass (
	sig from cisco.com/sjdkim1004 verified; ); 
Cc: l2vpn@ietf.org
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org


Bhupesh,

Currently BGP VPLS requires only a single VE-ID per VPLS instance per
PE. Now with this proposal, if there are 100 CE devices dual-homed to a
pair of PE devices for a given VPLS instance, you can end up with 100
VE-IDs for each PE for that VPLS instance and thus having 100 VPLS
routes being advertised in BGP versus a single route as done currently
!!=20

Cheers,
Ali =20

> -----Original Message-----
> From: Bhupesh Kothari [mailto:bhupesh@juniper.net]=20
> Sent: Friday, August 29, 2008 11:56 AM
> To: Ali Sajassi (sajassi)
> Cc: HENDERICKX Wim; l2vpn@ietf.org
> Subject: Re: Concern regarding=20
> draft-kompella-l2vpn-vpls-multihoming-01
>=20
> Ali,
>=20
>=20
> Ali Sajassi (sajassi) <sajassi@cisco.com> wrote:
>=20
> > =20
> > Bhupesh, Wim:
> >=20
> > I was just going through this draft and I have the=20
> following concern=20
> > regarding the dual-home mechanism described in this draft:
> >=20
> > Both BGP Path selection and VPLS Path selection described=20
> in section=20
> > 3.4 are based on the corresponding VSIs for the same VPLS=20
> on different=20
> > PE devices (AD =3D <RD, VE-ID, VBO, VBS, LB; DOM, BNH, CF,=20
> VP, LP>). In=20
> > other words, the BGP advertisement allows to pick one VSI=20
> for a given=20
> > VPLS instance among multiple PE devices that are dual/multi-homed.=20
> > However, redundancy needs to be provided at the port level=20
> and not VSI=20
> > level for dual/multi-homing of CE devices !!
>=20
> Not correct.  If you see other email thread on this topic on=20
> the mailing list, I described scenarios that have multiple=20
> sites (or VE IDs).  So, a PE can be a DF for a particular VE=20
> ID and non-DF for another VE ID for the same VSI.  So,=20
> redundancy is not at the VSI level; it is at the VE ID level.=20
>  A VE ID can map to a port or a set of port depending on the topology.
>=20
> >=20
> > In other words, if a CE1-PE1 link fails, you don't want to=20
> switch the=20
> > VSI from PE1 to PE2 and impact all other CE devices that are=20
> > dual-homed to PE1/PE2 for the same VPLS instance.
>=20
> Correct and thats not the case in BGP VPLS.=20
>=20
>=20
> > Besides, the current H-VPLS allows for dual-homing of the CE/u-PE=20
> > devices that are load balanced among dual/multi-homed PE=20
> devices for=20
> > the same VPLS instance !!
>=20
>=20
> For a particular VLAN, a CE multi-homed to two PEs would=20
> require STP or some other protocol in LDP VPLS to prevent=20
> loop as LDP VPLS PEs are unware of the multi-homing.
>=20
>=20
> >=20
> > -Ali
> >=20
>=20
> Thanks,
> Bhupesh
>=20


From l2vpn-bounces@ietf.org  Fri Aug 29 23:45:00 2008
Return-Path: <l2vpn-bounces@ietf.org>
X-Original-To: l2vpn-archive@megatron.ietf.org
Delivered-To: ietfarch-l2vpn-archive@core3.amsl.com
Received: from [127.0.0.1] (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id 783FA3A688C;
	Fri, 29 Aug 2008 23:45:00 -0700 (PDT)
X-Original-To: l2vpn@core3.amsl.com
Delivered-To: l2vpn@core3.amsl.com
Received: from localhost (localhost [127.0.0.1])
	by core3.amsl.com (Postfix) with ESMTP id DA3A53A688C
	for <l2vpn@core3.amsl.com>; Fri, 29 Aug 2008 23:44:59 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.249
X-Spam-Level: 
X-Spam-Status: No, score=-2.249 tagged_above=-999 required=5
	tests=[BAYES_00=-2.599, HELO_EQ_FR=0.35]
Received: from mail.ietf.org ([64.170.98.32])
	by localhost (core3.amsl.com [127.0.0.1]) (amavisd-new, port 10024)
	with ESMTP id GbLLAqiV4Kag for <l2vpn@core3.amsl.com>;
	Fri, 29 Aug 2008 23:44:59 -0700 (PDT)
Received: from smail6.alcatel.fr (gc-na5.alcatel.fr [64.208.49.5])
	by core3.amsl.com (Postfix) with ESMTP id E2DB83A681E
	for <l2vpn@ietf.org>; Fri, 29 Aug 2008 23:44:58 -0700 (PDT)
Received: from FRVELSBHS02.ad2.ad.alcatel.com (frvelsbhs02.ad2.ad.alcatel.com
	[155.132.6.74])
	by smail6.alcatel.fr (8.13.8/8.13.8/ICT) with ESMTP id m7U6iKhO020099; 
	Sat, 30 Aug 2008 08:44:20 +0200
Received: from FRVELSMBS22.ad2.ad.alcatel.com ([155.132.6.54]) by
	FRVELSBHS02.ad2.ad.alcatel.com with Microsoft
	SMTPSVC(6.0.3790.2499); Sat, 30 Aug 2008 08:44:20 +0200
X-MimeOLE: Produced By Microsoft Exchange V6.5
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Subject: RE: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
Date: Sat, 30 Aug 2008 08:44:50 +0200
Message-ID: <B128F666D4C8BD4FBF56CEAFB2DB66D70326585D@FRVELSMBS22.ad2.ad.alcatel.com>
In-Reply-To: <8633.1220036492@bhupesh-f8.jnpr.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01
Thread-Index: AckKEBtGLbn+HdYgRNuNqd0QCqpv5AAW3Bqg
References: <B128F666D4C8BD4FBF56CEAFB2DB66D702F859BF@FRVELSMBS22.ad2.ad.alcatel.com><21222.1217875881@bhupesh-f8.jnpr.net><B128F666D4C8BD4FBF56CEAFB2DB66D70302C011@FRVELSMBS22.ad2.ad.alcatel.com><27924.1217882240@bhupesh-f8.jnpr.net>
	<B128F666D4C8BD4FBF56CEAFB2DB66D70302C095@FRVELSMBS22.ad2.ad.alcatel.com>
	<483F121CD2155B43B56DB71228DBAF1502209660@xmb-sjc-239.amer.cisco.com>
	<B128F666D4C8BD4FBF56CEAFB2DB66D703265842@FRVELSMBS22.ad2.ad.alcatel.com>
	<8633.1220036492@bhupesh-f8.jnpr.net>
From: "HENDERICKX Wim" <wim.henderickx@alcatel-lucent.be>
To: "Bhupesh Kothari" <bhupesh@juniper.net>
X-OriginalArrivalTime: 30 Aug 2008 06:44:20.0924 (UTC)
	FILETIME=[D51177C0:01C90A6B]
X-Scanned-By: MIMEDefang 2.57 on 155.132.188.84
Cc: l2vpn@ietf.org, "Ali Sajassi \(sajassi\)" <sajassi@cisco.com>
X-BeenThere: l2vpn@ietf.org
X-Mailman-Version: 2.1.9
Precedence: list
List-Id: <l2vpn.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=unsubscribe>
List-Archive: <https://www.ietf.org/mailman/private/l2vpn>
List-Post: <mailto:l2vpn@ietf.org>
List-Help: <mailto:l2vpn-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/l2vpn>,
	<mailto:l2vpn-request@ietf.org?subject=subscribe>
Sender: l2vpn-bounces@ietf.org
Errors-To: l2vpn-bounces@ietf.org

Bhupesh,

My main open concern is the modeling of MAC flush in BGP. How do we
ensure with this proposal that the amount of messages remains low.

So in essence I would like to see a proposal how the MAC flush operation
in BGP would be modeled.

Cheers,
Wim

-----Original Message-----
From: Bhupesh Kothari [mailto:bhupesh@juniper.net]=20
Sent: vrijdag 29 augustus 2008 21:02
To: HENDERICKX Wim
Cc: Ali Sajassi (sajassi); l2vpn@ietf.org
Subject: Re: Concern regarding draft-kompella-l2vpn-vpls-multihoming-01

Hi Wim,

HENDERICKX Wim <wim.henderickx@alcatel-lucent.be> wrote:

> Ali,
>=20
> This is the same concern I raised in the emails before and will be=20
> difficult to solve if you let the remote PE(s) make a decision for=20
> multi-homing.

I thought I addressed all you MAC flush concerns. :-)=20

As I said, MAC flush capability in BGP based VPLS will be added to
address the concern you pointed out (with multiple VE IDs) and other
cases where MAC flush is needed.

Can you be specific on what is your open concern.=20

>=20
> Cheers,
> Wim

Thanks
Bhupesh


