
From nobody Wed Apr  1 13:21:05 2015
Return-Path: <ddolson@sandvine.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 7FF111A909B for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:21:02 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.091
X-Spam-Level: 
X-Spam-Status: No, score=0.091 tagged_above=-999 required=5 tests=[BAYES_50=0.8, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_LOW=-0.7, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4024eIhAtgBg for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:21:00 -0700 (PDT)
Received: from mail1.sandvine.com (Mail1.sandvine.com [64.7.137.134]) by ietfa.amsl.com (Postfix) with ESMTP id A169C1A909F for <aqm@ietf.org>; Wed,  1 Apr 2015 13:20:57 -0700 (PDT)
Received: from BLR-EXCHP-2.sandvine.com (192.168.196.172) by WTL-EXCHP-2.sandvine.com (192.168.194.177) with Microsoft SMTP Server (TLS) id 14.3.195.1; Wed, 1 Apr 2015 16:20:56 -0400
Received: from WTL-EXCHP-2.sandvine.com ([fe80::68ac:f071:19ff:3455]) by blr-exchp-2.sandvine.com ([fe80::6c6d:7108:c63c:9055%14]) with mapi id 14.03.0181.006; Wed, 1 Apr 2015 16:20:56 -0400
From: Dave Dolson <ddolson@sandvine.com>
To: "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: Turning PIE on and off
Thread-Index: AdBsuVurZp4RJEFxRD6VixLZQ7zUEw==
Date: Wed, 1 Apr 2015 20:20:56 +0000
Message-ID: <E8355113905631478EFF04F5AA706E9830BB0765@wtl-exchp-2.sandvine.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [192.168.200.63]
Content-Type: multipart/alternative; boundary="_000_E8355113905631478EFF04F5AA706E9830BB0765wtlexchp2sandvi_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/HC_8bsPgKQlAGRT9j2LjOzWvEBc>
Subject: [aqm] Turning PIE on and off
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 20:21:02 -0000

--_000_E8355113905631478EFF04F5AA706E9830BB0765wtlexchp2sandvi_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-5.1,
the text recommends turning PIE on only when buffer occupancy exceeds 1/3 o=
f the queue buffer size.

I think this recommendation unfortunately adds a new configuration paramete=
r, the "queue buffer size".  Which value should be chosen?
If PIE is defined properly, the behavior should be predictable regardless o=
f how large the packet memory is.
PIE should control average latency to target_delay even if the buffer size =
is theoretically infinite.


The pseudo-code has the same problem, defining QUEUE_SMALL =3D (1/3) * Buff=
er limit in bytes

Could the text be changed to specify some absolute size in bytes, or some a=
bsolute ratio of bytes/rate ?
Someone probably has good experimental outcomes, and they could check what =
size of "queue buffer size" they were using at the time.

Another approach could be to continuously measure arrival or departure rate=
, and enable PIE control when (qlen / rate) is some fraction of target_dela=
y.

Alternatively, don't recommend turning PIE on and off. Just always leave it=
 on. This requires updating 'p' even when there are few items in the queue.
You'd think that 'p' would go to zero when there is not congestion.
But I see that other aspects of the algorithm need an "off" state to work..=
.


David Dolson
Senior Software Architect, Sandvine Inc.


--_000_E8355113905631478EFF04F5AA706E9830BB0765wtlexchp2sandvi_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"=
>
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
span.EmailStyle17
	{mso-style-type:personal-compose;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
.MsoChpDefault
	{mso-style-type:export-only;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal">In <a href=3D"https://tools.ietf.org/html/draft-ietf=
-aqm-pie-00#section-5.1">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-5.1</a>,<o:p></o:=
p></p>
<p class=3D"MsoNormal">the text recommends turning PIE on only when buffer =
occupancy exceeds 1/3 of the queue buffer size.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">I think this recommendation unfortunately adds a new=
 configuration parameter, the &#8220;queue buffer size&#8221;.&nbsp; Which =
value should be chosen?<o:p></o:p></p>
<p class=3D"MsoNormal">If PIE is defined properly, the behavior should be p=
redictable regardless of how large the packet memory is.<o:p></o:p></p>
<p class=3D"MsoNormal">PIE should control average latency to target_delay e=
ven if the buffer size is theoretically infinite.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<pre><span style=3D"font-size:11.0pt;font-family:&quot;Calibri&quot;,&quot;=
sans-serif&quot;">The pseudo-code has the same problem, defining</span> QUE=
UE_SMALL =3D (1/3) * Buffer limit in bytes<o:p></o:p></pre>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Could the text be changed to specify some absolute s=
ize in bytes, or some absolute ratio of bytes/rate ?<o:p></o:p></p>
<p class=3D"MsoNormal">Someone probably has good experimental outcomes, and=
 they could check what size of &#8220;queue buffer size&#8221; they were us=
ing at the time.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Another approach could be to continuously measure ar=
rival or departure rate, and enable PIE control when (qlen / rate) is some =
fraction of target_delay.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Alternatively, don&#8217;t recommend turning PIE on =
and off. Just always leave it on. This requires updating &#8216;p&#8217; ev=
en when there are few items in the queue.<o:p></o:p></p>
<p class=3D"MsoNormal">You&#8217;d think that &#8216;p&#8217; would go to z=
ero when there is not congestion.<o:p></o:p></p>
<p class=3D"MsoNormal">But I see that other aspects of the algorithm need a=
n &#8220;off&#8221; state to work&#8230;<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">David Dolson<o:p></o:p></p>
<p class=3D"MsoNormal">Senior Software Architect, Sandvine Inc.<o:p></o:p><=
/p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
</div>
</body>
</html>

--_000_E8355113905631478EFF04F5AA706E9830BB0765wtlexchp2sandvi_--


From nobody Wed Apr  1 13:22:39 2015
Return-Path: <ddolson@sandvine.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 97ADF1A1B03 for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:22:37 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.909
X-Spam-Level: 
X-Spam-Status: No, score=-1.909 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Id1hMXT1sTJV for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:22:35 -0700 (PDT)
Received: from mail1.sandvine.com (mail1.sandvine.com [64.7.137.165]) by ietfa.amsl.com (Postfix) with ESMTP id 81F351A1A07 for <aqm@ietf.org>; Wed,  1 Apr 2015 13:22:34 -0700 (PDT)
Received: from WTL-EXCHP-2.sandvine.com ([fe80::68ac:f071:19ff:3455]) by WTL-EXCHP-3.sandvine.com ([::1]) with mapi id 14.03.0195.001; Wed, 1 Apr 2015 16:22:34 -0400
From: Dave Dolson <ddolson@sandvine.com>
To: "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: PIE departure rate estimation
Thread-Index: AdBsuZXDQTjdQGZHSkyfX7tnDoer/g==
Date: Wed, 1 Apr 2015 20:22:33 +0000
Message-ID: <E8355113905631478EFF04F5AA706E9830BB0786@wtl-exchp-2.sandvine.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [192.168.200.63]
Content-Type: multipart/alternative; boundary="_000_E8355113905631478EFF04F5AA706E9830BB0786wtlexchp2sandvi_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/3mFE-ZMjMowghPblmUcanbxZhIs>
Subject: [aqm] PIE departure rate estimation
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 20:22:37 -0000

--_000_E8355113905631478EFF04F5AA706E9830BB0786wtlexchp2sandvi_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3,

it says, "We only measure the departure rate when there are sufficient data=
 in the buffer"

Why can't departure rate be estimated regardless of queue size? Just count =
packets leaving over time? I'm wondering how to avoid the estimate getting =
stuck at the last value sampled when the queue had a certain quantity in it=
.

Section 4.2 cites Little's Law as "est_del =3D qlen/depart_rate",
but according to Wikipedia<http://en.wikipedia.org/wiki/Little%27s_law>, th=
e law uses arrival rate, not departure rate.
I don't know if it matters (I didn't read Little's proof), but this gives c=
redence to the suggestion in section 6 that the algorithm could use arrival=
 rate.
And I think it might be easier to measure when the queue has few items in i=
t.


David Dolson
Senior Software Architect, Sandvine Inc.


--_000_E8355113905631478EFF04F5AA706E9830BB0786wtlexchp2sandvi_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"=
>
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
span.EmailStyle17
	{mso-style-type:personal-compose;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
.MsoChpDefault
	{mso-style-type:export-only;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal">In <a href=3D"https://tools.ietf.org/html/draft-ietf=
-aqm-pie-00#section-4.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3</a>,<o:p></o:=
p></p>
<pre><span style=3D"font-size:11.0pt;font-family:&quot;Calibri&quot;,&quot;=
sans-serif&quot;">it says, </span>&#8220;We only measure the departure rate=
 when there are sufficient data in the buffer&#8221;<o:p></o:p></pre>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Why can&#8217;t departure rate be estimated regardle=
ss of queue size? Just count packets leaving over time? I&#8217;m wondering=
 how to avoid the estimate getting stuck at the last value sampled when the=
 queue had a certain quantity in it.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Section 4.2 cites Little&#8217;s Law as &#8220;est_d=
el =3D qlen/depart_rate&#8221;,<o:p></o:p></p>
<p class=3D"MsoNormal">but <a href=3D"http://en.wikipedia.org/wiki/Little%2=
7s_law">according to Wikipedia</a>, the law uses arrival rate, not departur=
e rate.<o:p></o:p></p>
<p class=3D"MsoNormal">I don&#8217;t know if it matters (I didn&#8217;t rea=
d Little&#8217;s proof), but this gives credence to the suggestion in secti=
on 6 that the algorithm could use arrival rate.<o:p></o:p></p>
<p class=3D"MsoNormal">And I think it might be easier to measure when the q=
ueue has few items in it.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">David Dolson<o:p></o:p></p>
<p class=3D"MsoNormal">Senior Software Architect, Sandvine Inc.<o:p></o:p><=
/p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
</div>
</body>
</html>

--_000_E8355113905631478EFF04F5AA706E9830BB0786wtlexchp2sandvi_--


From nobody Wed Apr  1 13:33:38 2015
Return-Path: <ddolson@sandvine.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 6BA001A908A for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:33:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.909
X-Spam-Level: 
X-Spam-Status: No, score=-1.909 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id r_ivhjAHurka for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:33:25 -0700 (PDT)
Received: from mail1.sandvine.com (mail1.sandvine.com [64.7.137.165]) by ietfa.amsl.com (Postfix) with ESMTP id 9B7B81A910F for <aqm@ietf.org>; Wed,  1 Apr 2015 13:31:12 -0700 (PDT)
Received: from WTL-EXCHP-2.sandvine.com ([fe80::68ac:f071:19ff:3455]) by WTL-EXCHP-3.sandvine.com ([::1]) with mapi id 14.03.0195.001; Wed, 1 Apr 2015 16:31:12 -0400
From: Dave Dolson <ddolson@sandvine.com>
To: "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: burst_allowance calculation in PIE
Thread-Index: AdBsusqqaI4pXvS0T+ioPH50ToijMg==
Date: Wed, 1 Apr 2015 20:31:11 +0000
Message-ID: <E8355113905631478EFF04F5AA706E9830BB07CA@wtl-exchp-2.sandvine.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [192.168.200.63]
Content-Type: multipart/alternative; boundary="_000_E8355113905631478EFF04F5AA706E9830BB07CAwtlexchp2sandvi_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/ObWafyoTyi5z2XqsifMS-dfDiX4>
Subject: [aqm] burst_allowance calculation in PIE
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 20:33:35 -0000

--_000_E8355113905631478EFF04F5AA706E9830BB07CAwtlexchp2sandvi_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-5.3,

It appears that burst_allowance is intended to allow a certain amount of tr=
affic to get queued before the algorithm kicks in. (Because the algorithm i=
sn't always running; it gets turned off when the queue gets near empty.)

However, burst_allowance isn't decremented according to packet arrival, rat=
her it is decremented each update period.
So whether 100 bytes or 1M bytes pass through, burst_allowance is decrement=
ed at the same rate.
Doesn't this seem wrong to not account for the quantity that is enqueued?

As for the wording itself,
"The burst allowance, noted by burst_allowance, is initialized to
max_burst. As long as burst_allowance is above zero, an incoming packet
will be enqueued bypassing the random drop process. During each update
instance, the value of burst_allowance is decremented by the update
period, Tupdate. When the congestion goes away, defined by us as p
equals to 0 and both the current and previous samples of estimated delay
are less than target_del, we reset burst_allowance to max_burst."

It says that burst_allowance is to be decremented "during each update insta=
nce". Does an "update instance" correspond to a "measurement cycle"? Can th=
is be made more precise? (I couldn't find "update instance" mentioned elsew=
here.)
>From the pseudo-code, it appears that it is decremented every 16ms; it only=
 gets set to BURST_ALLOWANCE at the transition from INACTIVE-->ACTIVE.
(In Pseudo-code, I think BURST_ALLOWANCE means max_burst and burst_count me=
ans burst_allowance.)



David Dolson
Senior Software Architect, Sandvine Inc.


--_000_E8355113905631478EFF04F5AA706E9830BB07CAwtlexchp2sandvi_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"=
>
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
span.EmailStyle17
	{mso-style-type:personal-compose;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
.MsoChpDefault
	{mso-style-type:export-only;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal">In <a href=3D"https://tools.ietf.org/html/draft-ietf=
-aqm-pie-00#section-5.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-5.3</a>, <o:p></o=
:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">It appears that burst_allowance is intended to allow=
 a certain amount of traffic to get queued before the algorithm kicks in. (=
Because the algorithm isn&#8217;t always running; it gets turned off when t=
he queue gets near empty.)<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">However, burst_allowance isn&#8217;t decremented acc=
ording to packet arrival, rather it is decremented each update period.<o:p>=
</o:p></p>
<p class=3D"MsoNormal">So whether 100 bytes or 1M bytes pass through, burst=
_allowance is decremented at the same rate.<o:p></o:p></p>
<p class=3D"MsoNormal">Doesn&#8217;t this seem wrong to not account for the=
 quantity that is enqueued?<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">As for the wording itself,<o:p></o:p></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">&#8220;The burst allowance, noted by burst_allowance, is i=
nitialized to<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">max_burst. As long as burst_allowance is above zero, an in=
coming packet<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">will be enqueued bypassing the random drop process. During=
 each update<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">instance, the value of burst_allowance is decremented by t=
he update<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">period, Tupdate. When the congestion goes away, defined by=
 us as p<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">equals to 0 and both the current and previous samples of e=
stimated delay<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:10.0pt;font-family:&quot;Co=
urier New&quot;">are less than target_del, we reset burst_allowance to max_=
burst.&#8221;<o:p></o:p></span></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">It says that burst_allowance is to be decremented &#=
8220;during each update instance&#8221;. Does an &#8220;update instance&#82=
21; correspond to a &#8220;measurement cycle&#8221;? Can this be made more =
precise? (I couldn&#8217;t find &#8220;update instance&#8221; mentioned els=
ewhere.)<o:p></o:p></p>
<p class=3D"MsoNormal">From the pseudo-code, it appears that it is decremen=
ted every 16ms; it only gets set to BURST_ALLOWANCE at the transition from =
INACTIVE--&gt;ACTIVE.<o:p></o:p></p>
<p class=3D"MsoNormal">(In Pseudo-code, I think BURST_ALLOWANCE means max_b=
urst and burst_count means burst_allowance.)<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">David Dolson<o:p></o:p></p>
<p class=3D"MsoNormal">Senior Software Architect, Sandvine Inc.<o:p></o:p><=
/p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
</div>
</body>
</html>

--_000_E8355113905631478EFF04F5AA706E9830BB07CAwtlexchp2sandvi_--


From nobody Wed Apr  1 13:51:25 2015
Return-Path: <ddolson@sandvine.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 0D9981A9165 for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:51:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.61
X-Spam-Level: 
X-Spam-Status: No, score=-2.61 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_LOW=-0.7, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id WX0dPOl2460m for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 13:51:22 -0700 (PDT)
Received: from mail1.sandvine.com (Mail1.sandvine.com [64.7.137.134]) by ietfa.amsl.com (Postfix) with ESMTP id DDFDF1A9114 for <aqm@ietf.org>; Wed,  1 Apr 2015 13:51:21 -0700 (PDT)
Received: from BLR-EXCHP-2.sandvine.com (192.168.196.172) by WTL-EXCHP-2.sandvine.com (192.168.194.177) with Microsoft SMTP Server (TLS) id 14.3.195.1; Wed, 1 Apr 2015 16:51:21 -0400
Received: from WTL-EXCHP-2.sandvine.com ([fe80::68ac:f071:19ff:3455]) by blr-exchp-2.sandvine.com ([fe80::6c6d:7108:c63c:9055%14]) with mapi id 14.03.0181.006; Wed, 1 Apr 2015 16:51:20 -0400
From: Dave Dolson <ddolson@sandvine.com>
To: "Scheffenegger, Richard" <rs@netapp.com>, Greg White <g.white@CableLabs.com>, Mikael Abrahamsson <swmike@swm.pp.se>, "Szilveszter Nadas" <Szilveszter.Nadas@ericsson.com>
Thread-Topic: [aqm] adoption of draft-white-aqm-docsis-pie-01
Thread-Index: AdBR0qmLD+JlLAXWSZym1tRD+kojwQXBCieAAIs1spA=
Date: Wed, 1 Apr 2015 20:51:19 +0000
Message-ID: <E8355113905631478EFF04F5AA706E9830BB08C1@wtl-exchp-2.sandvine.com>
References: <96cffe6791f54a67ae6b6359de6a3d6e@hioexcmbx05-prd.hq.netapp.com> <e9e3d9dc0435474283180936c371e0db@hioexcmbx05-prd.hq.netapp.com>
In-Reply-To: <e9e3d9dc0435474283180936c371e0db@hioexcmbx05-prd.hq.netapp.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [192.168.200.63]
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/9kUa496it8jcXJvhi7vCIWJpUFU>
Cc: "'aqm@ietf.org'" <aqm@ietf.org>, "aqm-chairs@tools.ietf.org" <aqm-chairs@tools.ietf.org>
Subject: Re: [aqm] adoption of draft-white-aqm-docsis-pie-01
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 20:51:24 -0000

I'm also reviewing draft-ietf-aqm-docsis-pie-00
(https://tools.ietf.org/html/draft-ietf-aqm-docsis-pie-00)=20

In section 3.4, an exponential delay aspect of PIE is mentioned. I just rea=
d https://tools.ietf.org/html/draft-ietf-aqm-pie-00 and saw no mention of t=
his exponential delay. Did I miss it, or is it a mistake for the docsis doc=
ument to reference this?


draft-ietf-aqm-docsis-pie-00 documents enabling and disabling of PIE contro=
l. Is enable/disable used with docsis-pie, or is it always on?


Otherwise, the document seems to contain enough information to document doc=
sis-pie.
The algorithm in section A.2 has quite a few aspects that are not explained=
, but it seems precise.
(I don't have the knowledge to comment on whether it is accurate.)


-Dave


-----Original Message-----
From: Scheffenegger, Richard [mailto:rs@netapp.com]=20
Sent: Friday, March 27, 2015 5:53 PM
To: Greg White; Mikael Abrahamsson; Dave Dolson; Szilveszter Nadas
Cc: aqm-chairs@tools.ietf.org; 'aqm@ietf.org'
Subject: RE: [aqm] adoption of draft-white-aqm-docsis-pie-01

Hi group,

as there haven't been any objections, but some indications of support on th=
e list, and based on the responses in the IETF92 meeting in Dallas, we chai=
rs think this document can be adopted as WG-item at this time.


Greg, can you please upload the most recent version as draft-ietf-aqm-docsi=
s-pie-00?

Also, as mentioned in the meeting, and to make true of my promise, I would =
like to invite the following individuals to review this draft, once the upd=
ated version becomes available.

Mikael Abrahamsson
Dave Dolson
Szilveszter N=E1das

Mostly everybody else who has commented during the meeting is already assig=
ned to other documents (Nobody will be left out :)

Thanks,
  Richard (co-chair)



> -----Original Message-----
> From: aqm [mailto:aqm-bounces@ietf.org] On Behalf Of Scheffenegger,
> Richard
> Sent: Donnerstag, 26. Februar 2015 08:45
> To: Greg White; 'aqm@ietf.org'
> Cc: Rong Pan (ropan); aqm-chairs@tools.ietf.org
> Subject: Re: [aqm] adoption of draft-white-aqm-docsis-pie-01
>=20
> Hi Greg, group,
>=20
>=20
>=20
> Regarding the adoption call, this is something we can actually (and
> should) start on the list.
>=20
> We can confirm that during the Dallas meeting in the room, but even befor=
e
> that, we'd like to get responses on the list now...
>=20
> Perhaps we have some volunteers to review this new version as well...
>=20
>=20
> Best regards,
>   Richard
>=20
>=20
>=20
> > -----Original Message-----
> > From: Greg White [mailto:g.white@CableLabs.com]
> > Sent: Donnerstag, 26. Februar 2015 01:13
> > To: aqm-chairs@tools.ietf.org
> > Cc: Rong Pan (ropan)
> > Subject: Re: [aqm] agenda for IETF 92 / Dallas
> >
> > Wes & Richard,
> >
> > Unfortunately I will not be at IETF92 in person, but will attend
> remotely.
> >  For draft-white-aqm-docsis-pie, Rong & I updated it in January and
> > included a new appendix to give a change log, which reads:
> >
> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> > Appendix B.  Change Log
> > B.1.  Since draft-white-aqm-docsis-pie-01
> >
> >    Added Change Log.
> >
> >    Removed discussion of Packet drop de-randomization, Enhanced burst
> >    protection, and 16ms update interval, as these are now included in
> >    [I-D.ietf-aqm-pie].
> >
> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> >
> >
> > Regarding WG adoption, I saw some support from Lars (and no objections)
> to
> > adopting it on an Informational track.   Will you do an official
> adoption
> > call at the Dallas meeting?
> >
> > -Greg
> >
> >
> > On 2/23/15, 8:40 PM, "Wesley Eddy" <wes@mti-systems.com> wrote:
> >
> > >Greetings AQMers!  We requested a short AQM meeting slot at the
> > >upcoming IETF 92 meeting in Dallas, and we received a 1-hour slot on
> Tuesday:
> > >https://datatracker.ietf.org/meeting/92/agenda.html
> > >
> > >Since this is not a lot of time, we'll need to prioritize the work
> > >that is discussed to what requires the face-to-face time in order to
> > >make progress on.
> > >
> > >For active draft editors, please let us know via this list or
> > >aqm-chairs@tools.ietf.org what you think your meeting time needs are,
> > >so we can put together an agenda.  If you don't need meeting time
> > >now, or would like to use an interim telecon to assist in gathering
> > >feedback, please also let us know.
> > >
> > >For others, please read the drafts, comment on this mailing list, and
> > >help us to review and complete them:
> > >https://datatracker.ietf.org/wg/aqm/documents/
> > >
> > >Thanks in advance!
> > >
> > >--
> > >Wes Eddy
> > >MTI Systems
> > >
> > >_______________________________________________
> > >aqm mailing list
> > >aqm@ietf.org
> > >https://www.ietf.org/mailman/listinfo/aqm
>=20
> _______________________________________________
> aqm mailing list
> aqm@ietf.org
> https://www.ietf.org/mailman/listinfo/aqm


From nobody Wed Apr  1 14:23:17 2015
Return-Path: <ropan@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id E71471A1A27 for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 14:23:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -14.51
X-Spam-Level: 
X-Spam-Status: No, score=-14.51 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id V46HxfTLqdFH for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 14:23:03 -0700 (PDT)
Received: from rcdn-iport-2.cisco.com (rcdn-iport-2.cisco.com [173.37.86.73]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 3A2941A1B23 for <aqm@ietf.org>; Wed,  1 Apr 2015 14:23:00 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=9325; q=dns/txt; s=iport; t=1427923381; x=1429132981; h=from:to:subject:date:message-id:mime-version; bh=xYuADPF2rw4nLJqMVahjClkKoIDUolkcabmurP7NjbY=; b=J58sAD5097z3XfCmTrdEQ+DDBifY8e8eUSoGjK1IfZWVkZ07CZi/rXam h/1oV7TNHKKHuk1kKFHGbItzKQ+xwNixK217Xgm/e3Kgejob4QUUzS/mR O7Iz9bjXhXbal0pUU2zV6fS3S5paCfwB9AUje6NYsOq/EV9+tne4YuqI0 A=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0A5BgD3YBxV/4YNJK1cgkNDUlwFw12CAIVzgUJMAQEBAQEBfYQUAQIELV4BCA4DAwECKDkUCQoEARKILgENzmIBAQEBBgEBAQEBAQEbiymEaA2EOAWQY4NyhgSBHYMyglyNFSKDbm8BgUN/AQEB
X-IronPort-AV: E=Sophos;i="5.11,506,1422921600";  d="scan'208,217";a="408856398"
Received: from alln-core-12.cisco.com ([173.36.13.134]) by rcdn-iport-2.cisco.com with ESMTP; 01 Apr 2015 21:23:00 +0000
Received: from xhc-aln-x07.cisco.com (xhc-aln-x07.cisco.com [173.36.12.81]) by alln-core-12.cisco.com (8.14.5/8.14.5) with ESMTP id t31LMxTK026813 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Wed, 1 Apr 2015 21:22:59 GMT
Received: from xmb-aln-x15.cisco.com ([169.254.9.109]) by xhc-aln-x07.cisco.com ([173.36.12.81]) with mapi id 14.03.0195.001; Wed, 1 Apr 2015 16:22:59 -0500
From: "Rong Pan (ropan)" <ropan@cisco.com>
To: Dave Dolson <ddolson@sandvine.com>, "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: [aqm] PIE departure rate estimation
Thread-Index: AQHQbMIGW+ot0iZ1OU6OZ2/eKhmRQQ==
Date: Wed, 1 Apr 2015 21:22:58 +0000
Message-ID: <D141ACCB.D99D%ropan@cisco.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
user-agent: Microsoft-MacOutlook/14.4.7.141117
x-originating-ip: [171.70.247.185]
Content-Type: multipart/alternative; boundary="_000_D141ACCBD99Dropanciscocom_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/Rsjg7vauvTF6Ueh2a_e9a5KEKl4>
Subject: Re: [aqm] PIE departure rate estimation
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 21:23:08 -0000

--_000_D141ACCBD99Dropanciscocom_
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

Dave,

Thanks for the valuable comments. We will see how we can incorporate them. =
For the comments below, please see inline.

Regards,

Rong


From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 1:22 PM
To: "aqm@ietf.org<mailto:aqm@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: [aqm] PIE departure rate estimation

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3,

it says, =93We only measure the departure rate when there are sufficient da=
ta in the buffer=94

Why can=92t departure rate be estimated regardless of queue size? Just coun=
t packets leaving over time? I=92m wondering how to avoid the estimate gett=
ing stuck at the last value sampled when the queue had a certain quantity i=
n it.

>>>>>>>>>>>>RP: We need to make sure that there are enough data in the queu=
e to guarantee a rate sample. The reason is that if a queue is empty from t=
ime to time, we can't measure its true draining rate. The time when the que=
ue empty should be cut out of the drain time calculation. For simplicity, i=
t is better to make sure we have enough data in the queue to ensure accurat=
e rate measurement sample.

Section 4.2 cites Little=92s Law as =93est_del =3D qlen/depart_rate=94,
but according to Wikipedia<http://en.wikipedia.org/wiki/Little%27s_law>, th=
e law uses arrival rate, not departure rate.
I don=92t know if it matters (I didn=92t read Little=92s proof), but this g=
ives credence to the suggestion in section 6 that the algorithm could use a=
rrival rate.
And I think it might be easier to measure when the queue has few items in i=
t.

>>>>>>>>>>>>RP: what you mentioned is to calculate average number of custom=
ers in a system using the arrival. http://en.wikipedia.org/wiki/Little's_la=
w Wikipedia also mentions that mean responseTime =3D MeanNumberInSystem / M=
eanThroughput. What we measure is the mean response time (latency). Hence, =
it is correct in our draft.

David Dolson
Senior Software Architect, Sandvine Inc.


--_000_D141ACCBD99Dropanciscocom_
Content-Type: text/html; charset="Windows-1252"
Content-ID: <7B8AB95D14D1CF4593BA33002A621955@emea.cisco.com>
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3DWindows-1=
252">
</head>
<body style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-lin=
e-break: after-white-space; color: rgb(0, 0, 0); font-size: 16px; font-fami=
ly: Calibri, sans-serif; ">
<div>Dave,</div>
<div><br>
</div>
<div>Thanks for the valuable comments. We will see how we can incorporate t=
hem. For the comments below, please see inline.</div>
<div><br>
</div>
<div>Regards,</div>
<div><br>
</div>
<div>Rong</div>
<div><br>
</div>
<div><br>
</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div style=3D"font-family:Calibri; font-size:11pt; text-align:left; color:b=
lack; BORDER-BOTTOM: medium none; BORDER-LEFT: medium none; PADDING-BOTTOM:=
 0in; PADDING-LEFT: 0in; PADDING-RIGHT: 0in; BORDER-TOP: #b5c4df 1pt solid;=
 BORDER-RIGHT: medium none; PADDING-TOP: 3pt">
<span style=3D"font-weight:bold">From: </span>Dave Dolson &lt;<a href=3D"ma=
ilto:ddolson@sandvine.com">ddolson@sandvine.com</a>&gt;<br>
<span style=3D"font-weight:bold">Date: </span>Wednesday, April 1, 2015 1:22=
 PM<br>
<span style=3D"font-weight:bold">To: </span>&quot;<a href=3D"mailto:aqm@iet=
f.org">aqm@ietf.org</a>&quot; &lt;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.=
org</a>&gt;<br>
<span style=3D"font-weight:bold">Subject: </span>[aqm] PIE departure rate e=
stimation<br>
</div>
<div><br>
</div>
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
span.EmailStyle17
	{mso-style-type:personal-compose;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
.MsoChpDefault
	{mso-style-type:export-only;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal">In <a href=3D"https://tools.ietf.org/html/draft-ietf=
-aqm-pie-00#section-4.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3</a>,<o:p></o:=
p></p>
<pre><span style=3D"font-size: 11pt; font-family: Calibri, sans-serif; ">it=
 says, </span>=93We only measure the departure rate when there are sufficie=
nt data in the buffer=94<o:p></o:p></pre>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Why can=92t departure rate be estimated regardless o=
f queue size? Just count packets leaving over time? I=92m wondering how to =
avoid the estimate getting stuck at the last value sampled when the queue h=
ad a certain quantity in it.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
</div>
</div>
</div>
</span>
<div>&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: We need to make su=
re that there are enough data in the queue to guarantee a rate sample. The =
reason is that if a queue is empty from time to time, we can't measure its =
true draining rate. The time when the queue empty should be cut out
 of the drain time calculation. For simplicity, it is better to make sure w=
e have enough data in the queue to ensure accurate rate measurement sample.=
</div>
<div><br>
</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal">Section 4.2 cites Little=92s Law as =93est_del =3D q=
len/depart_rate=94,<o:p></o:p></p>
<p class=3D"MsoNormal">but <a href=3D"http://en.wikipedia.org/wiki/Little%2=
7s_law">according to Wikipedia</a>, the law uses arrival rate, not departur=
e rate.<o:p></o:p></p>
<p class=3D"MsoNormal">I don=92t know if it matters (I didn=92t read Little=
=92s proof), but this gives credence to the suggestion in section 6 that th=
e algorithm could use arrival rate.<o:p></o:p></p>
<p class=3D"MsoNormal">And I think it might be easier to measure when the q=
ueue has few items in it.<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
</div>
</div>
</div>
</span>
<div>&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: what you mentioned=
 is to calculate average number of customers in a system using the arrival.
<a href=3D"http://en.wikipedia.org/wiki/Little's_law">http://en.wikipedia.o=
rg/wiki/Little's_law</a>&nbsp;Wikipedia also mentions that mean responseTim=
e =3D&nbsp;<span style=3D"background-color: rgb(255, 255, 255); color: rgb(=
37, 37, 37); font-family: sans-serif; font-size: 14px; line-height: 22px; "=
>MeanNumberInSystem
 / MeanThroughput. What we measure is the mean response time (latency). Hen=
ce, it is correct in our draft.&nbsp;</span></div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">David Dolson<o:p></o:p></p>
<p class=3D"MsoNormal">Senior Software Architect, Sandvine Inc.<o:p></o:p><=
/p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
</div>
</div>
</div>
</span>
</body>
</html>

--_000_D141ACCBD99Dropanciscocom_--


From nobody Wed Apr  1 14:43:09 2015
Return-Path: <ddolson@sandvine.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 01CD01A1E0B for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 14:43:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.309
X-Spam-Level: 
X-Spam-Status: No, score=-1.309 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, J_CHICKENPOX_84=0.6, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 3cvDzAk-j3nP for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 14:43:04 -0700 (PDT)
Received: from mail1.sandvine.com (mail1.sandvine.com [64.7.137.165]) by ietfa.amsl.com (Postfix) with ESMTP id 9288E1A19F4 for <aqm@ietf.org>; Wed,  1 Apr 2015 14:43:04 -0700 (PDT)
Received: from WTL-EXCHP-2.sandvine.com ([fe80::68ac:f071:19ff:3455]) by WTL-EXCHP-3.sandvine.com ([::1]) with mapi id 14.03.0195.001; Wed, 1 Apr 2015 17:43:04 -0400
From: Dave Dolson <ddolson@sandvine.com>
To: "Rong Pan (ropan)" <ropan@cisco.com>, "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: [aqm] PIE departure rate estimation
Thread-Index: AQHQbMIGW+ot0iZ1OU6OZ2/eKhmRQZ04qyFA
Date: Wed, 1 Apr 2015 21:43:03 +0000
Message-ID: <E8355113905631478EFF04F5AA706E9830BB0BF4@wtl-exchp-2.sandvine.com>
References: <D141ACCB.D99D%ropan@cisco.com>
In-Reply-To: <D141ACCB.D99D%ropan@cisco.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [192.168.200.63]
Content-Type: multipart/alternative; boundary="_000_E8355113905631478EFF04F5AA706E9830BB0BF4wtlexchp2sandvi_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/wsAuH5iDyywdSdUUavXJ9s8tGnM>
Subject: Re: [aqm] PIE departure rate estimation
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 21:43:08 -0000

--_000_E8355113905631478EFF04F5AA706E9830BB0BF4wtlexchp2sandvi_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Inline [DD]

From: Rong Pan (ropan) [mailto:ropan@cisco.com]
Sent: Wednesday, April 01, 2015 5:23 PM
To: Dave Dolson; aqm@ietf.org
Subject: Re: [aqm] PIE departure rate estimation

Dave,

Thanks for the valuable comments. We will see how we can incorporate them. =
For the comments below, please see inline.

Regards,

Rong


From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 1:22 PM
To: "aqm@ietf.org<mailto:aqm@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: [aqm] PIE departure rate estimation

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3,

it says, "We only measure the departure rate when there are sufficient data=
 in the buffer"

Why can't departure rate be estimated regardless of queue size? Just count =
packets leaving over time? I'm wondering how to avoid the estimate getting =
stuck at the last value sampled when the queue had a certain quantity in it=
.

>>>>>>>>>>>>RP: We need to make sure that there are enough data in the queu=
e to guarantee a rate sample. The reason is that if a queue is empty from t=
ime to time, we can't measure its true draining rate. The time when the que=
ue empty should be cut out of the drain time calculation. For simplicity, i=
t is better to make sure we have enough data in the queue to ensure accurat=
e rate measurement sample.
[DD] I understand the rationale, but supposing p=3D0.1, but the queue has v=
ery few items in it, too few to obtain a rate calculation. What causes p to=
 go to zero, since an update is not permitted? It isn't obvious to me that =
the algorithm always stops for all input traffic patterns.


Section 4.2 cites Little's Law as "est_del =3D qlen/depart_rate",
but according to Wikipedia<http://en.wikipedia.org/wiki/Little%27s_law>, th=
e law uses arrival rate, not departure rate.
I don't know if it matters (I didn't read Little's proof), but this gives c=
redence to the suggestion in section 6 that the algorithm could use arrival=
 rate.
And I think it might be easier to measure when the queue has few items in i=
t.

>>>>>>>>>>>>RP: what you mentioned is to calculate average number of custom=
ers in a system using the arrival. http://en.wikipedia.org/wiki/Little's_la=
w Wikipedia also mentions that mean responseTime =3D MeanNumberInSystem / M=
eanThroughput. What we measure is the mean response time (latency). Hence, =
it is correct in our draft.
[DD] Arrival (post drop) and departure rates are almost never the same at a=
ny instant. I guess they must average to the same value. Practically, it mi=
ght not matter which is used, but I thought Little's law should be cited co=
rrectly. In Wikipedia the law is cited one way and the example is shown ano=
ther way. Furthermore there is this ominous comment: "When exploring Little=
's law and learning to trust it, be aware of the common mistakes of using a=
rrivals(work arriving) when throughput(work completed) is called for..."


David Dolson
Senior Software Architect, Sandvine Inc.


--_000_E8355113905631478EFF04F5AA706E9830BB0BF4wtlexchp2sandvi_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"=
>
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
	{font-family:Tahoma;
	panose-1:2 11 6 4 3 5 4 4 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
p.MsoAcetate, li.MsoAcetate, div.MsoAcetate
	{mso-style-priority:99;
	mso-style-link:"Balloon Text Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:8.0pt;
	font-family:"Tahoma","sans-serif";}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
span.EmailStyle19
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.BalloonTextChar
	{mso-style-name:"Balloon Text Char";
	mso-style-priority:99;
	mso-style-link:"Balloon Text";
	font-family:"Tahoma","sans-serif";}
span.EmailStyle22
	{mso-style-type:personal-reply;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
.MsoChpDefault
	{mso-style-type:export-only;
	font-size:10.0pt;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">Inline [DD]<o:p></o:p>=
</span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"font-size:10.0pt;font-family:&quot=
;Tahoma&quot;,&quot;sans-serif&quot;">From:</span></b><span style=3D"font-s=
ize:10.0pt;font-family:&quot;Tahoma&quot;,&quot;sans-serif&quot;"> Rong Pan=
 (ropan) [mailto:ropan@cisco.com]
<br>
<b>Sent:</b> Wednesday, April 01, 2015 5:23 PM<br>
<b>To:</b> Dave Dolson; aqm@ietf.org<br>
<b>Subject:</b> Re: [aqm] PIE departure rate estimation<o:p></o:p></span></=
p>
</div>
</div>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Dave,<o:p></o:p></span><=
/p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Thanks for the valuable =
comments. We will see how we can incorporate them. For the comments below, =
please see inline.<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Regards,<o:p></o:p></spa=
n></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Rong<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"color:black">From: </span></b><spa=
n style=3D"color:black">Dave Dolson &lt;<a href=3D"mailto:ddolson@sandvine.=
com">ddolson@sandvine.com</a>&gt;<br>
<b>Date: </b>Wednesday, April 1, 2015 1:22 PM<br>
<b>To: </b>&quot;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&quot; &lt=
;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&gt;<br>
<b>Subject: </b>[aqm] PIE departure rate estimation<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">In <a href=3D"https://to=
ols.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3</a>,<o:p></o:=
p></span></p>
<pre><span style=3D"font-size:11.0pt;font-family:&quot;Calibri&quot;,&quot;=
sans-serif&quot;;color:black">it says, </span><span style=3D"color:black">&=
#8220;We only measure the departure rate when there are sufficient data in =
the buffer&#8221;<o:p></o:p></span></pre>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">Why can&#8217;t departur=
e rate be estimated regardless of queue size? Just count packets leaving ov=
er time? I&#8217;m wondering how to avoid the estimate getting stuck at the=
 last value sampled when the queue had a certain
 quantity in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: We need to make sure that ther=
e are enough data in the queue to guarantee a rate sample. The reason is th=
at if a queue is empty from time to time, we can't measure its true drainin=
g
 rate. The time when the queue empty should be cut out of the drain time ca=
lculation. For simplicity, it is better to make sure we have enough data in=
 the queue to ensure accurate rate measurement sample.<o:p></o:p></span></p=
>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] I understand the =
rationale, but supposing p=3D0.1, but the queue has very few items in it, t=
oo few to obtain a rate calculation. What causes p to go to zero, since an =
update is not permitted? It isn&#8217;t obvious
 to me that the algorithm always stops for all input traffic patterns.<o:p>=
</o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Section 4.2 cites Little=
&#8217;s Law as &#8220;est_del =3D qlen/depart_rate&#8221;,<o:p></o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:black">but <a href=3D"http://en=
.wikipedia.org/wiki/Little%27s_law">
according to Wikipedia</a>, the law uses arrival rate, not departure rate.<=
o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">I don&#8217;t know if it=
 matters (I didn&#8217;t read Little&#8217;s proof), but this gives credenc=
e to the suggestion in section 6 that the algorithm could use arrival rate.=
<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">And I think it might be =
easier to measure when the queue has few items in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: what you mentioned is to calcu=
late average number of customers in a system using the arrival.
<a href=3D"http://en.wikipedia.org/wiki/Little's_law">http://en.wikipedia.o=
rg/wiki/Little's_law</a>&nbsp;Wikipedia also mentions that mean responseTim=
e =3D&nbsp;</span><span style=3D"font-size:10.5pt;font-family:&quot;Arial&q=
uot;,&quot;sans-serif&quot;;color:#252525;background:white">MeanNumberInSys=
tem
 / MeanThroughput. What we measure is the mean response time (latency). Hen=
ce, it is correct in our draft.&nbsp;</span><span style=3D"font-size:10.5pt=
;font-family:&quot;Arial&quot;,&quot;sans-serif&quot;;color:black;backgroun=
d:white"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] Arrival (post dro=
p) and departure rates are almost never the same at any instant. I guess th=
ey must average to the same value. Practically, it might not matter which i=
s used, but I thought Little&#8217;s law should
 be cited correctly. In Wikipedia the law is cited one way and the example =
is shown another way. Furthermore there is this ominous comment: &#8220;Whe=
n exploring Little&#8217;s law and learning to trust it, be aware of the co=
mmon mistakes of using arrivals(work arriving)
 when throughput(work completed) is called for&#8230;&#8221;<o:p></o:p></sp=
an></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">David Dolson<o:p></o:p><=
/span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">Senior Software Architec=
t, Sandvine Inc.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
</div>
</body>
</html>

--_000_E8355113905631478EFF04F5AA706E9830BB0BF4wtlexchp2sandvi_--


From nobody Wed Apr  1 15:27:41 2015
Return-Path: <ropan@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id BE3D51A87A1 for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 15:27:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -13.91
X-Spam-Level: 
X-Spam-Status: No, score=-13.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=0.001, J_CHICKENPOX_84=0.6, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id tqspeCDbGbbk for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 15:27:37 -0700 (PDT)
Received: from rcdn-iport-2.cisco.com (rcdn-iport-2.cisco.com [173.37.86.73]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 1E48C1A0451 for <aqm@ietf.org>; Wed,  1 Apr 2015 15:27:37 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=17680; q=dns/txt; s=iport; t=1427927258; x=1429136858; h=from:to:subject:date:message-id:references:in-reply-to: mime-version; bh=dFeKMjphRSgRfRe22tSOcjJuNxZzMW2H1vZhyqbsXcw=; b=LH5rudNU+Dmrf+MlQ96dRYehh86GR1+/U5XjeK4yQzSsvEaN/p69LmWz hfMr8/w/OEzTWwOcV6MPl04dFe178djZ7VRQcMVfnNdeLPUYQWdE3RsrZ UCdI4G0ZR9eYC4VQqEzmrHmsLNrnb9q5EWTnHHsMxoJd/ZLWYg8Bg7pbL 8=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0A7BgAUcBxV/5FdJa1cgkNDUlwFw12CAIVzAoFBTAEBAQEBAX2EFAEBAQQtQBwCAQgOAwMBAQEoBzIUCQgCBAESiC4BDc5VAQEBAQEBAQEBAQEBAQEBAQEBAQEBF4sphGgNCgGELQWQY4NyhgSBHYMyglyGGoMzg0gig25vAYFDfwEBAQ
X-IronPort-AV: E=Sophos;i="5.11,506,1422921600";  d="scan'208,217";a="408874123"
Received: from rcdn-core-9.cisco.com ([173.37.93.145]) by rcdn-iport-2.cisco.com with ESMTP; 01 Apr 2015 22:27:37 +0000
Received: from xhc-aln-x04.cisco.com (xhc-aln-x04.cisco.com [173.36.12.78]) by rcdn-core-9.cisco.com (8.14.5/8.14.5) with ESMTP id t31MRaR4003054 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Wed, 1 Apr 2015 22:27:36 GMT
Received: from xmb-aln-x15.cisco.com ([169.254.9.109]) by xhc-aln-x04.cisco.com ([173.36.12.78]) with mapi id 14.03.0195.001; Wed, 1 Apr 2015 17:27:36 -0500
From: "Rong Pan (ropan)" <ropan@cisco.com>
To: Dave Dolson <ddolson@sandvine.com>, "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: [aqm] PIE departure rate estimation
Thread-Index: AQHQbMIGW+ot0iZ1OU6OZ2/eKhmRQZ04qyFA///v6gA=
Date: Wed, 1 Apr 2015 22:27:35 +0000
Message-ID: <D141BCEA.D9C1%ropan@cisco.com>
References: <D141ACCB.D99D%ropan@cisco.com> <E8355113905631478EFF04F5AA706E9830BB0BF4@wtl-exchp-2.sandvine.com>
In-Reply-To: <E8355113905631478EFF04F5AA706E9830BB0BF4@wtl-exchp-2.sandvine.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
user-agent: Microsoft-MacOutlook/14.4.7.141117
x-originating-ip: [171.70.247.185]
Content-Type: multipart/alternative; boundary="_000_D141BCEAD9C1ropanciscocom_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/r6zH-RFTJYVhZn44PeMTcodGUhY>
Subject: Re: [aqm] PIE departure rate estimation
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 01 Apr 2015 22:27:39 -0000

--_000_D141BCEAD9C1ropanciscocom_
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

Please see inline=85

Rong

From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 2:43 PM
To: rong <ropan@cisco.com<mailto:ropan@cisco.com>>, "aqm@ietf.org<mailto:aq=
m@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: RE: [aqm] PIE departure rate estimation

Inline [DD]

From: Rong Pan (ropan) [mailto:ropan@cisco.com]
Sent: Wednesday, April 01, 2015 5:23 PM
To: Dave Dolson; aqm@ietf.org<mailto:aqm@ietf.org>
Subject: Re: [aqm] PIE departure rate estimation

Dave,

Thanks for the valuable comments. We will see how we can incorporate them. =
For the comments below, please see inline.

Regards,

Rong


From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 1:22 PM
To: "aqm@ietf.org<mailto:aqm@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: [aqm] PIE departure rate estimation

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3,

it says, =93We only measure the departure rate when there are sufficient da=
ta in the buffer=94

Why can=92t departure rate be estimated regardless of queue size? Just coun=
t packets leaving over time? I=92m wondering how to avoid the estimate gett=
ing stuck at the last value sampled when the queue had a certain quantity i=
n it.

>>>>>>>>>>>>RP: We need to make sure that there are enough data in the queu=
e to guarantee a rate sample. The reason is that if a queue is empty from t=
ime to time, we can't measure its true draining rate. The time when the que=
ue empty should be cut out of the drain time calculation. For simplicity, i=
t is better to make sure we have enough data in the queue to ensure accurat=
e rate measurement sample.
[DD] I understand the rationale, but supposing p=3D0.1, but the queue has v=
ery few items in it, too few to obtain a rate calculation. What causes p to=
 go to zero, since an update is not permitted? It isn=92t obvious to me tha=
t the algorithm always stops for all input traffic patterns.

>>>>>>>>>>>>>>RP: There are two calculations in the PIE algorithm: one is f=
or p update and one is for calculate the rate. They are independent from ea=
ch other. If there are not enough packets in the buffer, we simply don't up=
date depart rate estimation,  instead we use the previous value. However, t=
he drop probability calculation is always updated. Hence, the "p" in your e=
xample will go down to 0.


Section 4.2 cites Little=92s Law as =93est_del =3D qlen/depart_rate=94,
but according to Wikipedia<http://en.wikipedia.org/wiki/Little%27s_law>, th=
e law uses arrival rate, not departure rate.
I don=92t know if it matters (I didn=92t read Little=92s proof), but this g=
ives credence to the suggestion in section 6 that the algorithm could use a=
rrival rate.
And I think it might be easier to measure when the queue has few items in i=
t.

>>>>>>>>>>>>RP: what you mentioned is to calculate average number of custom=
ers in a system using the arrival. http://en.wikipedia.org/wiki/Little's_la=
w Wikipedia also mentions that mean responseTime =3D MeanNumberInSystem / M=
eanThroughput. What we measure is the mean response time (latency). Hence, =
it is correct in our draft.
[DD] Arrival (post drop) and departure rates are almost never the same at a=
ny instant. I guess they must average to the same value. Practically, it mi=
ght not matter which is used, but I thought Little=92s law should be cited =
correctly. In Wikipedia the law is cited one way and the example is shown a=
nother way. Furthermore there is this ominous comment: =93When exploring Li=
ttle=92s law and learning to trust it, be aware of the common mistakes of u=
sing arrivals(work arriving) when throughput(work completed) is called for=
=85=94

>>>>>>>>>>>>>RP: I understand the concern. But I think we have used it in t=
he right way. Arrival is related to how many customers/packets in a system.=
 However, given #of customers/packets in a system, the wait time is directl=
y related to the service time (inverse of throughput), not arrival rate any=
 more.

David Dolson
Senior Software Architect, Sandvine Inc.


--_000_D141BCEAD9C1ropanciscocom_
Content-Type: text/html; charset="Windows-1252"
Content-ID: <C6F98555E3DE8045BDB4A0903DC30C08@emea.cisco.com>
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3DWindows-1=
252">
</head>
<body style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-lin=
e-break: after-white-space; color: rgb(0, 0, 0); font-size: 16px; font-fami=
ly: Calibri, sans-serif; ">
<div>Please see inline=85</div>
<div><br>
</div>
<div>Rong</div>
<div><br>
</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div style=3D"font-family:Calibri; font-size:11pt; text-align:left; color:b=
lack; BORDER-BOTTOM: medium none; BORDER-LEFT: medium none; PADDING-BOTTOM:=
 0in; PADDING-LEFT: 0in; PADDING-RIGHT: 0in; BORDER-TOP: #b5c4df 1pt solid;=
 BORDER-RIGHT: medium none; PADDING-TOP: 3pt">
<span style=3D"font-weight:bold">From: </span>Dave Dolson &lt;<a href=3D"ma=
ilto:ddolson@sandvine.com">ddolson@sandvine.com</a>&gt;<br>
<span style=3D"font-weight:bold">Date: </span>Wednesday, April 1, 2015 2:43=
 PM<br>
<span style=3D"font-weight:bold">To: </span>rong &lt;<a href=3D"mailto:ropa=
n@cisco.com">ropan@cisco.com</a>&gt;, &quot;<a href=3D"mailto:aqm@ietf.org"=
>aqm@ietf.org</a>&quot; &lt;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a=
>&gt;<br>
<span style=3D"font-weight:bold">Subject: </span>RE: [aqm] PIE departure ra=
te estimation<br>
</div>
<div><br>
</div>
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
	{font-family:Tahoma;
	panose-1:2 11 6 4 3 5 4 4 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
p.MsoAcetate, li.MsoAcetate, div.MsoAcetate
	{mso-style-priority:99;
	mso-style-link:"Balloon Text Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:8.0pt;
	font-family:"Tahoma","sans-serif";}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
span.EmailStyle19
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.BalloonTextChar
	{mso-style-name:"Balloon Text Char";
	mso-style-priority:99;
	mso-style-link:"Balloon Text";
	font-family:"Tahoma","sans-serif";}
span.EmailStyle22
	{mso-style-type:personal-reply;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
.MsoChpDefault
	{mso-style-type:export-only;
	font-size:10.0pt;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">Inline [DD]<o:p></o:p>=
</span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: Taho=
ma, sans-serif; ">From:</span></b><span style=3D"font-size: 10pt; font-fami=
ly: Tahoma, sans-serif; "> Rong Pan (ropan) [<a href=3D"mailto:ropan@cisco.=
com">mailto:ropan@cisco.com</a>]
<br>
<b>Sent:</b> Wednesday, April 01, 2015 5:23 PM<br>
<b>To:</b> Dave Dolson; <a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a><br=
>
<b>Subject:</b> Re: [aqm] PIE departure rate estimation<o:p></o:p></span></=
p>
</div>
</div>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Dave,<o:p></o:p></span><=
/p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Thanks for the valuable =
comments. We will see how we can incorporate them. For the comments below, =
please see inline.<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Regards,<o:p></o:p></spa=
n></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Rong<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"color:black">From: </span></b><spa=
n style=3D"color:black">Dave Dolson &lt;<a href=3D"mailto:ddolson@sandvine.=
com">ddolson@sandvine.com</a>&gt;<br>
<b>Date: </b>Wednesday, April 1, 2015 1:22 PM<br>
<b>To: </b>&quot;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&quot; &lt=
;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&gt;<br>
<b>Subject: </b>[aqm] PIE departure rate estimation<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">In <a href=3D"https://to=
ols.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3</a>,<o:p></o:=
p></span></p>
<pre><span style=3D"font-size: 11pt; font-family: Calibri, sans-serif; colo=
r: black; ">it says, </span><span style=3D"color:black">=93We only measure =
the departure rate when there are sufficient data in the buffer=94<o:p></o:=
p></span></pre>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">Why can=92t departure ra=
te be estimated regardless of queue size? Just count packets leaving over t=
ime? I=92m wondering how to avoid the estimate getting stuck at the last va=
lue sampled when the queue had a certain
 quantity in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: We need to make sure that ther=
e are enough data in the queue to guarantee a rate sample. The reason is th=
at if a queue is empty from time to time, we can't measure its true drainin=
g
 rate. The time when the queue empty should be cut out of the drain time ca=
lculation. For simplicity, it is better to make sure we have enough data in=
 the queue to ensure accurate rate measurement sample.<o:p></o:p></span></p=
>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] I understand the =
rationale, but supposing p=3D0.1, but the queue has very few items in it, t=
oo few to obtain a rate calculation. What causes p to go to zero, since an =
update is not permitted? It isn=92t obvious
 to me that the algorithm always stops for all input traffic patterns.</spa=
n></p>
</div>
</div>
</div>
</div>
</span>
<div><br>
</div>
<div>&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: There are =
two calculations in the PIE algorithm: one is for p update and one is for c=
alculate the rate. They are independent from each other. If there are not e=
nough packets in the buffer, we simply don't update depart rate estimation,
 &nbsp;instead we use the previous value. However, the drop probability cal=
culation is always updated. Hence, the &quot;p&quot; in your example will g=
o down to 0.</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<div>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Section 4.2 cites Little=
=92s Law as =93est_del =3D qlen/depart_rate=94,<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">but <a href=3D"http://en=
.wikipedia.org/wiki/Little%27s_law">
according to Wikipedia</a>, the law uses arrival rate, not departure rate.<=
o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">I don=92t know if it mat=
ters (I didn=92t read Little=92s proof), but this gives credence to the sug=
gestion in section 6 that the algorithm could use arrival rate.<o:p></o:p><=
/span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">And I think it might be =
easier to measure when the queue has few items in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: what you mentioned is to calcu=
late average number of customers in a system using the arrival.
<a href=3D"http://en.wikipedia.org/wiki/Little's_law">http://en.wikipedia.o=
rg/wiki/Little's_law</a>&nbsp;Wikipedia also mentions that mean responseTim=
e =3D&nbsp;</span><span style=3D"font-size: 10.5pt; font-family: Arial, san=
s-serif; color: rgb(37, 37, 37); background-color: white; background-positi=
on: initial initial; background-repeat: initial initial; ">MeanNumberInSyst=
em
 / MeanThroughput. What we measure is the mean response time (latency). Hen=
ce, it is correct in our draft.&nbsp;</span><span style=3D"font-size: 10.5p=
t; font-family: Arial, sans-serif; color: black; background-color: white; b=
ackground-position: initial initial; background-repeat: initial initial; ">=
<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] Arrival (post dro=
p) and departure rates are almost never the same at any instant. I guess th=
ey must average to the same value. Practically, it might not matter which i=
s used, but I thought Little=92s law should
 be cited correctly. In Wikipedia the law is cited one way and the example =
is shown another way. Furthermore there is this ominous comment: =93When ex=
ploring Little=92s law and learning to trust it, be aware of the common mis=
takes of using arrivals(work arriving)
 when throughput(work completed) is called for=85=94<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
</div>
</div>
</div>
</div>
</span>
<div>&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: I understand t=
he concern. But I think we have used it in the right way. Arrival is relate=
d to how many customers/packets in a system. However, given #of customers/p=
ackets in a system, the wait time is directly related to the service time
 (inverse of throughput), not arrival rate any more.</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">David Dolson<o:p></o:p><=
/span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">Senior Software Architec=
t, Sandvine Inc.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
</div>
</div>
</div>
</span>
</body>
</html>

--_000_D141BCEAD9C1ropanciscocom_--


From nobody Wed Apr  1 20:34:19 2015
Return-Path: <fred@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 324331A0378 for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 20:34:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -112.612
X-Spam-Level: 
X-Spam-Status: No, score=-112.612 tagged_above=-999 required=5 tests=[BAYES_40=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5, USER_IN_WHITELIST=-100] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 04iMVRU_8o8s for <aqm@ietfa.amsl.com>; Wed,  1 Apr 2015 20:34:16 -0700 (PDT)
Received: from rcdn-iport-5.cisco.com (rcdn-iport-5.cisco.com [173.37.86.76]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id BF7671B2A0F for <aqm@ietf.org>; Wed,  1 Apr 2015 20:34:13 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=9015; q=dns/txt; s=iport; t=1427945653; x=1429155253; h=from:to:cc:subject:date:message-id:mime-version; bh=38dZzwCMg9Xr5M5tG97kIJtfXZoMQibkVmwuLJHg7wg=; b=lDkAQOLd8+MWVs+2LDaS34FTNW3HKwIveZzaw9cEabcYJ3/N2lJC75uv MKiBvZCHFqKLJQlGe4rDUrlvo8/DZHyBWy/MJyS5lJCAP3VX7YBo9WHat wTvx5riyK7HprxNPWL9wXjVeNci9PbGGGndYRXWQoXWrL5zapvSCwYDl6 0=;
X-Files: signature.asc : 487
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0DmBQDEtxxV/5xdJa1cgwhSXQSDEL9+gk+Fc4FMTAEBAQEBAX6EJSNWEgFKAjQfCAQOE4ghDbRemCwBAQEBAQEBAQEBAQEBAQEBAQEBGYoqhRUKBwECT4JvL4EWBZBkgWyBMlSGBJRBIoF/H4FQgXEJFwQefwEBAQ
X-IronPort-AV: E=Sophos;i="5.11,507,1422921600";  d="asc'?scan'208";a="408568371"
Received: from rcdn-core-5.cisco.com ([173.37.93.156]) by rcdn-iport-5.cisco.com with ESMTP; 02 Apr 2015 03:34:13 +0000
Received: from xhc-aln-x10.cisco.com (xhc-aln-x10.cisco.com [173.36.12.84]) by rcdn-core-5.cisco.com (8.14.5/8.14.5) with ESMTP id t323YC4Y027842 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Thu, 2 Apr 2015 03:34:12 GMT
Received: from xmb-rcd-x09.cisco.com ([169.254.9.67]) by xhc-aln-x10.cisco.com ([173.36.12.84]) with mapi id 14.03.0195.001; Wed, 1 Apr 2015 22:34:12 -0500
From: "Fred Baker (fred)" <fred@cisco.com>
To: "draft-ietf-aqm-codel.all@tools.ietf.org" <draft-ietf-aqm-codel.all@tools.ietf.org>
Thread-Topic: codel spelling nits
Thread-Index: AQHQbPXiqqnVGgRyH0eaSueIRNJeHw==
Date: Thu, 2 Apr 2015 03:34:11 +0000
Message-ID: <5F53F60C-8374-4A38-9D68-25647ED8428F@cisco.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: yes
X-MS-TNEF-Correlator: 
x-originating-ip: [10.19.64.117]
Content-Type: multipart/signed; boundary="Apple-Mail=_1178BEA6-4B98-43AB-97B1-DDD58C8F16D6"; protocol="application/pgp-signature"; micalg=pgp-sha1
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/6TxkrVRoMKCPlqegLXas-Yz_08I>
Cc: "aqm@ietf.org list" <aqm@ietf.org>
Subject: [aqm] codel spelling nits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 02 Apr 2015 03:34:18 -0000

--Apple-Mail=_1178BEA6-4B98-43AB-97B1-DDD58C8F16D6
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

I=E2=80=99m in the process, as I volunteered to do, of reviewing the =
Codel draft. It is very well and readably written, so I don=E2=80=99t =
expect to have a lot to complain about. I do have three points, though.

One is a few spelling errors, noted in this email.

A second is the matter of marking and dropping. The world =E2=80=9Cdrop=E2=
=80=9D is used in one or another of its forms 120 times, and ECN marking =
twice, both times in passing. I=E2=80=99d like to understand better how =
ECN is handled, if only to have a statement that it is handled the same =
way as dropping.

The third is the definition of a =E2=80=9Cnormal TCP session=E2=80=9D, =
which is referred to a number of times but not nailed down that I have =
found other than that it appears to have a less than 100ms RTT. Later in =
the draft, it talks about accommodations for data center use, which =
tells me (consistent with our testing) that there is a lower bound below =
which is is less than optimal, and in our testing RTTs greater than 100 =
ms didn=E2=80=99t behave so well. The sweet spot might have been around =
25 ms, with minor degradation at 50 ms, and more degradation at longer =
RTTs.

I=E2=80=99ll come back if I find more issues,

*** draft/draft-ietf-aqm-codel-00.txt	Fri Oct 24 08:58:40 2014
--- draft-ietf-aqm-codel-00.txt	Wed Apr  1 18:43:33 2015
***************
*** 104,111 ****
      5.1.  Data Types  . . . . . . . . . . . . . . . . . . . . . . .  =
18
      5.2.  Per-queue state (codel_queue_t instance variables)  . . .  =
19
      5.3.  Constants . . . . . . . . . . . . . . . . . . . . . . . .  =
19
!      5.4.  Enque routine . . . . . . . . . . . . . . . . . . . . . .  =
19
!      5.5.  Deque routine . . . . . . . . . . . . . . . . . . . . . .  =
19



--- 104,111 ----
      5.1.  Data Types  . . . . . . . . . . . . . . . . . . . . . . .  =
18
      5.2.  Per-queue state (codel_queue_t instance variables)  . . .  =
19
      5.3.  Constants . . . . . . . . . . . . . . . . . . . . . . . .  =
19
!      5.4.  Enqueue routine . . . . . . . . . . . . . . . . . . . . . . =
 19
!      5.5.  Dequeue routine . . . . . . . . . . . . . . . . . . . . . . =
 19



***************
*** 283,289 ****


    is much slower than the link that feeds it (say, a high-speed
!    ethernet link into a limited DSL uplink) a 20 packet buffer at the
    bottleneck might be necessary to temporarily hold the 20 packets in
    flight to keep the utilization high.  The burst of packets should
    drain completely (to 0 or 1 packets) within a round trip time and
--- 283,289 ----


    is much slower than the link that feeds it (say, a high-speed
!    Ethernet link into a limited DSL uplink) a 20 packet buffer at the
    bottleneck might be necessary to temporarily hold the 20 packets in
    flight to keep the utilization high.  The burst of packets should
    drain completely (to 0 or 1 packets) within a round trip time and
***************
*** 326,332 ****

    The use of queue length is further complicated in networks that are
    subject to both short and long term changes in available link rate
!    (as in wifi).  Link rate drops can result in a spike in queue =
length
    that should be ignored unless it persists.  The length metric is
    problematic when what we really want to control is the amount of
    excess delay packets experience due to a persistent or standing
--- 326,332 ----

    The use of queue length is further complicated in networks that are
    subject to both short and long term changes in available link rate
!    (as in WiFi).  Link rate drops can result in a spike in queue =
length
    that should be ignored unless it persists.  The length metric is
    problematic when what we really want to control is the amount of
    excess delay packets experience due to a persistent or standing
***************
*** 455,461 ****
    As Kleinrock observed, the best operating point, in terms of
    bandwidth / delay tradeoff, is the peak power point since points off
    the peak represent a higher cost (in delay) per unit of bandwidth.
!    The power vs. _f_ curve for any AIMD TCP is monotone decreasing.  =
But
    the curve is very flat for _f_ < 0.1 followed by a increasing
    curvature with a knee around .2 then a steep, almost linear fall off
    [TSV84] [VJTARG14].  Since the previous equation showed that goodput
--- 455,461 ----
    As Kleinrock observed, the best operating point, in terms of
    bandwidth / delay tradeoff, is the peak power point since points off
    the peak represent a higher cost (in delay) per unit of bandwidth.
!    The power vs. _f_ curve for any AIMD TCP is monotonically =
decreasing.  But
    the curve is very flat for _f_ < 0.1 followed by a increasing
    curvature with a knee around .2 then a steep, almost linear fall off
    [TSV84] [VJTARG14].  Since the previous equation showed that goodput
***************
*** 929,935 ****
    management as described here or to adapt its principles to other
    applications.

!    Implementors are strongly encouraged to also look at Eric Dumazet's
    Linux kernel version of CoDel - a well-written, well tested, real-
    world, C-based implementation.  As of this writing, it is at:

--- 929,935 ----
    management as described here or to adapt its principles to other
    applications.

!    Implementers are strongly encouraged to also look at Eric Dumazet's
    Linux kernel version of CoDel - a well-written, well tested, real-
    world, C-based implementation.  As of this writing, it is at:

***************
*** 999,1006 ****

    "packet_t*" is a pointer to a packet descriptor.  We assume it has a
    tstamp field capable of holding a time_t and that field is available
!    for use by CoDel (it will be set by the enque routine and used by =
the
!    deque routine).



--- 999,1006 ----

    "packet_t*" is a pointer to a packet descriptor.  We assume it has a
    tstamp field capable of holding a time_t and that field is available
!    for use by CoDel (it will be set by the enqueue routine and used by =
the
!    dequeue routine).



***************
*** 1033,1039 ****
    u_int maxpacket =3D 512; // Maximum packet size in bytes
                     // (should use interface MTU)

! 5.4.  Enque routine

    All the work of CoDel is done in the deque routine.  The only CoDel
    addition to enque is putting the current time in the packet's tstamp
--- 1033,1039 ----
    u_int maxpacket =3D 512; // Maximum packet size in bytes
                     // (should use interface MTU)

! 5.4.  Enqueue routine

    All the work of CoDel is done in the deque routine.  The only CoDel
    addition to enque is putting the current time in the packet's tstamp
***************
*** 1046,1052 ****
        queue_t::enque(pkt);
    }

! 5.5.  Deque routine

    This is the heart of CoDel.  There are two branches: In packet-
    dropping state (meaning that the queue-sojourn time has gone above
--- 1046,1052 ----
        queue_t::enque(pkt);
    }

! 5.5.  Dequeue routine

    This is the heart of CoDel.  There are two branches: In packet-
    dropping state (meaning that the queue-sojourn time has gone above
***************
*** 1260,1266 ****

    An experiment by Stanford graduate students successfully used the
    linux CoDel to duplicate our published simulation work on CoDel's
!    ability to following drastic link rate changes which can be found =
at:
    http://reproducingnetworkresearch.wordpress.com/2012/06/06/solving-
    bufferbloat-the-codel-way/ .

--- 1260,1266 ----

    An experiment by Stanford graduate students successfully used the
    linux CoDel to duplicate our published simulation work on CoDel's
!    ability to follow drastic link rate changes which can be found at:
    http://reproducingnetworkresearch.wordpress.com/2012/06/06/solving-
    bufferbloat-the-codel-way/ .


--Apple-Mail=_1178BEA6-4B98-43AB-97B1-DDD58C8F16D6
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment; filename="signature.asc"
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Message signed with OpenPGP using GPGMail

-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org

iQEVAwUBVRy0YZ9ieig10VPpAQJs+ggAiijJEi02hZJHfDSw62U+Eqcejzx9p6zL
ziTsr8EMGuYCqCQ4ZxUGcLPQmQSrnPa/u+czTuCXYtkdroU0AtB/NxFQPkVd2QLY
GwcCyHOQ8NafCK7HJHlqAs9rWuoV6QFf1DAYPW9UylK+Bl8za27I24/cfkTQfmeW
zGBF6Nf/++Yg6D9eDrHpOzKd3p/rEaEWObQQUQPV7X+hte9uzpGF74raCfe3VXg4
vfsVAbLGC92HqCSzJPJ9s0TCcDCfeNqn4r1FRAncjOBTcb1YEZjSPZHKQz6PI98D
RlcosJkRiVhzAhlViuGOfBUaie+OqmK/JoLiFP2A4vcUaKQmbHSYLw==
=9V7B
-----END PGP SIGNATURE-----

--Apple-Mail=_1178BEA6-4B98-43AB-97B1-DDD58C8F16D6--


From nobody Thu Apr  2 05:09:22 2015
Return-Path: <internet-drafts@ietf.org>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 555931B2C91; Thu,  2 Apr 2015 05:09:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.9
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 tagged_above=-999 required=5 tests=[BAYES_00=-1.9] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nXetIbKyHMhb; Thu,  2 Apr 2015 05:09:20 -0700 (PDT)
Received: from ietfa.amsl.com (localhost [IPv6:::1]) by ietfa.amsl.com (Postfix) with ESMTP id 04FCC1A1A59; Thu,  2 Apr 2015 05:09:20 -0700 (PDT)
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
From: internet-drafts@ietf.org
To: <i-d-announce@ietf.org>
X-Test-IDTracker: no
X-IETF-IDTracker: 5.13.0
Auto-Submitted: auto-generated
Precedence: bulk
Message-ID: <20150402120920.32599.29704.idtracker@ietfa.amsl.com>
Date: Thu, 02 Apr 2015 05:09:20 -0700
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/lyRq2v8tu7RasVIz4-kCGwKfDhM>
Cc: aqm@ietf.org
Subject: [aqm] I-D Action: draft-ietf-aqm-ecn-benefits-03.txt
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 02 Apr 2015 12:09:21 -0000

A New Internet-Draft is available from the on-line Internet-Drafts directories.
 This draft is a work item of the Active Queue Management and Packet Scheduling Working Group of the IETF.

        Title           : The Benefits of using Explicit Congestion Notification (ECN)
        Authors         : Godred Fairhurst
                          Michael Welzl
	Filename        : draft-ietf-aqm-ecn-benefits-03.txt
	Pages           : 14
	Date            : 2015-04-02

Abstract:
   This document describes the potential benefits when applications
   enable Explicit Congestion Notification (ECN).  It outlines the
   principal gains in terms of increased throughput, reduced delay and
   other benefits when ECN is used over network paths that include
   equipment that supports ECN-marking.  It also identifies some
   potential problems that might occur when ECN is used.  The document
   does not propose new algorithms that may be able to use ECN or
   describe the details of implementation of ECN in endpoint devices,
   routers and other network devices.


The IETF datatracker status page for this draft is:
https://datatracker.ietf.org/doc/draft-ietf-aqm-ecn-benefits/

There's also a htmlized version available at:
http://tools.ietf.org/html/draft-ietf-aqm-ecn-benefits-03

A diff from the previous version is available at:
http://www.ietf.org/rfcdiff?url2=draft-ietf-aqm-ecn-benefits-03


Please note that it may take a couple of minutes from the time of submission
until the htmlized version and diff are available at tools.ietf.org.

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


From nobody Thu Apr  2 08:28:44 2015
Return-Path: <ddolson@sandvine.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id EF10E1AC405 for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 08:28:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.11
X-Spam-Level: 
X-Spam-Status: No, score=-0.11 tagged_above=-999 required=5 tests=[BAYES_20=-0.001, HTML_MESSAGE=0.001, J_CHICKENPOX_84=0.6, RCVD_IN_DNSWL_LOW=-0.7, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gV0PjPBwyHv1 for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 08:28:33 -0700 (PDT)
Received: from mail1.sandvine.com (Mail1.sandvine.com [64.7.137.134]) by ietfa.amsl.com (Postfix) with ESMTP id C8B031AC410 for <aqm@ietf.org>; Thu,  2 Apr 2015 08:28:32 -0700 (PDT)
Received: from BLR-EXCHP-2.sandvine.com (192.168.196.172) by WTL-EXCHP-2.sandvine.com (192.168.194.177) with Microsoft SMTP Server (TLS) id 14.3.195.1; Thu, 2 Apr 2015 11:28:32 -0400
Received: from WTL-EXCHP-2.sandvine.com ([fe80::68ac:f071:19ff:3455]) by blr-exchp-2.sandvine.com ([fe80::6c6d:7108:c63c:9055%14]) with mapi id 14.03.0181.006; Thu, 2 Apr 2015 11:28:31 -0400
From: Dave Dolson <ddolson@sandvine.com>
To: "Rong Pan (ropan)" <ropan@cisco.com>, "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: [aqm] PIE departure rate estimation
Thread-Index: AQHQbMIGW+ot0iZ1OU6OZ2/eKhmRQZ04qyFA///v6gCAARmgAA==
Date: Thu, 2 Apr 2015 15:28:31 +0000
Message-ID: <E8355113905631478EFF04F5AA706E9830BB34D5@wtl-exchp-2.sandvine.com>
References: <D141ACCB.D99D%ropan@cisco.com> <E8355113905631478EFF04F5AA706E9830BB0BF4@wtl-exchp-2.sandvine.com> <D141BCEA.D9C1%ropan@cisco.com>
In-Reply-To: <D141BCEA.D9C1%ropan@cisco.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [192.168.200.63]
Content-Type: multipart/alternative; boundary="_000_E8355113905631478EFF04F5AA706E9830BB34D5wtlexchp2sandvi_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/ifkBe6uvME5Nqm-J6ulZWlooBxo>
Subject: Re: [aqm] PIE departure rate estimation
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 02 Apr 2015 15:28:43 -0000

--_000_E8355113905631478EFF04F5AA706E9830BB34D5wtlexchp2sandvi_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

Rong,
(Please keep in mind I am only trying to provide constructive criticism, as=
 someone new to the PIE algorithm.)

>>>>>>>>>>>>>>RP: There are two calculations in the PIE algorithm: one is f=
or p update and one is for calculate the rate. They are independent from ea=
ch other. If there are not enough packets in the buffer, we simply don't up=
date depart rate estimation,  instead we use the previous value. However, t=
he drop probability calculation is always updated. Hence, the "p" in your e=
xample will go down to 0.

I didn't get this point from the draft, that 'p' is periodically updated re=
gardless of whether rate measurements are available.
May I suggest a statement in section 4.3: "Note that p is updated per secti=
on 4.2 even when not in a measurement cycle; in this case the previously la=
tched value of depart_rate is used."


[Sorry, the following is really pedantic, but I couldn't help myself...]

I tried to prove to myself that "p" will go down to 0 when the queue size i=
s below a certain size.
I found that it will not go down to 0 in a contrived situation when packet =
arrival rate equals transmit rate, keeping qlen constant
target_del * depart_rate <=3D qlen <=3D dq_threshold

In this range, depart_rate is never recomputed, so it might have latched an=
y arbitrary value, making the lower bound a constant.

In this range, p is not reduced because the beta*(est_del - est_del_old) te=
rm is zero (constant qlen) and the alpha*(est_del-target_del) is zero or po=
sitive.

I didn't work out the details, but because it takes a while for 'p' to get =
to zero, I think there are a variety of traffic patterns that can allow 'p'=
 to meander around in non-zero space even though the queue is nearly empty.=
 (p could be fairly large, which is the reason the queue is nearly empty).

The key to this scenario is latching a low value of depart_rate and never u=
pdating it.

In practice this might never happen, but also in practice anything that can=
 happen will. At least it probably won't be reproducible :-)

Is there a fix? I think it is to remove the restriction that a minimum queu=
e size is required for a measurement cycle. It's better to measure the actu=
al rate (even if it is not the potential rate) than have a bad estimate. Th=
en depart_rate is allowed to increase to the actual rate, and the estimated=
 delay is allowed to decrease.

-Dave



From: Rong Pan (ropan) [mailto:ropan@cisco.com]
Sent: Wednesday, April 01, 2015 6:28 PM
To: Dave Dolson; aqm@ietf.org
Subject: Re: [aqm] PIE departure rate estimation

Please see inline...

Rong

From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 2:43 PM
To: rong <ropan@cisco.com<mailto:ropan@cisco.com>>, "aqm@ietf.org<mailto:aq=
m@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: RE: [aqm] PIE departure rate estimation

Inline [DD]

From: Rong Pan (ropan) [mailto:ropan@cisco.com]
Sent: Wednesday, April 01, 2015 5:23 PM
To: Dave Dolson; aqm@ietf.org<mailto:aqm@ietf.org>
Subject: Re: [aqm] PIE departure rate estimation

Dave,

Thanks for the valuable comments. We will see how we can incorporate them. =
For the comments below, please see inline.

Regards,

Rong


From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 1:22 PM
To: "aqm@ietf.org<mailto:aqm@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: [aqm] PIE departure rate estimation

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3,

it says, "We only measure the departure rate when there are sufficient data=
 in the buffer"

Why can't departure rate be estimated regardless of queue size? Just count =
packets leaving over time? I'm wondering how to avoid the estimate getting =
stuck at the last value sampled when the queue had a certain quantity in it=
.

>>>>>>>>>>>>RP: We need to make sure that there are enough data in the queu=
e to guarantee a rate sample. The reason is that if a queue is empty from t=
ime to time, we can't measure its true draining rate. The time when the que=
ue empty should be cut out of the drain time calculation. For simplicity, i=
t is better to make sure we have enough data in the queue to ensure accurat=
e rate measurement sample.
[DD] I understand the rationale, but supposing p=3D0.1, but the queue has v=
ery few items in it, too few to obtain a rate calculation. What causes p to=
 go to zero, since an update is not permitted? It isn't obvious to me that =
the algorithm always stops for all input traffic patterns.

>>>>>>>>>>>>>>RP: There are two calculations in the PIE algorithm: one is f=
or p update and one is for calculate the rate. They are independent from ea=
ch other. If there are not enough packets in the buffer, we simply don't up=
date depart rate estimation,  instead we use the previous value. However, t=
he drop probability calculation is always updated. Hence, the "p" in your e=
xample will go down to 0.


Section 4.2 cites Little's Law as "est_del =3D qlen/depart_rate",
but according to Wikipedia<http://en.wikipedia.org/wiki/Little%27s_law>, th=
e law uses arrival rate, not departure rate.
I don't know if it matters (I didn't read Little's proof), but this gives c=
redence to the suggestion in section 6 that the algorithm could use arrival=
 rate.
And I think it might be easier to measure when the queue has few items in i=
t.

>>>>>>>>>>>>RP: what you mentioned is to calculate average number of custom=
ers in a system using the arrival. http://en.wikipedia.org/wiki/Little's_la=
w Wikipedia also mentions that mean responseTime =3D MeanNumberInSystem / M=
eanThroughput. What we measure is the mean response time (latency). Hence, =
it is correct in our draft.
[DD] Arrival (post drop) and departure rates are almost never the same at a=
ny instant. I guess they must average to the same value. Practically, it mi=
ght not matter which is used, but I thought Little's law should be cited co=
rrectly. In Wikipedia the law is cited one way and the example is shown ano=
ther way. Furthermore there is this ominous comment: "When exploring Little=
's law and learning to trust it, be aware of the common mistakes of using a=
rrivals(work arriving) when throughput(work completed) is called for..."

>>>>>>>>>>>>>RP: I understand the concern. But I think we have used it in t=
he right way. Arrival is related to how many customers/packets in a system.=
 However, given #of customers/packets in a system, the wait time is directl=
y related to the service time (inverse of throughput), not arrival rate any=
 more.

David Dolson
Senior Software Architect, Sandvine Inc.


--_000_E8355113905631478EFF04F5AA706E9830BB34D5wtlexchp2sandvi_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"=
>
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
	{font-family:Tahoma;
	panose-1:2 11 6 4 3 5 4 4 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
p.MsoAcetate, li.MsoAcetate, div.MsoAcetate
	{mso-style-priority:99;
	mso-style-link:"Balloon Text Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:8.0pt;
	font-family:"Tahoma","sans-serif";}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
span.BalloonTextChar
	{mso-style-name:"Balloon Text Char";
	mso-style-priority:99;
	mso-style-link:"Balloon Text";
	font-family:"Tahoma","sans-serif";}
span.EmailStyle21
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.EmailStyle22
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
span.EmailStyle23
	{mso-style-type:personal-reply;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
.MsoChpDefault
	{mso-style-type:export-only;
	font-size:10.0pt;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">Rong,<o=
:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">(Please=
 keep in mind I am only trying to provide constructive criticism, as someon=
e new to the PIE algorithm.)<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: There are two calculat=
ions in the PIE algorithm: one is for p update and one is for calculate the=
 rate. They are independent from each other. If there are not enough packet=
s in
 the buffer, we simply don't update depart rate estimation, &nbsp;instead w=
e use the previous value. However, the drop probability calculation is alwa=
ys updated. Hence, the &quot;p&quot; in your example will go down to 0.<o:p=
></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I didn&#8217;t get thi=
s point from the draft, that &#8216;p&#8217; is periodically updated regard=
less of whether rate measurements are available.
<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">May I suggest a statem=
ent in section 4.3: &#8220;Note that p is updated per section 4.2 even when=
 not in a measurement cycle; in this case the previously latched value of d=
epart_rate is used.&#8221;<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[Sorry, the following =
is really pedantic, but I couldn&#8217;t help myself&#8230;]<o:p></o:p></sp=
an></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I tried to prove to my=
self that &#8220;p&#8221; will go down to 0 when the queue size is below a =
certain size.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I found that it will n=
ot go down to 0 in a contrived situation when packet arrival rate equals tr=
ansmit rate, keeping qlen constant<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">target_del * depart_ra=
te &lt;=3D qlen &lt;=3D dq_threshold<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">In this range, depart_=
rate is never recomputed, so it might have latched any arbitrary value, mak=
ing the lower bound a constant.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">In this range, p is no=
t reduced because the beta*(est_del - est_del_old) term is zero (constant q=
len) and the alpha*(est_del-target_del) is zero or positive.<o:p></o:p></sp=
an></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I didn&#8217;t work ou=
t the details, but because it takes a while for &#8216;p&#8217; to get to z=
ero, I think there are a variety of traffic patterns that can allow &#8216;=
p&#8217; to meander around in non-zero space even though the queue
 is nearly empty. (p could be fairly large, which is the reason the queue i=
s nearly empty).<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">The key to this scenar=
io is latching a low value of depart_rate and never updating it.<o:p></o:p>=
</span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">In practice this might=
 never happen, but also in practice anything that can happen will. At least=
 it probably won&#8217;t be reproducible :-)<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">Is there a fix? I thin=
k it is to remove the restriction that a minimum queue size is required for=
 a measurement cycle. It&#8217;s better to measure the actual rate (even if=
 it is not the potential rate) than have a
 bad estimate. Then depart_rate is allowed to increase to the actual rate, =
and the estimated delay is allowed to decrease.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">-Dave<o:p></o:p></span=
></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"font-size:10.0pt;font-family:&quot=
;Tahoma&quot;,&quot;sans-serif&quot;">From:</span></b><span style=3D"font-s=
ize:10.0pt;font-family:&quot;Tahoma&quot;,&quot;sans-serif&quot;"> Rong Pan=
 (ropan) [mailto:ropan@cisco.com]
<br>
<b>Sent:</b> Wednesday, April 01, 2015 6:28 PM<br>
<b>To:</b> Dave Dolson; aqm@ietf.org<br>
<b>Subject:</b> Re: [aqm] PIE departure rate estimation<o:p></o:p></span></=
p>
</div>
</div>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Please see inline&#8230;=
<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Rong<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"color:black">From: </span></b><spa=
n style=3D"color:black">Dave Dolson &lt;<a href=3D"mailto:ddolson@sandvine.=
com">ddolson@sandvine.com</a>&gt;<br>
<b>Date: </b>Wednesday, April 1, 2015 2:43 PM<br>
<b>To: </b>rong &lt;<a href=3D"mailto:ropan@cisco.com">ropan@cisco.com</a>&=
gt;, &quot;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&quot; &lt;<a hr=
ef=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&gt;<br>
<b>Subject: </b>RE: [aqm] PIE departure rate estimation<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">Inline [DD]</span><spa=
n style=3D"color:black"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">&nbsp;</span><span sty=
le=3D"color:black"><o:p></o:p></span></p>
<div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"font-size:10.0pt;font-family:&quot=
;Tahoma&quot;,&quot;sans-serif&quot;;color:black">From:</span></b><span sty=
le=3D"font-size:10.0pt;font-family:&quot;Tahoma&quot;,&quot;sans-serif&quot=
;;color:black"> Rong Pan (ropan) [<a href=3D"mailto:ropan@cisco.com">mailto=
:ropan@cisco.com</a>]
<br>
<b>Sent:</b> Wednesday, April 01, 2015 5:23 PM<br>
<b>To:</b> Dave Dolson; <a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a><br=
>
<b>Subject:</b> Re: [aqm] PIE departure rate estimation</span><span style=
=3D"color:black"><o:p></o:p></span></p>
</div>
</div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Dave,<o:p></o:p></span><=
/p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Thanks for the valuable =
comments. We will see how we can incorporate them. For the comments below, =
please see inline.<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Regards,<o:p></o:p></spa=
n></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Rong<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"color:black">From: </span></b><spa=
n style=3D"color:black">Dave Dolson &lt;<a href=3D"mailto:ddolson@sandvine.=
com">ddolson@sandvine.com</a>&gt;<br>
<b>Date: </b>Wednesday, April 1, 2015 1:22 PM<br>
<b>To: </b>&quot;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&quot; &lt=
;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&gt;<br>
<b>Subject: </b>[aqm] PIE departure rate estimation<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&nbsp;<=
/span><span style=3D"color:black"><o:p></o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">In <a href=3D"https://to=
ols.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3</a>,<o:p></o:=
p></span></p>
<pre><span style=3D"font-size:11.0pt;font-family:&quot;Calibri&quot;,&quot;=
sans-serif&quot;;color:black">it says, </span><span style=3D"color:black">&=
#8220;We only measure the departure rate when there are sufficient data in =
the buffer&#8221;<o:p></o:p></span></pre>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">Why can&#8217;t departur=
e rate be estimated regardless of queue size? Just count packets leaving ov=
er time? I&#8217;m wondering how to avoid the estimate getting stuck at the=
 last value sampled when the queue had a certain
 quantity in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: We need to make sure that ther=
e are enough data in the queue to guarantee a rate sample. The reason is th=
at if a queue is empty from time to time, we can't measure its true drainin=
g
 rate. The time when the queue empty should be cut out of the drain time ca=
lculation. For simplicity, it is better to make sure we have enough data in=
 the queue to ensure accurate rate measurement sample.</span><span style=3D=
"color:black"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] I understand the =
rationale, but supposing p=3D0.1, but the queue has very few items in it, t=
oo few to obtain a rate calculation. What causes p to go to zero, since an =
update is not permitted? It isn&#8217;t obvious
 to me that the algorithm always stops for all input traffic patterns.</spa=
n><span style=3D"color:black"><o:p></o:p></span></p>
</div>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: There are two calculat=
ions in the PIE algorithm: one is for p update and one is for calculate the=
 rate. They are independent from each other. If there are not enough packet=
s in
 the buffer, we simply don't update depart rate estimation, &nbsp;instead w=
e use the previous value. However, the drop probability calculation is alwa=
ys updated. Hence, the &quot;p&quot; in your example will go down to 0.<o:p=
></o:p></span></p>
</div>
<div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">&nbsp;</span><span sty=
le=3D"color:black"><o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&nbsp;<=
/span><span style=3D"color:black"><o:p></o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Section 4.2 cites Little=
&#8217;s Law as &#8220;est_del =3D qlen/depart_rate&#8221;,<o:p></o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:black">but <a href=3D"http://en=
.wikipedia.org/wiki/Little%27s_law">
according to Wikipedia</a>, the law uses arrival rate, not departure rate.<=
o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">I don&#8217;t know if it=
 matters (I didn&#8217;t read Little&#8217;s proof), but this gives credenc=
e to the suggestion in section 6 that the algorithm could use arrival rate.=
<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">And I think it might be =
easier to measure when the queue has few items in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: what you mentioned is to calcu=
late average number of customers in a system using the arrival.
<a href=3D"http://en.wikipedia.org/wiki/Little's_law">http://en.wikipedia.o=
rg/wiki/Little's_law</a>&nbsp;Wikipedia also mentions that mean responseTim=
e =3D&nbsp;</span><span style=3D"font-size:10.5pt;font-family:&quot;Arial&q=
uot;,&quot;sans-serif&quot;;color:#252525;background:white">MeanNumberInSys=
tem
 / MeanThroughput. What we measure is the mean response time (latency). Hen=
ce, it is correct in our draft.&nbsp;</span><span style=3D"color:black"><o:=
p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] Arrival (post dro=
p) and departure rates are almost never the same at any instant. I guess th=
ey must average to the same value. Practically, it might not matter which i=
s used, but I thought Little&#8217;s law should
 be cited correctly. In Wikipedia the law is cited one way and the example =
is shown another way. Furthermore there is this ominous comment: &#8220;Whe=
n exploring Little&#8217;s law and learning to trust it, be aware of the co=
mmon mistakes of using arrivals(work arriving)
 when throughput(work completed) is called for&#8230;&#8221;</span><span st=
yle=3D"color:black"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">&nbsp;</span><span sty=
le=3D"color:black"><o:p></o:p></span></p>
</div>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: I understand the concern. =
But I think we have used it in the right way. Arrival is related to how man=
y customers/packets in a system. However, given #of customers/packets in a
 system, the wait time is directly related to the service time (inverse of =
throughput), not arrival rate any more.<o:p></o:p></span></p>
</div>
<div>
<div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">David Dolson<o:p></o:p><=
/span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">Senior Software Architec=
t, Sandvine Inc.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
</div>
</div>
</div>
</body>
</html>

--_000_E8355113905631478EFF04F5AA706E9830BB34D5wtlexchp2sandvi_--


From nobody Thu Apr  2 10:31:44 2015
Return-Path: <ropan@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 08ECD1AD05C for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 10:31:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -13.91
X-Spam-Level: 
X-Spam-Status: No, score=-13.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=0.001, J_CHICKENPOX_84=0.6, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZMqnEAbNvoEs for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 10:31:38 -0700 (PDT)
Received: from rcdn-iport-5.cisco.com (rcdn-iport-5.cisco.com [173.37.86.76]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 7B19D1A014D for <aqm@ietf.org>; Thu,  2 Apr 2015 10:31:38 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=29931; q=dns/txt; s=iport; t=1427995898; x=1429205498; h=from:to:subject:date:message-id:references:in-reply-to: mime-version; bh=HEO9fLQ7tW5ghnWHyThzA868Vu8riSrJsyAUYR8LPNI=; b=JhD7nLW3/pgDwlT+BaiP/2GaqAT7zKRFspr1pivvnf5SzOpfF8L1xB3F dhaMpa+x8fZFJiPKfjGcTClc4qqYMbTH8PdjLWZSyVeOjbGQ9b3pYG9MD II32y/EEuAv21l4pGkY4ydi6qDdyBMjbh/Gi0uiIgZN46f+lY3l/hgghi M=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0ClBgC9ex1V/4cNJK1cgkVDUlwFw1qCA4VzAoFKTAEBAQEBAX6EHgEBAQQMIUAcAgEIDgMDAQEBIQcHMhQJCAIEARIfiA8BDc4LAQEBAQEBAQEBAQEBAQEBAQEBAQEBF4oqf4RoDQoBhC0FkGqDdYYLgR2DNIJehhyDNYNIIoNubwGBQ38BAQE
X-IronPort-AV: E=Sophos;i="5.11,512,1422921600";  d="scan'208,217";a="408748843"
Received: from alln-core-2.cisco.com ([173.36.13.135]) by rcdn-iport-5.cisco.com with ESMTP; 02 Apr 2015 17:31:37 +0000
Received: from xhc-rcd-x12.cisco.com (xhc-rcd-x12.cisco.com [173.37.183.86]) by alln-core-2.cisco.com (8.14.5/8.14.5) with ESMTP id t32HVbho009159 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Thu, 2 Apr 2015 17:31:37 GMT
Received: from xmb-aln-x15.cisco.com ([169.254.9.109]) by xhc-rcd-x12.cisco.com ([173.37.183.86]) with mapi id 14.03.0195.001; Thu, 2 Apr 2015 12:31:37 -0500
From: "Rong Pan (ropan)" <ropan@cisco.com>
To: Dave Dolson <ddolson@sandvine.com>, "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: [aqm] PIE departure rate estimation
Thread-Index: AQHQbMIGW+ot0iZ1OU6OZ2/eKhmRQZ04qyFA///v6gCAARmgAIAAJgQA
Date: Thu, 2 Apr 2015 17:31:36 +0000
Message-ID: <D142C6AE.D9F1%ropan@cisco.com>
References: <D141ACCB.D99D%ropan@cisco.com> <E8355113905631478EFF04F5AA706E9830BB0BF4@wtl-exchp-2.sandvine.com> <D141BCEA.D9C1%ropan@cisco.com> <E8355113905631478EFF04F5AA706E9830BB34D5@wtl-exchp-2.sandvine.com>
In-Reply-To: <E8355113905631478EFF04F5AA706E9830BB34D5@wtl-exchp-2.sandvine.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
user-agent: Microsoft-MacOutlook/14.4.7.141117
x-originating-ip: [171.70.247.185]
Content-Type: multipart/alternative; boundary="_000_D142C6AED9F1ropanciscocom_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/a14PkURL7jpaFaiF1Xf2nbTLog4>
Subject: Re: [aqm] PIE departure rate estimation
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 02 Apr 2015 17:31:42 -0000

--_000_D142C6AED9F1ropanciscocom_
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

Dave,

Thanks for your time and effort in reviewing the PIE draft. We really appre=
ciate it and we will update according to many of your valuable comments (yo=
ur other emails). For the ones that I think need clarification, I try to ex=
plain below.

Best Regards,

Rong

From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Thursday, April 2, 2015 8:28 AM
To: rong <ropan@cisco.com<mailto:ropan@cisco.com>>, "aqm@ietf.org<mailto:aq=
m@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: RE: [aqm] PIE departure rate estimation

Rong,
(Please keep in mind I am only trying to provide constructive criticism, as=
 someone new to the PIE algorithm.)

>>>>>>>>>>>>>>RP: There are two calculations in the PIE algorithm: one is f=
or p update and one is for calculate the rate. They are independent from ea=
ch other. If there are not enough packets in the buffer, we simply don't up=
date depart rate estimation,  instead we use the previous value. However, t=
he drop probability calculation is always updated. Hence, the "p" in your e=
xample will go down to 0.

I didn=92t get this point from the draft, that =91p=92 is periodically upda=
ted regardless of whether rate measurements are available.
May I suggest a statement in section 4.3: =93Note that p is updated per sec=
tion 4.2 even when not in a measurement cycle; in this case the previously =
latched value of depart_rate is used.=94

>>>>>>>>>>>>>>>RP: will do.


[Sorry, the following is really pedantic, but I couldn=92t help myself=85]

I tried to prove to myself that =93p=94 will go down to 0 when the queue si=
ze is below a certain size.
I found that it will not go down to 0 in a contrived situation when packet =
arrival rate equals transmit rate, keeping qlen constant
target_del * depart_rate <=3D qlen <=3D dq_threshold


>>>>>>>>>>>>>>>RP: if arrival rate =3D departure rate and p is not equal to=
 0, the net arrival rate to the queue =3D arrival rate *(1-p) < departure r=
ate, the queue will goes to zero until p =3D 0. In any case, there is one s=
afeguard clause where p will exponentially go to 0 when queue is empty, whi=
ch is accidentally deleted in the pseudo code, which I will update.

In this range, depart_rate is never recomputed, so it might have latched an=
y arbitrary value, making the lower bound a constant.

In this range, p is not reduced because the beta*(est_del - est_del_old) te=
rm is zero (constant qlen) and the alpha*(est_del-target_del) is zero or po=
sitive.

I didn=92t work out the details, but because it takes a while for =91p=92 t=
o get to zero, I think there are a variety of traffic patterns that can all=
ow =91p=92 to meander around in non-zero space even though the queue is nea=
rly empty. (p could be fairly large, which is the reason the queue is nearl=
y empty).

The key to this scenario is latching a low value of depart_rate and never u=
pdating it.

In practice this might never happen, but also in practice anything that can=
 happen will. At least it probably won=92t be reproducible :-)

Is there a fix? I think it is to remove the restriction that a minimum queu=
e size is required for a measurement cycle. It=92s better to measure the ac=
tual rate (even if it is not the potential rate) than have a bad estimate. =
Then depart_rate is allowed to increase to the actual rate, and the estimat=
ed delay is allowed to decrease.

-Dave



From: Rong Pan (ropan) [mailto:ropan@cisco.com]
Sent: Wednesday, April 01, 2015 6:28 PM
To: Dave Dolson; aqm@ietf.org<mailto:aqm@ietf.org>
Subject: Re: [aqm] PIE departure rate estimation

Please see inline=85

Rong

From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 2:43 PM
To: rong <ropan@cisco.com<mailto:ropan@cisco.com>>, "aqm@ietf.org<mailto:aq=
m@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: RE: [aqm] PIE departure rate estimation

Inline [DD]

From: Rong Pan (ropan) [mailto:ropan@cisco.com]
Sent: Wednesday, April 01, 2015 5:23 PM
To: Dave Dolson; aqm@ietf.org<mailto:aqm@ietf.org>
Subject: Re: [aqm] PIE departure rate estimation

Dave,

Thanks for the valuable comments. We will see how we can incorporate them. =
For the comments below, please see inline.

Regards,

Rong


From: Dave Dolson <ddolson@sandvine.com<mailto:ddolson@sandvine.com>>
Date: Wednesday, April 1, 2015 1:22 PM
To: "aqm@ietf.org<mailto:aqm@ietf.org>" <aqm@ietf.org<mailto:aqm@ietf.org>>
Subject: [aqm] PIE departure rate estimation

In https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3,

it says, =93We only measure the departure rate when there are sufficient da=
ta in the buffer=94

Why can=92t departure rate be estimated regardless of queue size? Just coun=
t packets leaving over time? I=92m wondering how to avoid the estimate gett=
ing stuck at the last value sampled when the queue had a certain quantity i=
n it.

>>>>>>>>>>>>RP: We need to make sure that there are enough data in the queu=
e to guarantee a rate sample. The reason is that if a queue is empty from t=
ime to time, we can't measure its true draining rate. The time when the que=
ue empty should be cut out of the drain time calculation. For simplicity, i=
t is better to make sure we have enough data in the queue to ensure accurat=
e rate measurement sample.
[DD] I understand the rationale, but supposing p=3D0.1, but the queue has v=
ery few items in it, too few to obtain a rate calculation. What causes p to=
 go to zero, since an update is not permitted? It isn=92t obvious to me tha=
t the algorithm always stops for all input traffic patterns.

>>>>>>>>>>>>>>RP: There are two calculations in the PIE algorithm: one is f=
or p update and one is for calculate the rate. They are independent from ea=
ch other. If there are not enough packets in the buffer, we simply don't up=
date depart rate estimation,  instead we use the previous value. However, t=
he drop probability calculation is always updated. Hence, the "p" in your e=
xample will go down to 0.


Section 4.2 cites Little=92s Law as =93est_del =3D qlen/depart_rate=94,
but according to Wikipedia<http://en.wikipedia.org/wiki/Little%27s_law>, th=
e law uses arrival rate, not departure rate.
I don=92t know if it matters (I didn=92t read Little=92s proof), but this g=
ives credence to the suggestion in section 6 that the algorithm could use a=
rrival rate.
And I think it might be easier to measure when the queue has few items in i=
t.

>>>>>>>>>>>>RP: what you mentioned is to calculate average number of custom=
ers in a system using the arrival. http://en.wikipedia.org/wiki/Little's_la=
w Wikipedia also mentions that mean responseTime =3D MeanNumberInSystem / M=
eanThroughput. What we measure is the mean response time (latency). Hence, =
it is correct in our draft.
[DD] Arrival (post drop) and departure rates are almost never the same at a=
ny instant. I guess they must average to the same value. Practically, it mi=
ght not matter which is used, but I thought Little=92s law should be cited =
correctly. In Wikipedia the law is cited one way and the example is shown a=
nother way. Furthermore there is this ominous comment: =93When exploring Li=
ttle=92s law and learning to trust it, be aware of the common mistakes of u=
sing arrivals(work arriving) when throughput(work completed) is called for=
=85=94

>>>>>>>>>>>>>RP: I understand the concern. But I think we have used it in t=
he right way. Arrival is related to how many customers/packets in a system.=
 However, given #of customers/packets in a system, the wait time is directl=
y related to the service time (inverse of throughput), not arrival rate any=
 more.

David Dolson
Senior Software Architect, Sandvine Inc.


--_000_D142C6AED9F1ropanciscocom_
Content-Type: text/html; charset="Windows-1252"
Content-ID: <C48F60883175254A828E18E641F47FD3@emea.cisco.com>
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3DWindows-1=
252">
</head>
<body style=3D"word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-lin=
e-break: after-white-space; color: rgb(0, 0, 0); font-size: 16px; font-fami=
ly: Calibri, sans-serif; ">
<div>Dave,</div>
<div><br>
</div>
<div>Thanks for your time and effort in reviewing the PIE draft. We really =
appreciate it and we will update according to many of your valuable comment=
s (your other emails). For the ones that I think need clarification, I try =
to explain below.</div>
<div><br>
</div>
<div>Best Regards,</div>
<div><br>
</div>
<div>Rong&nbsp;</div>
<div><br>
</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div style=3D"font-family:Calibri; font-size:11pt; text-align:left; color:b=
lack; BORDER-BOTTOM: medium none; BORDER-LEFT: medium none; PADDING-BOTTOM:=
 0in; PADDING-LEFT: 0in; PADDING-RIGHT: 0in; BORDER-TOP: #b5c4df 1pt solid;=
 BORDER-RIGHT: medium none; PADDING-TOP: 3pt">
<span style=3D"font-weight:bold">From: </span>Dave Dolson &lt;<a href=3D"ma=
ilto:ddolson@sandvine.com">ddolson@sandvine.com</a>&gt;<br>
<span style=3D"font-weight:bold">Date: </span>Thursday, April 2, 2015 8:28 =
AM<br>
<span style=3D"font-weight:bold">To: </span>rong &lt;<a href=3D"mailto:ropa=
n@cisco.com">ropan@cisco.com</a>&gt;, &quot;<a href=3D"mailto:aqm@ietf.org"=
>aqm@ietf.org</a>&quot; &lt;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a=
>&gt;<br>
<span style=3D"font-weight:bold">Subject: </span>RE: [aqm] PIE departure ra=
te estimation<br>
</div>
<div><br>
</div>
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
@font-face
	{font-family:Tahoma;
	panose-1:2 11 6 4 3 5 4 4 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
pre
	{mso-style-priority:99;
	mso-style-link:"HTML Preformatted Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:10.0pt;
	font-family:"Courier New";}
p.MsoAcetate, li.MsoAcetate, div.MsoAcetate
	{mso-style-priority:99;
	mso-style-link:"Balloon Text Char";
	margin:0in;
	margin-bottom:.0001pt;
	font-size:8.0pt;
	font-family:"Tahoma","sans-serif";}
span.HTMLPreformattedChar
	{mso-style-name:"HTML Preformatted Char";
	mso-style-priority:99;
	mso-style-link:"HTML Preformatted";
	font-family:"Courier New";}
span.BalloonTextChar
	{mso-style-name:"Balloon Text Char";
	mso-style-priority:99;
	mso-style-link:"Balloon Text";
	font-family:"Tahoma","sans-serif";}
span.EmailStyle21
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
span.EmailStyle22
	{mso-style-type:personal;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
span.EmailStyle23
	{mso-style-type:personal-reply;
	font-family:"Calibri","sans-serif";
	color:#1F497D;}
.MsoChpDefault
	{mso-style-type:export-only;
	font-size:10.0pt;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">Rong,<o=
:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">(Please=
 keep in mind I am only trying to provide constructive criticism, as someon=
e new to the PIE algorithm.)<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: There are two calculat=
ions in the PIE algorithm: one is for p update and one is for calculate the=
 rate. They are independent from each other. If there are not enough packet=
s in
 the buffer, we simply don't update depart rate estimation, &nbsp;instead w=
e use the previous value. However, the drop probability calculation is alwa=
ys updated. Hence, the &quot;p&quot; in your example will go down to 0.<o:p=
></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I didn=92t get this po=
int from the draft, that =91p=92 is periodically updated regardless of whet=
her rate measurements are available.
<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">May I suggest a statem=
ent in section 4.3: =93Note that p is updated per section 4.2 even when not=
 in a measurement cycle; in this case the previously latched value of depar=
t_rate is used.=94</span></p>
</div>
</div>
</div>
</span>
<div><br>
</div>
<div>&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: will d=
o.&nbsp;</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[Sorry, the following =
is really pedantic, but I couldn=92t help myself=85]<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I tried to prove to my=
self that =93p=94 will go down to 0 when the queue size is below a certain =
size.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I found that it will n=
ot go down to 0 in a contrived situation when packet arrival rate equals tr=
ansmit rate, keeping qlen constant<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">target_del * depart_ra=
te &lt;=3D qlen &lt;=3D dq_threshold<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
</div>
</div>
</div>
</span>
<div><br>
</div>
<div>&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: if arr=
ival rate =3D departure rate and p is not equal to 0, the net arrival rate =
to the queue =3D arrival rate *(1-p) &lt; departure rate, the queue will go=
es to zero until p =3D 0. In any case, there is one safeguard clause where =
p will exponentially
 go to 0 when queue is empty, which is accidentally deleted in the pseudo c=
ode, which I will update.</div>
<div><br>
</div>
<span id=3D"OLK_SRC_BODY_SECTION">
<div xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micro=
soft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" x=
mlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:/=
/www.w3.org/TR/REC-html40">
<div lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">In this range, depart_=
rate is never recomputed, so it might have latched any arbitrary value, mak=
ing the lower bound a constant.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">In this range, p is no=
t reduced because the beta*(est_del - est_del_old) term is zero (constant q=
len) and the alpha*(est_del-target_del) is zero or positive.<o:p></o:p></sp=
an></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">I didn=92t work out th=
e details, but because it takes a while for =91p=92 to get to zero, I think=
 there are a variety of traffic patterns that can allow =91p=92 to meander =
around in non-zero space even though the queue
 is nearly empty. (p could be fairly large, which is the reason the queue i=
s nearly empty).<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">The key to this scenar=
io is latching a low value of depart_rate and never updating it.<o:p></o:p>=
</span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">In practice this might=
 never happen, but also in practice anything that can happen will. At least=
 it probably won=92t be reproducible :-)<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">Is there a fix? I thin=
k it is to remove the restriction that a minimum queue size is required for=
 a measurement cycle. It=92s better to measure the actual rate (even if it =
is not the potential rate) than have a
 bad estimate. Then depart_rate is allowed to increase to the actual rate, =
and the estimated delay is allowed to decrease.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">-Dave<o:p></o:p></span=
></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D"><o:p>&nbsp;</o:p></spa=
n></p>
<div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: Taho=
ma, sans-serif; ">From:</span></b><span style=3D"font-size: 10pt; font-fami=
ly: Tahoma, sans-serif; "> Rong Pan (ropan) [<a href=3D"mailto:ropan@cisco.=
com">mailto:ropan@cisco.com</a>]
<br>
<b>Sent:</b> Wednesday, April 01, 2015 6:28 PM<br>
<b>To:</b> Dave Dolson; <a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a><br=
>
<b>Subject:</b> Re: [aqm] PIE departure rate estimation<o:p></o:p></span></=
p>
</div>
</div>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Please see inline=85<o:p=
></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Rong<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black"><o:p>&nbsp;</o:p></span>=
</p>
</div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"color:black">From: </span></b><spa=
n style=3D"color:black">Dave Dolson &lt;<a href=3D"mailto:ddolson@sandvine.=
com">ddolson@sandvine.com</a>&gt;<br>
<b>Date: </b>Wednesday, April 1, 2015 2:43 PM<br>
<b>To: </b>rong &lt;<a href=3D"mailto:ropan@cisco.com">ropan@cisco.com</a>&=
gt;, &quot;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&quot; &lt;<a hr=
ef=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&gt;<br>
<b>Subject: </b>RE: [aqm] PIE departure rate estimation<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">Inline [DD]</span><spa=
n style=3D"color:black"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">&nbsp;</span><span sty=
le=3D"color:black"><o:p></o:p></span></p>
<div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"font-size: 10pt; font-family: Taho=
ma, sans-serif; color: black; ">From:</span></b><span style=3D"font-size: 1=
0pt; font-family: Tahoma, sans-serif; color: black; "> Rong Pan (ropan) [<a=
 href=3D"mailto:ropan@cisco.com">mailto:ropan@cisco.com</a>]
<br>
<b>Sent:</b> Wednesday, April 01, 2015 5:23 PM<br>
<b>To:</b> Dave Dolson; <a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a><br=
>
<b>Subject:</b> Re: [aqm] PIE departure rate estimation</span><span style=
=3D"color:black"><o:p></o:p></span></p>
</div>
</div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Dave,<o:p></o:p></span><=
/p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Thanks for the valuable =
comments. We will see how we can incorporate them. For the comments below, =
please see inline.<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Regards,<o:p></o:p></spa=
n></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Rong<o:p></o:p></span></=
p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
<div style=3D"border:none;border-top:solid #B5C4DF 1.0pt;padding:3.0pt 0in =
0in 0in">
<p class=3D"MsoNormal"><b><span style=3D"color:black">From: </span></b><spa=
n style=3D"color:black">Dave Dolson &lt;<a href=3D"mailto:ddolson@sandvine.=
com">ddolson@sandvine.com</a>&gt;<br>
<b>Date: </b>Wednesday, April 1, 2015 1:22 PM<br>
<b>To: </b>&quot;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&quot; &lt=
;<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a>&gt;<br>
<b>Subject: </b>[aqm] PIE departure rate estimation<o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&nbsp;<=
/span><span style=3D"color:black"><o:p></o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">In <a href=3D"https://to=
ols.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3">
https://tools.ietf.org/html/draft-ietf-aqm-pie-00#section-4.3</a>,<o:p></o:=
p></span></p>
<pre><span style=3D"font-size: 11pt; font-family: Calibri, sans-serif; colo=
r: black; ">it says, </span><span style=3D"color:black">=93We only measure =
the departure rate when there are sufficient data in the buffer=94<o:p></o:=
p></span></pre>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">Why can=92t departure ra=
te be estimated regardless of queue size? Just count packets leaving over t=
ime? I=92m wondering how to avoid the estimate getting stuck at the last va=
lue sampled when the queue had a certain
 quantity in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: We need to make sure that ther=
e are enough data in the queue to guarantee a rate sample. The reason is th=
at if a queue is empty from time to time, we can't measure its true drainin=
g
 rate. The time when the queue empty should be cut out of the drain time ca=
lculation. For simplicity, it is better to make sure we have enough data in=
 the queue to ensure accurate rate measurement sample.</span><span style=3D=
"color:black"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] I understand the =
rationale, but supposing p=3D0.1, but the queue has very few items in it, t=
oo few to obtain a rate calculation. What causes p to go to zero, since an =
update is not permitted? It isn=92t obvious
 to me that the algorithm always stops for all input traffic patterns.</spa=
n><span style=3D"color:black"><o:p></o:p></span></p>
</div>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black"><o:p>&n=
bsp;</o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: There are two calculat=
ions in the PIE algorithm: one is for p update and one is for calculate the=
 rate. They are independent from each other. If there are not enough packet=
s in
 the buffer, we simply don't update depart rate estimation, &nbsp;instead w=
e use the previous value. However, the drop probability calculation is alwa=
ys updated. Hence, the &quot;p&quot; in your example will go down to 0.<o:p=
></o:p></span></p>
</div>
<div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">&nbsp;</span><span sty=
le=3D"color:black"><o:p></o:p></span></p>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&nbsp;<=
/span><span style=3D"color:black"><o:p></o:p></span></p>
</div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">Section 4.2 cites Little=
=92s Law as =93est_del =3D qlen/depart_rate=94,<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">but <a href=3D"http://en=
.wikipedia.org/wiki/Little%27s_law">
according to Wikipedia</a>, the law uses arrival rate, not departure rate.<=
o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">I don=92t know if it mat=
ters (I didn=92t read Little=92s proof), but this gives credence to the sug=
gestion in section 6 that the algorithm could use arrival rate.<o:p></o:p><=
/span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">And I think it might be =
easier to measure when the queue has few items in it.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: what you mentioned is to calcu=
late average number of customers in a system using the arrival.
<a href=3D"http://en.wikipedia.org/wiki/Little's_law">http://en.wikipedia.o=
rg/wiki/Little's_law</a>&nbsp;Wikipedia also mentions that mean responseTim=
e =3D&nbsp;</span><span style=3D"font-size: 10.5pt; font-family: Arial, san=
s-serif; color: rgb(37, 37, 37); background-color: white; background-positi=
on: initial initial; background-repeat: initial initial; ">MeanNumberInSyst=
em
 / MeanThroughput. What we measure is the mean response time (latency). Hen=
ce, it is correct in our draft.&nbsp;</span><span style=3D"color:black"><o:=
p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">[DD] Arrival (post dro=
p) and departure rates are almost never the same at any instant. I guess th=
ey must average to the same value. Practically, it might not matter which i=
s used, but I thought Little=92s law should
 be cited correctly. In Wikipedia the law is cited one way and the example =
is shown another way. Furthermore there is this ominous comment: =93When ex=
ploring Little=92s law and learning to trust it, be aware of the common mis=
takes of using arrivals(work arriving)
 when throughput(work completed) is called for=85=94</span><span style=3D"c=
olor:black"><o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:#1F497D">&nbsp;</span><span sty=
le=3D"color:black"><o:p></o:p></span></p>
</div>
</div>
</div>
<div>
<p class=3D"MsoNormal"><span style=3D"font-size:12.0pt;color:black">&gt;&gt=
;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;&gt;RP: I understand the concern. =
But I think we have used it in the right way. Arrival is related to how man=
y customers/packets in a system. However, given #of customers/packets in a
 system, the wait time is directly related to the service time (inverse of =
throughput), not arrival rate any more.<o:p></o:p></span></p>
</div>
<div>
<div>
<div>
<div>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
<p class=3D"MsoNormal"><span style=3D"color:black">David Dolson<o:p></o:p><=
/span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">Senior Software Architec=
t, Sandvine Inc.<o:p></o:p></span></p>
<p class=3D"MsoNormal"><span style=3D"color:black">&nbsp;<o:p></o:p></span>=
</p>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</span>
</body>
</html>

--_000_D142C6AED9F1ropanciscocom_--


From nobody Thu Apr  2 13:51:20 2015
Return-Path: <g.white@CableLabs.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id CF9A11A1ADD for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 13:51:18 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.226
X-Spam-Level: 
X-Spam-Status: No, score=0.226 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HELO_EQ_MODEMCABLE=0.768, HOST_EQ_MODEMCABLE=1.368, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gou5rolNQGft for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 13:51:17 -0700 (PDT)
Received: from ondar.cablelabs.com (ondar.cablelabs.com [192.160.73.61]) by ietfa.amsl.com (Postfix) with ESMTP id 25D311A1ABC for <aqm@ietf.org>; Thu,  2 Apr 2015 13:51:17 -0700 (PDT)
Received: from kyzyl.cablelabs.com (kyzyl [10.253.0.7]) by ondar.cablelabs.com (8.14.7/8.14.7) with ESMTP id t32KobXs024277; Thu, 2 Apr 2015 14:50:38 -0600
Received: from exchange.cablelabs.com (10.5.0.19) by kyzyl.cablelabs.com (F-Secure/fsigk_smtp/407/kyzyl.cablelabs.com); Thu, 02 Apr 2015 14:50:37 -0600 (MDT)
X-Virus-Status: clean(F-Secure/fsigk_smtp/407/kyzyl.cablelabs.com)
Received: from EXCHANGE.cablelabs.com ([::1]) by EXCHANGE.cablelabs.com ([::1]) with mapi id 14.03.0224.002; Thu, 2 Apr 2015 14:48:49 -0600
From: Greg White <g.white@CableLabs.com>
To: Dave Dolson <ddolson@sandvine.com>, "Scheffenegger, Richard" <rs@netapp.com>, Mikael Abrahamsson <swmike@swm.pp.se>, Szilveszter Nadas <Szilveszter.Nadas@ericsson.com>
Thread-Topic: [aqm] adoption of draft-white-aqm-docsis-pie-01
Thread-Index: AdBR0qmLD+JlLAXWSZym1tRD+kojwQXBCieAAIs1spAAoLDJAA==
Date: Thu, 2 Apr 2015 20:48:48 +0000
Message-ID: <D142ED5C.47F52%g.white@cablelabs.com>
References: <96cffe6791f54a67ae6b6359de6a3d6e@hioexcmbx05-prd.hq.netapp.com> <e9e3d9dc0435474283180936c371e0db@hioexcmbx05-prd.hq.netapp.com> <E8355113905631478EFF04F5AA706E9830BB08C1@wtl-exchp-2.sandvine.com>
In-Reply-To: <E8355113905631478EFF04F5AA706E9830BB08C1@wtl-exchp-2.sandvine.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
user-agent: Microsoft-MacOutlook/14.4.8.150116
x-originating-ip: [10.5.0.27]
Content-Type: text/plain; charset="iso-8859-1"
Content-ID: <9B6C0FC05F33BB459110F2ACF3EB1AE7@cablelabs.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/Ss8_WdcaBOFGBg8wtOkCfod6NyY>
Cc: "'aqm@ietf.org'" <aqm@ietf.org>, "aqm-chairs@tools.ietf.org" <aqm-chairs@tools.ietf.org>
Subject: Re: [aqm] adoption of draft-white-aqm-docsis-pie-01
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 02 Apr 2015 20:51:19 -0000

Thanks for the review Dave.

I double checked with Rong, and the exponential decay of drop probability
was inadvertently removed from the PIE pseudocode (I think she mentioned
this in the other thread as well).

In terms of enable/disable, there are two aspects.  One is that the
DOCSIS-PIE algorithm is enabled by default, but can be disabled via
explicit configuration by the network operator. The other is that the PIE
algorithm has a built-in concept of being ACTIVE vs INACTIVE.  In the
pseudocode of the PIE draft these capitalized terms are used, but in the
description in section 5.1 of the PIE draft, this same functionality is
referred to as "Turning PIE on and off".  Maybe this is an opportunity to
clean up the language in the PIE draft so it is more consistent.
DOCSIS-PIE supports the Active/Inactive aspect, but makes some adjustments
to the logic from what is included in the PIE pseudocode (and uses
different terminology).  I'll add a section that describes the deltas
there and also I'll try to re-align the terminology in the DOCSIS-PIE
pseudocode so that it matches PIE.

-Greg


On 4/1/15, 2:51 PM, "Dave Dolson" <ddolson@sandvine.com> wrote:

>I'm also reviewing draft-ietf-aqm-docsis-pie-00
>(https://tools.ietf.org/html/draft-ietf-aqm-docsis-pie-00)
>
>In section 3.4, an exponential delay aspect of PIE is mentioned. I just
>read https://tools.ietf.org/html/draft-ietf-aqm-pie-00 and saw no mention
>of this exponential delay. Did I miss it, or is it a mistake for the
>docsis document to reference this?
>
>
>draft-ietf-aqm-docsis-pie-00 documents enabling and disabling of PIE
>control. Is enable/disable used with docsis-pie, or is it always on?
>
>
>Otherwise, the document seems to contain enough information to document
>docsis-pie.
>The algorithm in section A.2 has quite a few aspects that are not
>explained, but it seems precise.
>(I don't have the knowledge to comment on whether it is accurate.)
>
>
>-Dave
>
>
>-----Original Message-----
>From: Scheffenegger, Richard [mailto:rs@netapp.com]
>Sent: Friday, March 27, 2015 5:53 PM
>To: Greg White; Mikael Abrahamsson; Dave Dolson; Szilveszter Nadas
>Cc: aqm-chairs@tools.ietf.org; 'aqm@ietf.org'
>Subject: RE: [aqm] adoption of draft-white-aqm-docsis-pie-01
>
>Hi group,
>
>as there haven't been any objections, but some indications of support on
>the list, and based on the responses in the IETF92 meeting in Dallas, we
>chairs think this document can be adopted as WG-item at this time.
>
>
>Greg, can you please upload the most recent version as
>draft-ietf-aqm-docsis-pie-00?
>
>Also, as mentioned in the meeting, and to make true of my promise, I
>would like to invite the following individuals to review this draft, once
>the updated version becomes available.
>
>Mikael Abrahamsson
>Dave Dolson
>Szilveszter N=E1das
>
>Mostly everybody else who has commented during the meeting is already
>assigned to other documents (Nobody will be left out :)
>
>Thanks,
>  Richard (co-chair)
>
>
>
>> -----Original Message-----
>> From: aqm [mailto:aqm-bounces@ietf.org] On Behalf Of Scheffenegger,
>> Richard
>> Sent: Donnerstag, 26. Februar 2015 08:45
>> To: Greg White; 'aqm@ietf.org'
>> Cc: Rong Pan (ropan); aqm-chairs@tools.ietf.org
>> Subject: Re: [aqm] adoption of draft-white-aqm-docsis-pie-01
>>=20
>> Hi Greg, group,
>>=20
>>=20
>>=20
>> Regarding the adoption call, this is something we can actually (and
>> should) start on the list.
>>=20
>> We can confirm that during the Dallas meeting in the room, but even
>>before
>> that, we'd like to get responses on the list now...
>>=20
>> Perhaps we have some volunteers to review this new version as well...
>>=20
>>=20
>> Best regards,
>>   Richard
>>=20
>>=20
>>=20
>> > -----Original Message-----
>> > From: Greg White [mailto:g.white@CableLabs.com]
>> > Sent: Donnerstag, 26. Februar 2015 01:13
>> > To: aqm-chairs@tools.ietf.org
>> > Cc: Rong Pan (ropan)
>> > Subject: Re: [aqm] agenda for IETF 92 / Dallas
>> >
>> > Wes & Richard,
>> >
>> > Unfortunately I will not be at IETF92 in person, but will attend
>> remotely.
>> >  For draft-white-aqm-docsis-pie, Rong & I updated it in January and
>> > included a new appendix to give a change log, which reads:
>> >
>> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>> > Appendix B.  Change Log
>> > B.1.  Since draft-white-aqm-docsis-pie-01
>> >
>> >    Added Change Log.
>> >
>> >    Removed discussion of Packet drop de-randomization, Enhanced burst
>> >    protection, and 16ms update interval, as these are now included in
>> >    [I-D.ietf-aqm-pie].
>> >
>> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>> >
>> >
>> > Regarding WG adoption, I saw some support from Lars (and no
>>objections)
>> to
>> > adopting it on an Informational track.   Will you do an official
>> adoption
>> > call at the Dallas meeting?
>> >
>> > -Greg
>> >
>> >
>> > On 2/23/15, 8:40 PM, "Wesley Eddy" <wes@mti-systems.com> wrote:
>> >
>> > >Greetings AQMers!  We requested a short AQM meeting slot at the
>> > >upcoming IETF 92 meeting in Dallas, and we received a 1-hour slot on
>> Tuesday:
>> > >https://datatracker.ietf.org/meeting/92/agenda.html
>> > >
>> > >Since this is not a lot of time, we'll need to prioritize the work
>> > >that is discussed to what requires the face-to-face time in order to
>> > >make progress on.
>> > >
>> > >For active draft editors, please let us know via this list or
>> > >aqm-chairs@tools.ietf.org what you think your meeting time needs are,
>> > >so we can put together an agenda.  If you don't need meeting time
>> > >now, or would like to use an interim telecon to assist in gathering
>> > >feedback, please also let us know.
>> > >
>> > >For others, please read the drafts, comment on this mailing list, and
>> > >help us to review and complete them:
>> > >https://datatracker.ietf.org/wg/aqm/documents/
>> > >
>> > >Thanks in advance!
>> > >
>> > >--
>> > >Wes Eddy
>> > >MTI Systems
>> > >
>> > >_______________________________________________
>> > >aqm mailing list
>> > >aqm@ietf.org
>> > >https://www.ietf.org/mailman/listinfo/aqm
>>=20
>> _______________________________________________
>> aqm mailing list
>> aqm@ietf.org
>> https://www.ietf.org/mailman/listinfo/aqm


From nobody Thu Apr  2 14:49:59 2015
Return-Path: <ropan@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id B50BE1A7005 for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 14:49:57 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -14.511
X-Spam-Level: 
X-Spam-Status: No, score=-14.511 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Vw2a0_n3yEdc for <aqm@ietfa.amsl.com>; Thu,  2 Apr 2015 14:49:55 -0700 (PDT)
Received: from alln-iport-2.cisco.com (alln-iport-2.cisco.com [173.37.142.89]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 22BB21A6FEE for <aqm@ietf.org>; Thu,  2 Apr 2015 14:49:55 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=6909; q=dns/txt; s=iport; t=1428011395; x=1429220995; h=from:to:cc:subject:date:message-id:references: in-reply-to:content-id:content-transfer-encoding: mime-version; bh=hS1kTA8m55417xoj5tNHvDKDfn5fCfpjl8tAtUu1lRQ=; b=SQaP946q4nWzWG5FPHIA1wf2DPkiUjH28jcv9WcxTTxKrma1vj4PUlbg UN7GwP2tmle2Q1e/HyCR7KpL8/geJL261pa+q3LmVNvD4BCLNRs9rw5Rd E6W8bWLmurubPS6yKhKdBNnXtq33Aq/Kiok+LzrUnXglHDskqxKY+xKes g=;
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0AxBQCnuB1V/5tdJa1QDIMIUlwFxVoKhX0CgUxMAQEBAQEBfoQeAQEBBAEBARpRCwwEAgEIDgMEAQEBJwcnCxQJCAEBBAENBYguAQ3NUwEBAQEBAQEBAQEBAQEBAQEBAQEBARMEiip/hBEBZwIFBoQnBYpxg1mCIIN1gieDZIEdgzSCXlWFR4M1g0giggMcgVBvAYFDfwEBAQ
X-IronPort-AV: E=Sophos;i="5.11,513,1422921600"; d="scan'208";a="137827683"
Received: from rcdn-core-4.cisco.com ([173.37.93.155]) by alln-iport-2.cisco.com with ESMTP; 02 Apr 2015 21:49:54 +0000
Received: from xhc-aln-x05.cisco.com (xhc-aln-x05.cisco.com [173.36.12.79]) by rcdn-core-4.cisco.com (8.14.5/8.14.5) with ESMTP id t32LnsPA010538 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Thu, 2 Apr 2015 21:49:54 GMT
Received: from xmb-aln-x15.cisco.com ([169.254.9.109]) by xhc-aln-x05.cisco.com ([173.36.12.79]) with mapi id 14.03.0195.001; Thu, 2 Apr 2015 16:49:53 -0500
From: "Rong Pan (ropan)" <ropan@cisco.com>
To: Greg White <g.white@CableLabs.com>, Dave Dolson <ddolson@sandvine.com>, "Scheffenegger, Richard" <rs@netapp.com>, Mikael Abrahamsson <swmike@swm.pp.se>, Szilveszter Nadas <Szilveszter.Nadas@ericsson.com>
Thread-Topic: [aqm] adoption of draft-white-aqm-docsis-pie-01
Thread-Index: AdBR0qmLD+JlLAXWSZym1tRD+kojwQXBCieAAIs1spAAoLDJAP//74WA
Date: Thu, 2 Apr 2015 21:49:53 +0000
Message-ID: <D143071B.DA34%ropan@cisco.com>
References: <96cffe6791f54a67ae6b6359de6a3d6e@hioexcmbx05-prd.hq.netapp.com> <e9e3d9dc0435474283180936c371e0db@hioexcmbx05-prd.hq.netapp.com> <E8355113905631478EFF04F5AA706E9830BB08C1@wtl-exchp-2.sandvine.com> <D142ED5C.47F52%g.white@cablelabs.com>
In-Reply-To: <D142ED5C.47F52%g.white@cablelabs.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
user-agent: Microsoft-MacOutlook/14.4.7.141117
x-originating-ip: [171.70.247.185]
Content-Type: text/plain; charset="iso-8859-1"
Content-ID: <680A089E985D444C9715132B5341A79A@emea.cisco.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/LxahrRMWqXGTYy3ML895_YJ61pY>
Cc: "'aqm@ietf.org'" <aqm@ietf.org>, "aqm-chairs@tools.ietf.org" <aqm-chairs@tools.ietf.org>
Subject: Re: [aqm] adoption of draft-white-aqm-docsis-pie-01
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 02 Apr 2015 21:49:57 -0000

Greg,

Good suggestion of keeping DOCSIS-PIE and PIE drafts consistent. Let's
sync up and make sure they are consistent.

Thanks,

Rong

On 4/2/15 1:48 PM, "Greg White" <g.white@CableLabs.com> wrote:

>Thanks for the review Dave.
>
>I double checked with Rong, and the exponential decay of drop probability
>was inadvertently removed from the PIE pseudocode (I think she mentioned
>this in the other thread as well).
>
>In terms of enable/disable, there are two aspects.  One is that the
>DOCSIS-PIE algorithm is enabled by default, but can be disabled via
>explicit configuration by the network operator. The other is that the PIE
>algorithm has a built-in concept of being ACTIVE vs INACTIVE.  In the
>pseudocode of the PIE draft these capitalized terms are used, but in the
>description in section 5.1 of the PIE draft, this same functionality is
>referred to as "Turning PIE on and off".  Maybe this is an opportunity to
>clean up the language in the PIE draft so it is more consistent.
>DOCSIS-PIE supports the Active/Inactive aspect, but makes some adjustments
>to the logic from what is included in the PIE pseudocode (and uses
>different terminology).  I'll add a section that describes the deltas
>there and also I'll try to re-align the terminology in the DOCSIS-PIE
>pseudocode so that it matches PIE.
>
>-Greg
>
>
>On 4/1/15, 2:51 PM, "Dave Dolson" <ddolson@sandvine.com> wrote:
>
>>I'm also reviewing draft-ietf-aqm-docsis-pie-00
>>(https://tools.ietf.org/html/draft-ietf-aqm-docsis-pie-00)
>>
>>In section 3.4, an exponential delay aspect of PIE is mentioned. I just
>>read https://tools.ietf.org/html/draft-ietf-aqm-pie-00 and saw no mention
>>of this exponential delay. Did I miss it, or is it a mistake for the
>>docsis document to reference this?
>>
>>
>>draft-ietf-aqm-docsis-pie-00 documents enabling and disabling of PIE
>>control. Is enable/disable used with docsis-pie, or is it always on?
>>
>>
>>Otherwise, the document seems to contain enough information to document
>>docsis-pie.
>>The algorithm in section A.2 has quite a few aspects that are not
>>explained, but it seems precise.
>>(I don't have the knowledge to comment on whether it is accurate.)
>>
>>
>>-Dave
>>
>>
>>-----Original Message-----
>>From: Scheffenegger, Richard [mailto:rs@netapp.com]
>>Sent: Friday, March 27, 2015 5:53 PM
>>To: Greg White; Mikael Abrahamsson; Dave Dolson; Szilveszter Nadas
>>Cc: aqm-chairs@tools.ietf.org; 'aqm@ietf.org'
>>Subject: RE: [aqm] adoption of draft-white-aqm-docsis-pie-01
>>
>>Hi group,
>>
>>as there haven't been any objections, but some indications of support on
>>the list, and based on the responses in the IETF92 meeting in Dallas, we
>>chairs think this document can be adopted as WG-item at this time.
>>
>>
>>Greg, can you please upload the most recent version as
>>draft-ietf-aqm-docsis-pie-00?
>>
>>Also, as mentioned in the meeting, and to make true of my promise, I
>>would like to invite the following individuals to review this draft, once
>>the updated version becomes available.
>>
>>Mikael Abrahamsson
>>Dave Dolson
>>Szilveszter N=E1das
>>
>>Mostly everybody else who has commented during the meeting is already
>>assigned to other documents (Nobody will be left out :)
>>
>>Thanks,
>>  Richard (co-chair)
>>
>>
>>
>>> -----Original Message-----
>>> From: aqm [mailto:aqm-bounces@ietf.org] On Behalf Of Scheffenegger,
>>> Richard
>>> Sent: Donnerstag, 26. Februar 2015 08:45
>>> To: Greg White; 'aqm@ietf.org'
>>> Cc: Rong Pan (ropan); aqm-chairs@tools.ietf.org
>>> Subject: Re: [aqm] adoption of draft-white-aqm-docsis-pie-01
>>>=20
>>> Hi Greg, group,
>>>=20
>>>=20
>>>=20
>>> Regarding the adoption call, this is something we can actually (and
>>> should) start on the list.
>>>=20
>>> We can confirm that during the Dallas meeting in the room, but even
>>>before
>>> that, we'd like to get responses on the list now...
>>>=20
>>> Perhaps we have some volunteers to review this new version as well...
>>>=20
>>>=20
>>> Best regards,
>>>   Richard
>>>=20
>>>=20
>>>=20
>>> > -----Original Message-----
>>> > From: Greg White [mailto:g.white@CableLabs.com]
>>> > Sent: Donnerstag, 26. Februar 2015 01:13
>>> > To: aqm-chairs@tools.ietf.org
>>> > Cc: Rong Pan (ropan)
>>> > Subject: Re: [aqm] agenda for IETF 92 / Dallas
>>> >
>>> > Wes & Richard,
>>> >
>>> > Unfortunately I will not be at IETF92 in person, but will attend
>>> remotely.
>>> >  For draft-white-aqm-docsis-pie, Rong & I updated it in January and
>>> > included a new appendix to give a change log, which reads:
>>> >
>>> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>> > Appendix B.  Change Log
>>> > B.1.  Since draft-white-aqm-docsis-pie-01
>>> >
>>> >    Added Change Log.
>>> >
>>> >    Removed discussion of Packet drop de-randomization, Enhanced burst
>>> >    protection, and 16ms update interval, as these are now included in
>>> >    [I-D.ietf-aqm-pie].
>>> >
>>> > =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>> >
>>> >
>>> > Regarding WG adoption, I saw some support from Lars (and no
>>>objections)
>>> to
>>> > adopting it on an Informational track.   Will you do an official
>>> adoption
>>> > call at the Dallas meeting?
>>> >
>>> > -Greg
>>> >
>>> >
>>> > On 2/23/15, 8:40 PM, "Wesley Eddy" <wes@mti-systems.com> wrote:
>>> >
>>> > >Greetings AQMers!  We requested a short AQM meeting slot at the
>>> > >upcoming IETF 92 meeting in Dallas, and we received a 1-hour slot on
>>> Tuesday:
>>> > >https://datatracker.ietf.org/meeting/92/agenda.html
>>> > >
>>> > >Since this is not a lot of time, we'll need to prioritize the work
>>> > >that is discussed to what requires the face-to-face time in order to
>>> > >make progress on.
>>> > >
>>> > >For active draft editors, please let us know via this list or
>>> > >aqm-chairs@tools.ietf.org what you think your meeting time needs
>>>are,
>>> > >so we can put together an agenda.  If you don't need meeting time
>>> > >now, or would like to use an interim telecon to assist in gathering
>>> > >feedback, please also let us know.
>>> > >
>>> > >For others, please read the drafts, comment on this mailing list,
>>>and
>>> > >help us to review and complete them:
>>> > >https://datatracker.ietf.org/wg/aqm/documents/
>>> > >
>>> > >Thanks in advance!
>>> > >
>>> > >--
>>> > >Wes Eddy
>>> > >MTI Systems
>>> > >
>>> > >_______________________________________________
>>> > >aqm mailing list
>>> > >aqm@ietf.org
>>> > >https://www.ietf.org/mailman/listinfo/aqm
>>>=20
>>> _______________________________________________
>>> aqm mailing list
>>> aqm@ietf.org
>>> https://www.ietf.org/mailman/listinfo/aqm
>
>_______________________________________________
>aqm mailing list
>aqm@ietf.org
>https://www.ietf.org/mailman/listinfo/aqm


From nobody Wed Apr  8 23:51:28 2015
Return-Path: <gregskinner0@icloud.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 964321A00EF for <aqm@ietfa.amsl.com>; Wed,  8 Apr 2015 23:51:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.466
X-Spam-Level: *
X-Spam-Status: No, score=1.466 tagged_above=-999 required=5 tests=[BAYES_50=0.8, LOTS_OF_MONEY=0.001, SPF_SOFTFAIL=0.665] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2U3mC4s5x4Up for <aqm@ietfa.amsl.com>; Wed,  8 Apr 2015 23:51:25 -0700 (PDT)
Received: from gds.best.vwh.net (gds.best.vwh.net [192.220.66.135]) by ietfa.amsl.com (Postfix) with ESMTP id 8F0AA1A00E8 for <aqm@ietf.org>; Wed,  8 Apr 2015 23:51:25 -0700 (PDT)
Received: (qmail 74457 invoked by uid 17031); 9 Apr 2015 06:51:24 -0000
Received: from unknown (HELO [192.168.1.100]) (gds@[198.144.201.12]) (envelope-sender <gregskinner0@icloud.com>) by 192.220.66.135 (qmail-ldap-1.03) with AES256-SHA encrypted SMTP for <dave.taht@gmail.com>; 9 Apr 2015 06:51:24 -0000
Content-Type: text/plain; charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2070.6\))
From: Greg Skinner <gregskinner0@icloud.com>
In-Reply-To: <CAA93jw4p-jYt_XLN+NQbAGyLWVAHuRWcVab0CpJBkmpE5QoTBA@mail.gmail.com>
Date: Wed, 8 Apr 2015 23:51:05 -0700
Content-Transfer-Encoding: quoted-printable
Message-Id: <4077D77B-6C1A-47DF-989C-76B4B99AF863@icloud.com>
References: <CAA93jw4zZXGn_z2_dBBMsKN-XdwP88Wt-ChvfWJgCtp=J+i0rg@mail.gmail.com> <473265656416337848@unknownmsgid> <CAA93jw4p-jYt_XLN+NQbAGyLWVAHuRWcVab0CpJBkmpE5QoTBA@mail.gmail.com>
To: Dave Taht <dave.taht@gmail.com>
X-Mailer: Apple Mail (2.2070.6)
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/QswnLEADm-bhPgStiXqKf5gWbSM>
Cc: Vishal Misra <misra@cs.columbia.edu>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] the cisco pie patent and IETF IPR filing
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2015 06:51:27 -0000

Sorry I haven=E2=80=99t had a chance to comment on this until now.  =
Also, I=E2=80=99m not on the bloat list, but you may copy any responses =
to me there, if you wish.

I was thinking perhaps another approach to spreading the word could be =
presenting talks at conferences aimed at CIOs and CTOs.  They are =
generally interested in saving money and getting better performance from =
equipment that=E2=80=99s already in use.  The algorithms and techniques =
discussed here do that.  Getting sysadmins or netadmins to make config =
changes to enable the sysctl settings and/or upgrade kernels (in =
production) might be more difficult, because they don=E2=80=99t always =
have the authority to make these types of changes, even though they have =
the privileges.  But a CIO or CTO has the authority to request that =
these types of changes be made permanent, especially if it improves the =
bottom line.

=E2=80=94gregbo

> On Mar 5, 2015, at 5:58 AM, Dave Taht <dave.taht@gmail.com> wrote:
>=20
> On Wed, Mar 4, 2015 at 12:17 AM, Vishal Misra <misra@cs.columbia.edu> =
wrote:
>> Hi Dave,
>>=20
>> Thanks for your email. A few quick points:
>>=20
>> - I have actually sent a note already to someone on the Cisco PIE =
team
>> about the error in the IETF IPR filing and am sure they will get it
>> corrected. You have helpfully dug out the actual patent application
>> and it appears that one digit got inadvertently changed in the Cisco
>> IETF IPR declaration of the patent application.
>>=20
>> - I wish I had a "marketing department" that would do stories for me
>> :-). I work at Columbia University and that story that you point out
>> was done by a writer at the UMass-Amherst engineering school as an
>> example of academic research having practical impact. There is an
>> urgent need to support more academic research and I think stories =
like
>> this one support the cause.
>=20
> Well, yes and no. One thing I have tried really hard to do throughout
> this project is give credit where credit is due, at every talk for
> example, always mentioning pie, even before I actually had any data on
> it's performance.- I try to give every individual that has contributed
> something to this "stone soup" project, as here at uknof -
>=20
> https://plus.google.com/u/0/107942175615993706558/posts/immF8Pkj19C
>=20
> praise - for what they did to help out. There have been an amazing
> level of details to sort out along the way here at every level in the
> OS stack, and in the hardware and there is simply no one individual or
> company I would single out as truly key, except maybe George P.
> Burdell!
>=20
> A lesson I have learned is that folk in marketing are not particularly
> good at correctly distributing credit, and I assume that is how they
> are taught to write, to not look at any facts outside of their
> immediate objectives. [1]
>=20
> =
http://newsroom.cisco.com/feature-content?type=3Dwebcontent&articleId=3D14=
14442
>=20
> and 'course nobody in the press has shown up with a photographer to
> write puff pieces about the overall effort except, well, cringely's
> work is not puffy enough by marketing standards:  (
> http://www.cringely.com/tag/bufferbloat/ )
>=20
> I admit to a great deal of frustration when nick weaver writes an
> otherwise *excellent* piece in forbes,
> =
http://www.forbes.com/sites/valleyvoices/2015/02/27/this-one-clause-in-the=
-new-net-neutrality-regs-would-be-a-fiasco-for-the-internet/
>=20
> and expends 3+ paragraphs explaining bufferbloat, but never gives the
> reader a link back *to the word* so that maybe, some CTO or CEO that
> reads that rag would have some context and clue when an engineer comes
> up to him asking for permission to go implement a fix that is now,
> basically, off the shelf.
>=20
> *I* am going to keep giving credit to everyone I can, in every talk
> and presentation I do, and there are quite a few core contributors
> that I wish I had called out by name more - for example, I would have
> mentioned felix feitkau's contribution towards fixing wifi at the
> nznog talk if I could correctly pronounce his name! I struggled for
> years to be able to pronounce juliusz's!
>=20
> At the very least, I hope we can do more from a SEO perspective - and
> all *pull together* to get the message out - that bufferbloat is
> fixed, that solutions are being standardized in the ietf, and the code
> is widely available on a ton of platforms already - and move to
> somehow get to where ISPs are announcing settings for things like
> openwrt + sqm-scripts, and more importantly - schedules for rolling
> out fixes (like docsis 3.1 and better CPE) to their customers.
>=20
> everyone:
>=20
> What else more can we do here to cross the chasm?
>=20
>> - Indeed neither me nor any of the other PI authors had any idea of
>> the PIE work. I discovered it accidentally when I was at MIT giving a
>> talk on Network Neutrality and Dave Clark mentioned Cisco's PIE and
>> DOCSIS 3.1 to me. I later read up on PIE and was pleasantly surprised
>> that our PI work from more than a decade back evolved into it.
>>=20
>> - I had contributed the PI code to Sally Floyd back in 2001 and it =
has
>> been part of ns2 for the longest time (pi.cc). It shouldn't be
>> difficult to adapt that for a Linux implementation and I am happy to
>> help anyone who wishes to try it. Maybe that might affect your =
loyalty
>> to fq_codel.
>=20
> I let the data take me where it may. I (not) always have, but reformed
> about 15 years ago. [1] I hope that you and your students also,  do
> some experiments on the successors to PI and RED and DRR - and also
> follow the data where-ever it leads you.
>=20
> I was fiercely proud of sfqred - until fq_codel blew it away on every
> benchmark I could devise. I have long longed to find another
> independent expert in the field to create new experiments and/or
> recreate/reproduce/disprove our results.
>=20
> [1] "For a successful technology, reality must take precedence over
> public relations, for Nature cannot be fooled. " - Richard P. Feyman,
> Challenger Disaster report:
> https://www.youtube.com/watch?v=3D6Rwcbsn19c0
>=20
>=20
>> -Vishal
>> --
>> http://www.cs.columbia.edu/~misra/
>>=20
>>> On Mar 4, 2015, at 1:07 AM, Dave Taht <dave.taht@gmail.com> wrote:
>>>=20
>>> Two items:
>>>=20
>>> A) The IETF IPR filing   http://datatracker.ietf.org/ipr/2187/ =
points
>>> to the wrong patent: 13/874,500. A google search for that patent
>>> number brings up http://www.google.com/patents/US20130239255"
>>>=20
>>> It is ironically relevant to the discussions at hand, as that one =
concerns:
>>>=20
>>> Abstract:
>>>=20
>>> "Provided are methods of increasing the tolerance of a plant to
>>> abiotic stresses and/or increasing the biomass and/or increasing the
>>> yield of a plant by expressing within the plant an exogenous
>>> polynucleotide homologous to SEQ ID NO:13."
>>>=20
>>> ... As I consider myself a near-vegetable, and am 40 pounds heavier,
>>> and not responding particularly well to antibiotics, after
>>> participating for the past several years on all the ietf mailing =
lists
>>> I just got off of. I am sure that upon acceptance of pie in the =
ietf,
>>> that making that particular patent more generally available for all =
to
>>> use would probably have similar effects on others.
>>>=20
>>> The correct patent number for PIE, 13/874,600, is here:
>>>=20
>>> http://www.google.com/patents/US20140328175
>>>=20
>>> I would appreciate that the IPR filing be corrected.
>>>=20
>>> In the meantime, here's some more great NSFW george carlin routines!
>>>=20
>>> https://www.youtube.com/watch?v=3DtVlkxrNlp10
>>>=20
>>> B) Vishal Misra (author of PI) gave me pointers to his PI papers
>>> recently (and he had NO idea at all his work was used for pie! - he
>>> got his marketing department to issue a press release about it:
>>> =
http://engineering.umass.edu/news/got-bufferbloat-umass-amherst-research-b=
ehind-fix
>>> )
>>>=20
>>> I usually have a pretty strict policy about never reading patents, =
but
>>> I read all those papers [1], and both! patents above. I had not =
fully
>>> realized that the PI-AQM work went as far back as 2001. The PI =
update
>>> equation and the PIE update equation, look pretty darn similar, just
>>> the meanings of two variables, changed.
>>>=20
>>> C) I am kind of curious if any working code for the original PI
>>> algorithm exists for linux?
>>>=20
>>> D) oh, never mind, I will blog about the rest one day.
>>>=20
>>> [1] still prefer fq_codel.
>>>=20
>>> --
>>> Dave T=C3=A4ht
>>> Let's make wifi fast, less jittery and reliable again!
>>>=20
>>> https://plus.google.com/u/0/107942175615993706558/posts/TVX3o84jjmb
>>=20
>=20
>=20
>=20
> --=20
> Dave T=C3=A4ht
> Let's make wifi fast, less jittery and reliable again!
>=20
> https://plus.google.com/u/0/107942175615993706558/posts/TVX3o84jjmb
>=20
>=20


From nobody Thu Apr  9 15:47:04 2015
Return-Path: <dave.taht@gmail.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id B4D181B30C0 for <aqm@ietfa.amsl.com>; Thu,  9 Apr 2015 12:24:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.301
X-Spam-Level: *
X-Spam-Status: No, score=1.301 tagged_above=-999 required=5 tests=[BAYES_50=0.8, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, J_CHICKENPOX_16=0.6, LOTS_OF_MONEY=0.001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nAj7B0XA0Krn for <aqm@ietfa.amsl.com>; Thu,  9 Apr 2015 12:24:04 -0700 (PDT)
Received: from mail-ob0-x22b.google.com (mail-ob0-x22b.google.com [IPv6:2607:f8b0:4003:c01::22b]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id C88AF1B30CF for <aqm@ietf.org>; Thu,  9 Apr 2015 12:24:03 -0700 (PDT)
Received: by obbfy7 with SMTP id fy7so30236544obb.2 for <aqm@ietf.org>; Thu, 09 Apr 2015 12:24:03 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; bh=8AEfDn4J3Qhd3zmvgpWklpa+ufNnPoAvnK+0EjRzTKI=; b=frBBk5vJGIQygEBmqEllYcJmtA2AhRzbEQRNPmGbebBdck1t7hDl8POaCR94pY0JBc dIGFeOiwuJqu+Tl8cNQ+GrqOI161v7ZZFFcIXE8sTeKsIlSaCtBp1DDBox8ho0cnD95A cmL8M+h7a6pjBrr3kvsO6vGEE8HY6DlO1gglFCjIJ+uoAWvXI3SwPr2/7g+0bAlp7Vrr RaUMMAjT4keMtlwhb+5LtDqNBZQBo1KsSq+M9L0rhWrEnw2Jl99tDX4a6whm1jCTaajF Bk8MsPfUlt14yiNpXi5odi5iOhbZ+rlS4TxyKG7T3v819B4KnH2QwKlZ/VaXr0IcRQT7 dw2g==
MIME-Version: 1.0
X-Received: by 10.60.103.234 with SMTP id fz10mr40481575oeb.11.1428607443241;  Thu, 09 Apr 2015 12:24:03 -0700 (PDT)
Received: by 10.202.51.66 with HTTP; Thu, 9 Apr 2015 12:24:03 -0700 (PDT)
In-Reply-To: <4077D77B-6C1A-47DF-989C-76B4B99AF863@icloud.com>
References: <CAA93jw4zZXGn_z2_dBBMsKN-XdwP88Wt-ChvfWJgCtp=J+i0rg@mail.gmail.com> <473265656416337848@unknownmsgid> <CAA93jw4p-jYt_XLN+NQbAGyLWVAHuRWcVab0CpJBkmpE5QoTBA@mail.gmail.com> <4077D77B-6C1A-47DF-989C-76B4B99AF863@icloud.com>
Date: Thu, 9 Apr 2015 12:24:03 -0700
Message-ID: <CAA93jw65PVgCta-3yhEo3YfsAhJUE-9E0+oNoXmaT6C1c4DChA@mail.gmail.com>
From: Dave Taht <dave.taht@gmail.com>
To: Greg Skinner <gregskinner0@icloud.com>, bloat <bloat@lists.bufferbloat.net>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/1V1fLm3KQLBJSJ8MTZSlBMSusvo>
X-Mailman-Approved-At: Thu, 09 Apr 2015 15:46:58 -0700
Cc: Vishal Misra <misra@cs.columbia.edu>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] the cisco pie patent and IETF IPR filing
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2015 19:24:07 -0000

On Wed, Apr 8, 2015 at 11:51 PM, Greg Skinner <gregskinner0@icloud.com> wro=
te:
> Sorry I haven=E2=80=99t had a chance to comment on this until now.  Also,=
 I=E2=80=99m not on the bloat list, but you may copy any responses to me th=
ere, if you wish.
>
> I was thinking perhaps another approach to spreading the word could be pr=
esenting talks at conferences aimed at CIOs and CTOs.  They are generally i=
nterested in saving money and getting better performance from equipment tha=
t=E2=80=99s already in use.  The algorithms and techniques discussed here d=
o that.

I agree that we are more ready to talk to those sorts of folk now.
However I personally am tired of travelling and talking and would
prefer someone else(es) be "Speaker-to-CTOs", and get all the airline
miles, fame and fortune for a while, maybe getting in some golf time
as well.

I would really LOVE it if more people here gave talks and attended
conferences, user group meetings, etc.

A problem here is a lot of people take our bufferbloat-beating
algorithms as someone selling yet another form of snake oil - and it
is worse - here, we are not making snake oil - we are *giving it away*
- and if it doesn't work, you can have your money back! I have had to
go to sceptic=C2=B4s houses and install the software, after doing a
measurement, and forcing our tools on to j.random persons laptops, in
order to make the points we make. That does not scale, although it
usually turns a skeptic into an enthusiastic supporter.

I expended the majority of my budget on travel in the first few years,
primarily to find and talk to experts in the field, to make sure we
were not going to break the internet. I am pretty sure the Internet
will be less broken with more of our stuff in play, than not, now.

A concern to me, in finding a "Speaker-to-CTOs" is to find someone who
can present the solutions in a way even more neutrally than I do, and
yet convince people to actually go deploy them. I make no bones (these
days!) about how I feel that fq + aqm + ecn (fq_codel=C2=B4s default mode)
is safe to deploy, for example, and there are many people here that
disagree with me, thinking that aqm + ecn alone is safe to deploy, and
obviously I promote fq_codel derived stuff on routers (and sch_fq on
well protected servers), far more than pie, although I am careful to
maintain that I can=C2=B4t wait to see pie (without ecn) on cablemodems one
day RSN.

I have NO idea to what extent a product having or not having a given
shaping/fq/aqm algorithm could be used as a sales tool for people
selling it, and actually selling something that worked would be a big
motivator for some future speaker to chase after the CTO/CIO/CEOs at
their conferences. Certainly ubnt=C2=B4s customers are digging it...

>Getting sysadmins or netadmins to make config changes to enable the sysctl=
 settings and/or upgrade kernels (in production) might be more difficult, b=
ecause they don=E2=80=99t always have the authority to make these types of =
changes, even though they have the privileges.

In the good ole days sysadmins were eager to do anything they could to
improve network behavior and connectivity, and did not have to filter
every decision through management, a lawyer, and an accountant.

Recently at the NZNOG conference there was a whole session where
provider after provider announced their AS number and willingness to
peer with anybody. I nearly had to wipe away tears of remembered joy
for the good ole days where we cared more about connectivity and good
service rather than anything else.

In the general case, strategically, I have preferred to work bottom up
rather than top down, and to accumulate enough operational experience
to then be able to approach top management with a fait accompli (it is
easier to ask for forgiveness than permission) and permission to go
full scale on a rollout with results in hand.

This strategy is sort of working in multiple places, notably google
uses this approach effectively (and distros like openwrt, free.fr,
etc, fedora 22 and anything running systemd without patches, like
arch, all defaulting to fq_codel rather than pfifo_fast now!).

There are several target audiences:

A) hardware makers: I would like it if more firewall makers like
barracuda were paying attention, and we had something that worked on
pfsense. As happy as I am about ubnt=C2=B4s enthusiastic support for
fq_codel, there has not been a peep from multiple other places.
B) Chip designers: Here, aside from whatever is going on with
streamboost at qca, I am not sure if anyone is paying attention.
C) China: Folk making gear there, notably EEs, would benefit from more
awareness. I long for a good bufferbloat talk or article written in
multiple other languages, including the main ones there, japanese, and
spanish.
D) Big web service providers and CDN folk - where sch_fq is
potentially a big deal (being as it is so popular already)
E) IT staffs in general - nanog has an upcoming conference that would
be good to have a talk at.
F) VOIP and webrtc service providers - here we made a huge dent
already in the asterisk and freeswitch communities. I was asked to do
a followup at the next cluecon, but as I will be out of the country -
if anyone would like to give a talk there please contact me offlist.
G) Load balancer makers. There is some sign that F5 is paying
attention, at least, but what others are doing here is unknown.
F) ISPs

And way too many conferences to attend.

> But a CIO or CTO has the authority to request that these types of changes=
 be made permanent, especially if it improves the bottom line.

I agree that we need to start talking to more of these folk, at a
level they can understand, with arguments that involve saving money
with better, faster, less latent networks.

> =E2=80=94gregbo
>
>> On Mar 5, 2015, at 5:58 AM, Dave Taht <dave.taht@gmail.com> wrote:
>>
>> On Wed, Mar 4, 2015 at 12:17 AM, Vishal Misra <misra@cs.columbia.edu> wr=
ote:
>>> Hi Dave,
>>>
>>> Thanks for your email. A few quick points:
>>>
>>> - I have actually sent a note already to someone on the Cisco PIE team
>>> about the error in the IETF IPR filing and am sure they will get it
>>> corrected. You have helpfully dug out the actual patent application
>>> and it appears that one digit got inadvertently changed in the Cisco
>>> IETF IPR declaration of the patent application.
>>>
>>> - I wish I had a "marketing department" that would do stories for me
>>> :-). I work at Columbia University and that story that you point out
>>> was done by a writer at the UMass-Amherst engineering school as an
>>> example of academic research having practical impact. There is an
>>> urgent need to support more academic research and I think stories like
>>> this one support the cause.
>>
>> Well, yes and no. One thing I have tried really hard to do throughout
>> this project is give credit where credit is due, at every talk for
>> example, always mentioning pie, even before I actually had any data on
>> it's performance.- I try to give every individual that has contributed
>> something to this "stone soup" project, as here at uknof -
>>
>> https://plus.google.com/u/0/107942175615993706558/posts/immF8Pkj19C
>>
>> praise - for what they did to help out. There have been an amazing
>> level of details to sort out along the way here at every level in the
>> OS stack, and in the hardware and there is simply no one individual or
>> company I would single out as truly key, except maybe George P.
>> Burdell!
>>
>> A lesson I have learned is that folk in marketing are not particularly
>> good at correctly distributing credit, and I assume that is how they
>> are taught to write, to not look at any facts outside of their
>> immediate objectives. [1]
>>
>> http://newsroom.cisco.com/feature-content?type=3Dwebcontent&articleId=3D=
1414442
>>
>> and 'course nobody in the press has shown up with a photographer to
>> write puff pieces about the overall effort except, well, cringely's
>> work is not puffy enough by marketing standards:  (
>> http://www.cringely.com/tag/bufferbloat/ )
>>
>> I admit to a great deal of frustration when nick weaver writes an
>> otherwise *excellent* piece in forbes,
>> http://www.forbes.com/sites/valleyvoices/2015/02/27/this-one-clause-in-t=
he-new-net-neutrality-regs-would-be-a-fiasco-for-the-internet/
>>
>> and expends 3+ paragraphs explaining bufferbloat, but never gives the
>> reader a link back *to the word* so that maybe, some CTO or CEO that
>> reads that rag would have some context and clue when an engineer comes
>> up to him asking for permission to go implement a fix that is now,
>> basically, off the shelf.
>>
>> *I* am going to keep giving credit to everyone I can, in every talk
>> and presentation I do, and there are quite a few core contributors
>> that I wish I had called out by name more - for example, I would have
>> mentioned felix feitkau's contribution towards fixing wifi at the
>> nznog talk if I could correctly pronounce his name! I struggled for
>> years to be able to pronounce juliusz's!
>>
>> At the very least, I hope we can do more from a SEO perspective - and
>> all *pull together* to get the message out - that bufferbloat is
>> fixed, that solutions are being standardized in the ietf, and the code
>> is widely available on a ton of platforms already - and move to
>> somehow get to where ISPs are announcing settings for things like
>> openwrt + sqm-scripts, and more importantly - schedules for rolling
>> out fixes (like docsis 3.1 and better CPE) to their customers.
>>
>> everyone:
>>
>> What else more can we do here to cross the chasm?
>>
>>> - Indeed neither me nor any of the other PI authors had any idea of
>>> the PIE work. I discovered it accidentally when I was at MIT giving a
>>> talk on Network Neutrality and Dave Clark mentioned Cisco's PIE and
>>> DOCSIS 3.1 to me. I later read up on PIE and was pleasantly surprised
>>> that our PI work from more than a decade back evolved into it.
>>>
>>> - I had contributed the PI code to Sally Floyd back in 2001 and it has
>>> been part of ns2 for the longest time (pi.cc). It shouldn't be
>>> difficult to adapt that for a Linux implementation and I am happy to
>>> help anyone who wishes to try it. Maybe that might affect your loyalty
>>> to fq_codel.
>>
>> I let the data take me where it may. I (not) always have, but reformed
>> about 15 years ago. [1] I hope that you and your students also,  do
>> some experiments on the successors to PI and RED and DRR - and also
>> follow the data where-ever it leads you.
>>
>> I was fiercely proud of sfqred - until fq_codel blew it away on every
>> benchmark I could devise. I have long longed to find another
>> independent expert in the field to create new experiments and/or
>> recreate/reproduce/disprove our results.
>>
>> [1] "For a successful technology, reality must take precedence over
>> public relations, for Nature cannot be fooled. " - Richard P. Feyman,
>> Challenger Disaster report:
>> https://www.youtube.com/watch?v=3D6Rwcbsn19c0
>>
>>
>>> -Vishal
>>> --
>>> http://www.cs.columbia.edu/~misra/
>>>
>>>> On Mar 4, 2015, at 1:07 AM, Dave Taht <dave.taht@gmail.com> wrote:
>>>>
>>>> Two items:
>>>>
>>>> A) The IETF IPR filing   http://datatracker.ietf.org/ipr/2187/ points
>>>> to the wrong patent: 13/874,500. A google search for that patent
>>>> number brings up http://www.google.com/patents/US20130239255"
>>>>
>>>> It is ironically relevant to the discussions at hand, as that one conc=
erns:
>>>>
>>>> Abstract:
>>>>
>>>> "Provided are methods of increasing the tolerance of a plant to
>>>> abiotic stresses and/or increasing the biomass and/or increasing the
>>>> yield of a plant by expressing within the plant an exogenous
>>>> polynucleotide homologous to SEQ ID NO:13."
>>>>
>>>> ... As I consider myself a near-vegetable, and am 40 pounds heavier,
>>>> and not responding particularly well to antibiotics, after
>>>> participating for the past several years on all the ietf mailing lists
>>>> I just got off of. I am sure that upon acceptance of pie in the ietf,
>>>> that making that particular patent more generally available for all to
>>>> use would probably have similar effects on others.
>>>>
>>>> The correct patent number for PIE, 13/874,600, is here:
>>>>
>>>> http://www.google.com/patents/US20140328175
>>>>
>>>> I would appreciate that the IPR filing be corrected.
>>>>
>>>> In the meantime, here's some more great NSFW george carlin routines!
>>>>
>>>> https://www.youtube.com/watch?v=3DtVlkxrNlp10
>>>>
>>>> B) Vishal Misra (author of PI) gave me pointers to his PI papers
>>>> recently (and he had NO idea at all his work was used for pie! - he
>>>> got his marketing department to issue a press release about it:
>>>> http://engineering.umass.edu/news/got-bufferbloat-umass-amherst-resear=
ch-behind-fix
>>>> )
>>>>
>>>> I usually have a pretty strict policy about never reading patents, but
>>>> I read all those papers [1], and both! patents above. I had not fully
>>>> realized that the PI-AQM work went as far back as 2001. The PI update
>>>> equation and the PIE update equation, look pretty darn similar, just
>>>> the meanings of two variables, changed.
>>>>
>>>> C) I am kind of curious if any working code for the original PI
>>>> algorithm exists for linux?
>>>>
>>>> D) oh, never mind, I will blog about the rest one day.
>>>>
>>>> [1] still prefer fq_codel.
>>>>
>>>> --
>>>> Dave T=C3=A4ht
>>>> Let's make wifi fast, less jittery and reliable again!
>>>>
>>>> https://plus.google.com/u/0/107942175615993706558/posts/TVX3o84jjmb
>>>
>>
>>
>>
>> --
>> Dave T=C3=A4ht
>> Let's make wifi fast, less jittery and reliable again!
>>
>> https://plus.google.com/u/0/107942175615993706558/posts/TVX3o84jjmb
>>
>>
>



--=20
Dave T=C3=A4ht
Open Networking needs **Open Source Hardware**

https://plus.google.com/u/0/107942175615993706558/posts/N8mZ5F5iSPU


From nobody Thu Apr  9 15:47:05 2015
Return-Path: <dave.taht@gmail.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id BF1B31B30F6 for <aqm@ietfa.amsl.com>; Thu,  9 Apr 2015 12:29:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.6
X-Spam-Level: 
X-Spam-Status: No, score=-0.6 tagged_above=-999 required=5 tests=[BAYES_05=-0.5, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FREEMAIL_FROM=0.001, SPF_PASS=-0.001] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id DzqC42MKLnsQ for <aqm@ietfa.amsl.com>; Thu,  9 Apr 2015 12:28:59 -0700 (PDT)
Received: from mail-ob0-x235.google.com (mail-ob0-x235.google.com [IPv6:2607:f8b0:4003:c01::235]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id F33A61A8F40 for <aqm@ietf.org>; Thu,  9 Apr 2015 12:28:58 -0700 (PDT)
Received: by oblw8 with SMTP id w8so147488176obl.0 for <aqm@ietf.org>; Thu, 09 Apr 2015 12:28:58 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; bh=vJA3Gi1F/toYj4j2q0CqLXjY04ytvLQeAMmgns5eTnA=; b=wfnYM9BltmDzrB3JYdV7NhpEKco9Sslx825hEDVhT+UDMiXd67WeMply4scbbppNNK JXKlntkc9ejVJDLeRHLzWQphMEOM9cw55FEwCW/TwlImJBtOy3j4w0leYAeeykYAoq6k x+f9eEsTlGjVzSE7I17PzusuYnopSBPQJrS070DynJxJvlKx5ExVpYVTgTbj0brtmlXT QI2YdjYHOMuS+eKkL6AUJzYezUjQksL3qZIGgpzivYcMEERXxLhvttpPBnV7TFWWhWLa 3alkjQIaswwOFFJyUA8yKtdUcjKlK/N6upCjg6mWIaD9PfYvVJcxnbPNiacGAdy7xpWr h4kw==
MIME-Version: 1.0
X-Received: by 10.60.173.201 with SMTP id bm9mr40106956oec.34.1428607738515; Thu, 09 Apr 2015 12:28:58 -0700 (PDT)
Received: by 10.202.51.66 with HTTP; Thu, 9 Apr 2015 12:28:58 -0700 (PDT)
In-Reply-To: <4077D77B-6C1A-47DF-989C-76B4B99AF863@icloud.com>
References: <CAA93jw4zZXGn_z2_dBBMsKN-XdwP88Wt-ChvfWJgCtp=J+i0rg@mail.gmail.com> <473265656416337848@unknownmsgid> <CAA93jw4p-jYt_XLN+NQbAGyLWVAHuRWcVab0CpJBkmpE5QoTBA@mail.gmail.com> <4077D77B-6C1A-47DF-989C-76B4B99AF863@icloud.com>
Date: Thu, 9 Apr 2015 12:28:58 -0700
Message-ID: <CAA93jw5-yEwZ1otBoN=xYhhOktZ7i5kkzDtwKc_Hewy4PL+k=w@mail.gmail.com>
From: Dave Taht <dave.taht@gmail.com>
To: Greg Skinner <gregskinner0@icloud.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/MrpfDGvGjlxcyuYAkWYeTG1OQOM>
X-Mailman-Approved-At: Thu, 09 Apr 2015 15:46:59 -0700
Cc: bloat <bloat@lists.bufferbloat.net>, Vishal Misra <misra@cs.columbia.edu>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] the cisco pie patent and IETF IPR filing
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 09 Apr 2015 19:29:00 -0000

On Wed, Apr 8, 2015 at 11:51 PM, Greg Skinner <gregskinner0@icloud.com> wro=
te:
> Sorry I haven=E2=80=99t had a chance to comment on this until now.  Also,=
 I=E2=80=99m not on the bloat list, but you may copy any responses to me th=
ere, if you wish.
>
> I was thinking perhaps another approach to spreading the word could be pr=
esenting talks at conferences aimed at CIOs and CTOs.  They are generally i=
nterested in saving money and getting better performance from equipment tha=
t=E2=80=99s already in use.  The algorithms and techniques discussed here d=
o that.

Well, more directly to this point - they can do it on Linux, on modern
distros, on all cpus supported. Those using with other OSes or older
embedded gear or top brand switchs/routers like cisco, juniper, or
arista are all still in a hole, as are most of the users of head end
BRASes, CMTSes, etc. It would be nice to have the big folk announcing
solutions and plans and so on before nagging the CTOS and CIOS with
stuff they can=C2=B4t just buy. Telling someone that they have to switch
their entire edge network over to Linux is often a hard sell.

I do keep hoping for a tilera or a cavium to pour this stuff into
their chips and firmware, but so far no such luck.

Blatant plug: see my .sig

--=20
Dave T=C3=A4ht
Open Networking needs **Open Source Hardware**

https://plus.google.com/u/0/107942175615993706558/posts/N8mZ5F5iSPU


From nobody Sun Apr 12 21:33:43 2015
Return-Path: <andrewmcgr@google.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 65B771A0439 for <aqm@ietfa.amsl.com>; Sun, 12 Apr 2015 21:33:42 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.312
X-Spam-Level: *
X-Spam-Status: No, score=1.312 tagged_above=-999 required=5 tests=[BAYES_50=0.8, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id igTxo4lmeA6j for <aqm@ietfa.amsl.com>; Sun, 12 Apr 2015 21:33:30 -0700 (PDT)
Received: from mail-qk0-x22d.google.com (mail-qk0-x22d.google.com [IPv6:2607:f8b0:400d:c09::22d]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 2336B1A0145 for <aqm@ietf.org>; Sun, 12 Apr 2015 21:33:30 -0700 (PDT)
Received: by qku63 with SMTP id 63so159671438qku.3 for <aqm@ietf.org>; Sun, 12 Apr 2015 21:33:29 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=Lv1ZlB8T8/PuSYlCH96RujeLJS6QmDQJSGMeP0/mxM8=; b=k5tPVQBuWVfmZIPqlh8TA/wIxBa8aFHlAWI+mke/1ZFWotu+h9xpaMgmBcu7ywji5f fKCn8iOykdrguD7SPiwa2bz9pI5+zK5lsPjY8u3R31MeRVVdZVpEaR8zoNaPIGpqe1Rl gwXYVk+7II5lnWh/RN3v/zhv1CSG53jw9l4qczHN0bjfWVBqpqz58DMW6sfIETSwqabe wx5IeZ51OzMlBjinBAC4Pqoe8PrOodUopH+Kvct3VdMffQ7+D6Tp/ELmNS+5I/2fKJkd eE0U2QYmtfHU8gdXILUARbAJ+X0tyS5YnJw+fuwhuB0FsZZ1t7oddnKrwXJv5iHHWpIN eS+A==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:date :message-id:subject:from:to:cc:content-type; bh=Lv1ZlB8T8/PuSYlCH96RujeLJS6QmDQJSGMeP0/mxM8=; b=FbnWmOKhqQK99cGu8yrbldxjBHj2oa+QEvIh/kSf8cMeQDNNe8XS/hS08TGjNAb636 ssnkqAe6qNtG7zoQlIfmcpaNiP04Lj8BJ0sEjKG5BY4wEd5yqIqdq4msMQeyHsdbsvga FzPsmJ5vUTQdpAmn0KO1fSgXhEgxCHVHRYRtv5woaIFqBrOeRgX58VZBiYT4BwPZLV4a DELZ+WDqHbcxTqUBEedkLxRBNw4wQaLaxGSpVnOIPxjN2qdM+jW9PcGc0hw6xA8K7d10 ovrrcyUqea1rB2q5LiAwG0kNakJ78Xl2ipp5Sn6gIWzdwfHG/ZX3a7Kpjo7tPtGqzUCO t5xQ==
X-Gm-Message-State: ALoCoQlTRu6ANPCUDxAC/0B+BvLUCyufvEBfTX8bHSYoyx3DkEhrIXd+O5pDeBUIs9DFV1GtTQN7
MIME-Version: 1.0
X-Received: by 10.55.42.219 with SMTP id q88mr25820026qkq.20.1428899609065; Sun, 12 Apr 2015 21:33:29 -0700 (PDT)
Received: by 10.96.139.232 with HTTP; Sun, 12 Apr 2015 21:33:28 -0700 (PDT)
In-Reply-To: <5F53F60C-8374-4A38-9D68-25647ED8428F@cisco.com>
References: <5F53F60C-8374-4A38-9D68-25647ED8428F@cisco.com>
Date: Mon, 13 Apr 2015 14:33:28 +1000
Message-ID: <CAPRuP3m3b+YOeBgKqz5Y+eEVzYGkxGhL4x7H2R91b2Ua+dv7Ng@mail.gmail.com>
From: Andrew Mcgregor <andrewmcgr@google.com>
To: "Fred Baker (fred)" <fred@cisco.com>
Content-Type: multipart/alternative; boundary=001a1147ba8c9e59190513939d48
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/3g3MyBANb4MDZdBWZs1vs4kGD-Q>
Cc: "draft-ietf-aqm-codel.all@tools.ietf.org" <draft-ietf-aqm-codel.all@tools.ietf.org>, "aqm@ietf.org list" <aqm@ietf.org>
Subject: Re: [aqm] codel spelling nits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 04:33:42 -0000

--001a1147ba8c9e59190513939d48
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Thanks Fred, will incorporate those.

On 2 April 2015 at 14:34, Fred Baker (fred) <fred@cisco.com> wrote:

> I=E2=80=99m in the process, as I volunteered to do, of reviewing the Code=
l draft.
> It is very well and readably written, so I don=E2=80=99t expect to have a=
 lot to
> complain about. I do have three points, though.
>
> One is a few spelling errors, noted in this email.
>
> A second is the matter of marking and dropping. The world =E2=80=9Cdrop=
=E2=80=9D is used
> in one or another of its forms 120 times, and ECN marking twice, both tim=
es
> in passing. I=E2=80=99d like to understand better how ECN is handled, if =
only to
> have a statement that it is handled the same way as dropping.
>
> The third is the definition of a =E2=80=9Cnormal TCP session=E2=80=9D, wh=
ich is referred
> to a number of times but not nailed down that I have found other than tha=
t
> it appears to have a less than 100ms RTT. Later in the draft, it talks
> about accommodations for data center use, which tells me (consistent with
> our testing) that there is a lower bound below which is is less than
> optimal, and in our testing RTTs greater than 100 ms didn=E2=80=99t behav=
e so well.
> The sweet spot might have been around 25 ms, with minor degradation at 50
> ms, and more degradation at longer RTTs.
>
> I=E2=80=99ll come back if I find more issues,
>
> *** draft/draft-ietf-aqm-codel-00.txt   Fri Oct 24 08:58:40 2014
> --- draft-ietf-aqm-codel-00.txt Wed Apr  1 18:43:33 2015
> ***************
> *** 104,111 ****
>       5.1.  Data Types  . . . . . . . . . . . . . . . . . . . . . . .  18
>       5.2.  Per-queue state (codel_queue_t instance variables)  . . .  19
>       5.3.  Constants . . . . . . . . . . . . . . . . . . . . . . . .  19
> !      5.4.  Enque routine . . . . . . . . . . . . . . . . . . . . . .  1=
9
> !      5.5.  Deque routine . . . . . . . . . . . . . . . . . . . . . .  1=
9
>
>
>
> --- 104,111 ----
>       5.1.  Data Types  . . . . . . . . . . . . . . . . . . . . . . .  18
>       5.2.  Per-queue state (codel_queue_t instance variables)  . . .  19
>       5.3.  Constants . . . . . . . . . . . . . . . . . . . . . . . .  19
> !      5.4.  Enqueue routine . . . . . . . . . . . . . . . . . . . . . .
> 19
> !      5.5.  Dequeue routine . . . . . . . . . . . . . . . . . . . . . .
> 19
>
>
>
> ***************
> *** 283,289 ****
>
>
>     is much slower than the link that feeds it (say, a high-speed
> !    ethernet link into a limited DSL uplink) a 20 packet buffer at the
>     bottleneck might be necessary to temporarily hold the 20 packets in
>     flight to keep the utilization high.  The burst of packets should
>     drain completely (to 0 or 1 packets) within a round trip time and
> --- 283,289 ----
>
>
>     is much slower than the link that feeds it (say, a high-speed
> !    Ethernet link into a limited DSL uplink) a 20 packet buffer at the
>     bottleneck might be necessary to temporarily hold the 20 packets in
>     flight to keep the utilization high.  The burst of packets should
>     drain completely (to 0 or 1 packets) within a round trip time and
> ***************
> *** 326,332 ****
>
>     The use of queue length is further complicated in networks that are
>     subject to both short and long term changes in available link rate
> !    (as in wifi).  Link rate drops can result in a spike in queue length
>     that should be ignored unless it persists.  The length metric is
>     problematic when what we really want to control is the amount of
>     excess delay packets experience due to a persistent or standing
> --- 326,332 ----
>
>     The use of queue length is further complicated in networks that are
>     subject to both short and long term changes in available link rate
> !    (as in WiFi).  Link rate drops can result in a spike in queue length
>     that should be ignored unless it persists.  The length metric is
>     problematic when what we really want to control is the amount of
>     excess delay packets experience due to a persistent or standing
> ***************
> *** 455,461 ****
>     As Kleinrock observed, the best operating point, in terms of
>     bandwidth / delay tradeoff, is the peak power point since points off
>     the peak represent a higher cost (in delay) per unit of bandwidth.
> !    The power vs. _f_ curve for any AIMD TCP is monotone decreasing.  Bu=
t
>     the curve is very flat for _f_ < 0.1 followed by a increasing
>     curvature with a knee around .2 then a steep, almost linear fall off
>     [TSV84] [VJTARG14].  Since the previous equation showed that goodput
> --- 455,461 ----
>     As Kleinrock observed, the best operating point, in terms of
>     bandwidth / delay tradeoff, is the peak power point since points off
>     the peak represent a higher cost (in delay) per unit of bandwidth.
> !    The power vs. _f_ curve for any AIMD TCP is monotonically
> decreasing.  But
>     the curve is very flat for _f_ < 0.1 followed by a increasing
>     curvature with a knee around .2 then a steep, almost linear fall off
>     [TSV84] [VJTARG14].  Since the previous equation showed that goodput
> ***************
> *** 929,935 ****
>     management as described here or to adapt its principles to other
>     applications.
>
> !    Implementors are strongly encouraged to also look at Eric Dumazet's
>     Linux kernel version of CoDel - a well-written, well tested, real-
>     world, C-based implementation.  As of this writing, it is at:
>
> --- 929,935 ----
>     management as described here or to adapt its principles to other
>     applications.
>
> !    Implementers are strongly encouraged to also look at Eric Dumazet's
>     Linux kernel version of CoDel - a well-written, well tested, real-
>     world, C-based implementation.  As of this writing, it is at:
>
> ***************
> *** 999,1006 ****
>
>     "packet_t*" is a pointer to a packet descriptor.  We assume it has a
>     tstamp field capable of holding a time_t and that field is available
> !    for use by CoDel (it will be set by the enque routine and used by th=
e
> !    deque routine).
>
>
>
> --- 999,1006 ----
>
>     "packet_t*" is a pointer to a packet descriptor.  We assume it has a
>     tstamp field capable of holding a time_t and that field is available
> !    for use by CoDel (it will be set by the enqueue routine and used by
> the
> !    dequeue routine).
>
>
>
> ***************
> *** 1033,1039 ****
>     u_int maxpacket =3D 512; // Maximum packet size in bytes
>                      // (should use interface MTU)
>
> ! 5.4.  Enque routine
>
>     All the work of CoDel is done in the deque routine.  The only CoDel
>     addition to enque is putting the current time in the packet's tstamp
> --- 1033,1039 ----
>     u_int maxpacket =3D 512; // Maximum packet size in bytes
>                      // (should use interface MTU)
>
> ! 5.4.  Enqueue routine
>
>     All the work of CoDel is done in the deque routine.  The only CoDel
>     addition to enque is putting the current time in the packet's tstamp
> ***************
> *** 1046,1052 ****
>         queue_t::enque(pkt);
>     }
>
> ! 5.5.  Deque routine
>
>     This is the heart of CoDel.  There are two branches: In packet-
>     dropping state (meaning that the queue-sojourn time has gone above
> --- 1046,1052 ----
>         queue_t::enque(pkt);
>     }
>
> ! 5.5.  Dequeue routine
>
>     This is the heart of CoDel.  There are two branches: In packet-
>     dropping state (meaning that the queue-sojourn time has gone above
> ***************
> *** 1260,1266 ****
>
>     An experiment by Stanford graduate students successfully used the
>     linux CoDel to duplicate our published simulation work on CoDel's
> !    ability to following drastic link rate changes which can be found at=
:
>     http://reproducingnetworkresearch.wordpress.com/2012/06/06/solving-
>     bufferbloat-the-codel-way/ .
>
> --- 1260,1266 ----
>
>     An experiment by Stanford graduate students successfully used the
>     linux CoDel to duplicate our published simulation work on CoDel's
> !    ability to follow drastic link rate changes which can be found at:
>     http://reproducingnetworkresearch.wordpress.com/2012/06/06/solving-
>     bufferbloat-the-codel-way/ .
>
>


--=20
Andrew McGregor | SRE | andrewmcgr@google.com | +61 4 1071 2221

--001a1147ba8c9e59190513939d48
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Thanks Fred, will incorporate those.</div><div class=3D"gm=
ail_extra"><br><div class=3D"gmail_quote">On 2 April 2015 at 14:34, Fred Ba=
ker (fred) <span dir=3D"ltr">&lt;<a href=3D"mailto:fred@cisco.com" target=
=3D"_blank">fred@cisco.com</a>&gt;</span> wrote:<br><blockquote class=3D"gm=
ail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-le=
ft:1ex">I=E2=80=99m in the process, as I volunteered to do, of reviewing th=
e Codel draft. It is very well and readably written, so I don=E2=80=99t exp=
ect to have a lot to complain about. I do have three points, though.<br>
<br>
One is a few spelling errors, noted in this email.<br>
<br>
A second is the matter of marking and dropping. The world =E2=80=9Cdrop=E2=
=80=9D is used in one or another of its forms 120 times, and ECN marking tw=
ice, both times in passing. I=E2=80=99d like to understand better how ECN i=
s handled, if only to have a statement that it is handled the same way as d=
ropping.<br>
<br>
The third is the definition of a =E2=80=9Cnormal TCP session=E2=80=9D, whic=
h is referred to a number of times but not nailed down that I have found ot=
her than that it appears to have a less than 100ms RTT. Later in the draft,=
 it talks about accommodations for data center use, which tells me (consist=
ent with our testing) that there is a lower bound below which is is less th=
an optimal, and in our testing RTTs greater than 100 ms didn=E2=80=99t beha=
ve so well. The sweet spot might have been around 25 ms, with minor degrada=
tion at 50 ms, and more degradation at longer RTTs.<br>
<br>
I=E2=80=99ll come back if I find more issues,<br>
<br>
*** draft/draft-ietf-aqm-codel-00.txt=C2=A0 =C2=A0Fri Oct 24 08:58:40 2014<=
br>
--- draft-ietf-aqm-codel-00.txt Wed Apr=C2=A0 1 18:43:33 2015<br>
***************<br>
*** 104,111 ****<br>
=C2=A0 =C2=A0 =C2=A0 5.1.=C2=A0 Data Types=C2=A0 . . . . . . . . . . . . . =
. . . . . . . . . .=C2=A0 18<br>
=C2=A0 =C2=A0 =C2=A0 5.2.=C2=A0 Per-queue state (codel_queue_t instance var=
iables)=C2=A0 . . .=C2=A0 19<br>
=C2=A0 =C2=A0 =C2=A0 5.3.=C2=A0 Constants . . . . . . . . . . . . . . . . .=
 . . . . . . .=C2=A0 19<br>
!=C2=A0 =C2=A0 =C2=A0 5.4.=C2=A0 Enque routine . . . . . . . . . . . . . . =
. . . . . . . .=C2=A0 19<br>
!=C2=A0 =C2=A0 =C2=A0 5.5.=C2=A0 Deque routine . . . . . . . . . . . . . . =
. . . . . . . .=C2=A0 19<br>
<br>
<br>
<br>
--- 104,111 ----<br>
=C2=A0 =C2=A0 =C2=A0 5.1.=C2=A0 Data Types=C2=A0 . . . . . . . . . . . . . =
. . . . . . . . . .=C2=A0 18<br>
=C2=A0 =C2=A0 =C2=A0 5.2.=C2=A0 Per-queue state (codel_queue_t instance var=
iables)=C2=A0 . . .=C2=A0 19<br>
=C2=A0 =C2=A0 =C2=A0 5.3.=C2=A0 Constants . . . . . . . . . . . . . . . . .=
 . . . . . . .=C2=A0 19<br>
!=C2=A0 =C2=A0 =C2=A0 5.4.=C2=A0 Enqueue routine . . . . . . . . . . . . . =
. . . . . . . . .=C2=A0 19<br>
!=C2=A0 =C2=A0 =C2=A0 5.5.=C2=A0 Dequeue routine . . . . . . . . . . . . . =
. . . . . . . . .=C2=A0 19<br>
<br>
<br>
<br>
***************<br>
*** 283,289 ****<br>
<br>
<br>
=C2=A0 =C2=A0 is much slower than the link that feeds it (say, a high-speed=
<br>
!=C2=A0 =C2=A0 ethernet link into a limited DSL uplink) a 20 packet buffer =
at the<br>
=C2=A0 =C2=A0 bottleneck might be necessary to temporarily hold the 20 pack=
ets in<br>
=C2=A0 =C2=A0 flight to keep the utilization high.=C2=A0 The burst of packe=
ts should<br>
=C2=A0 =C2=A0 drain completely (to 0 or 1 packets) within a round trip time=
 and<br>
--- 283,289 ----<br>
<br>
<br>
=C2=A0 =C2=A0 is much slower than the link that feeds it (say, a high-speed=
<br>
!=C2=A0 =C2=A0 Ethernet link into a limited DSL uplink) a 20 packet buffer =
at the<br>
=C2=A0 =C2=A0 bottleneck might be necessary to temporarily hold the 20 pack=
ets in<br>
=C2=A0 =C2=A0 flight to keep the utilization high.=C2=A0 The burst of packe=
ts should<br>
=C2=A0 =C2=A0 drain completely (to 0 or 1 packets) within a round trip time=
 and<br>
***************<br>
*** 326,332 ****<br>
<br>
=C2=A0 =C2=A0 The use of queue length is further complicated in networks th=
at are<br>
=C2=A0 =C2=A0 subject to both short and long term changes in available link=
 rate<br>
!=C2=A0 =C2=A0 (as in wifi).=C2=A0 Link rate drops can result in a spike in=
 queue length<br>
=C2=A0 =C2=A0 that should be ignored unless it persists.=C2=A0 The length m=
etric is<br>
=C2=A0 =C2=A0 problematic when what we really want to control is the amount=
 of<br>
=C2=A0 =C2=A0 excess delay packets experience due to a persistent or standi=
ng<br>
--- 326,332 ----<br>
<br>
=C2=A0 =C2=A0 The use of queue length is further complicated in networks th=
at are<br>
=C2=A0 =C2=A0 subject to both short and long term changes in available link=
 rate<br>
!=C2=A0 =C2=A0 (as in WiFi).=C2=A0 Link rate drops can result in a spike in=
 queue length<br>
=C2=A0 =C2=A0 that should be ignored unless it persists.=C2=A0 The length m=
etric is<br>
=C2=A0 =C2=A0 problematic when what we really want to control is the amount=
 of<br>
=C2=A0 =C2=A0 excess delay packets experience due to a persistent or standi=
ng<br>
***************<br>
*** 455,461 ****<br>
=C2=A0 =C2=A0 As Kleinrock observed, the best operating point, in terms of<=
br>
=C2=A0 =C2=A0 bandwidth / delay tradeoff, is the peak power point since poi=
nts off<br>
=C2=A0 =C2=A0 the peak represent a higher cost (in delay) per unit of bandw=
idth.<br>
!=C2=A0 =C2=A0 The power vs. _f_ curve for any AIMD TCP is monotone decreas=
ing.=C2=A0 But<br>
=C2=A0 =C2=A0 the curve is very flat for _f_ &lt; 0.1 followed by a increas=
ing<br>
=C2=A0 =C2=A0 curvature with a knee around .2 then a steep, almost linear f=
all off<br>
=C2=A0 =C2=A0 [TSV84] [VJTARG14].=C2=A0 Since the previous equation showed =
that goodput<br>
--- 455,461 ----<br>
=C2=A0 =C2=A0 As Kleinrock observed, the best operating point, in terms of<=
br>
=C2=A0 =C2=A0 bandwidth / delay tradeoff, is the peak power point since poi=
nts off<br>
=C2=A0 =C2=A0 the peak represent a higher cost (in delay) per unit of bandw=
idth.<br>
!=C2=A0 =C2=A0 The power vs. _f_ curve for any AIMD TCP is monotonically de=
creasing.=C2=A0 But<br>
=C2=A0 =C2=A0 the curve is very flat for _f_ &lt; 0.1 followed by a increas=
ing<br>
=C2=A0 =C2=A0 curvature with a knee around .2 then a steep, almost linear f=
all off<br>
=C2=A0 =C2=A0 [TSV84] [VJTARG14].=C2=A0 Since the previous equation showed =
that goodput<br>
***************<br>
*** 929,935 ****<br>
=C2=A0 =C2=A0 management as described here or to adapt its principles to ot=
her<br>
=C2=A0 =C2=A0 applications.<br>
<br>
!=C2=A0 =C2=A0 Implementors are strongly encouraged to also look at Eric Du=
mazet&#39;s<br>
=C2=A0 =C2=A0 Linux kernel version of CoDel - a well-written, well tested, =
real-<br>
=C2=A0 =C2=A0 world, C-based implementation.=C2=A0 As of this writing, it i=
s at:<br>
<br>
--- 929,935 ----<br>
=C2=A0 =C2=A0 management as described here or to adapt its principles to ot=
her<br>
=C2=A0 =C2=A0 applications.<br>
<br>
!=C2=A0 =C2=A0 Implementers are strongly encouraged to also look at Eric Du=
mazet&#39;s<br>
=C2=A0 =C2=A0 Linux kernel version of CoDel - a well-written, well tested, =
real-<br>
=C2=A0 =C2=A0 world, C-based implementation.=C2=A0 As of this writing, it i=
s at:<br>
<br>
***************<br>
*** 999,1006 ****<br>
<br>
=C2=A0 =C2=A0 &quot;packet_t*&quot; is a pointer to a packet descriptor.=C2=
=A0 We assume it has a<br>
=C2=A0 =C2=A0 tstamp field capable of holding a time_t and that field is av=
ailable<br>
!=C2=A0 =C2=A0 for use by CoDel (it will be set by the enque routine and us=
ed by the<br>
!=C2=A0 =C2=A0 deque routine).<br>
<br>
<br>
<br>
--- 999,1006 ----<br>
<br>
=C2=A0 =C2=A0 &quot;packet_t*&quot; is a pointer to a packet descriptor.=C2=
=A0 We assume it has a<br>
=C2=A0 =C2=A0 tstamp field capable of holding a time_t and that field is av=
ailable<br>
!=C2=A0 =C2=A0 for use by CoDel (it will be set by the enqueue routine and =
used by the<br>
!=C2=A0 =C2=A0 dequeue routine).<br>
<br>
<br>
<br>
***************<br>
*** 1033,1039 ****<br>
=C2=A0 =C2=A0 u_int maxpacket =3D 512; // Maximum packet size in bytes<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0// (should use interface MTU)<br>
<br>
! 5.4.=C2=A0 Enque routine<br>
<br>
=C2=A0 =C2=A0 All the work of CoDel is done in the deque routine.=C2=A0 The=
 only CoDel<br>
=C2=A0 =C2=A0 addition to enque is putting the current time in the packet&#=
39;s tstamp<br>
--- 1033,1039 ----<br>
=C2=A0 =C2=A0 u_int maxpacket =3D 512; // Maximum packet size in bytes<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0// (should use interface MTU)<br>
<br>
! 5.4.=C2=A0 Enqueue routine<br>
<br>
=C2=A0 =C2=A0 All the work of CoDel is done in the deque routine.=C2=A0 The=
 only CoDel<br>
=C2=A0 =C2=A0 addition to enque is putting the current time in the packet&#=
39;s tstamp<br>
***************<br>
*** 1046,1052 ****<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 queue_t::enque(pkt);<br>
=C2=A0 =C2=A0 }<br>
<br>
! 5.5.=C2=A0 Deque routine<br>
<br>
=C2=A0 =C2=A0 This is the heart of CoDel.=C2=A0 There are two branches: In =
packet-<br>
=C2=A0 =C2=A0 dropping state (meaning that the queue-sojourn time has gone =
above<br>
--- 1046,1052 ----<br>
=C2=A0 =C2=A0 =C2=A0 =C2=A0 queue_t::enque(pkt);<br>
=C2=A0 =C2=A0 }<br>
<br>
! 5.5.=C2=A0 Dequeue routine<br>
<br>
=C2=A0 =C2=A0 This is the heart of CoDel.=C2=A0 There are two branches: In =
packet-<br>
=C2=A0 =C2=A0 dropping state (meaning that the queue-sojourn time has gone =
above<br>
***************<br>
*** 1260,1266 ****<br>
<br>
=C2=A0 =C2=A0 An experiment by Stanford graduate students successfully used=
 the<br>
=C2=A0 =C2=A0 linux CoDel to duplicate our published simulation work on CoD=
el&#39;s<br>
!=C2=A0 =C2=A0 ability to following drastic link rate changes which can be =
found at:<br>
=C2=A0 =C2=A0 <a href=3D"http://reproducingnetworkresearch.wordpress.com/20=
12/06/06/solving-" target=3D"_blank">http://reproducingnetworkresearch.word=
press.com/2012/06/06/solving-</a><br>
=C2=A0 =C2=A0 bufferbloat-the-codel-way/ .<br>
<br>
--- 1260,1266 ----<br>
<br>
=C2=A0 =C2=A0 An experiment by Stanford graduate students successfully used=
 the<br>
=C2=A0 =C2=A0 linux CoDel to duplicate our published simulation work on CoD=
el&#39;s<br>
!=C2=A0 =C2=A0 ability to follow drastic link rate changes which can be fou=
nd at:<br>
=C2=A0 =C2=A0 <a href=3D"http://reproducingnetworkresearch.wordpress.com/20=
12/06/06/solving-" target=3D"_blank">http://reproducingnetworkresearch.word=
press.com/2012/06/06/solving-</a><br>
=C2=A0 =C2=A0 bufferbloat-the-codel-way/ .<br>
<br>
</blockquote></div><br><br clear=3D"all"><div><br></div>-- <br><div class=
=3D"gmail_signature"><div dir=3D"ltr"><span style=3D"color:rgb(85,85,85);fo=
nt-family:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px=
 0px;border-style:solid;border-color:rgb(213,15,37);padding-top:2px;margin-=
top:2px">Andrew McGregor=C2=A0|</span><span style=3D"color:rgb(85,85,85);fo=
nt-family:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px=
 0px;border-style:solid;border-color:rgb(51,105,232);padding-top:2px;margin=
-top:2px">=C2=A0SRE=C2=A0|</span><span style=3D"color:rgb(85,85,85);font-fa=
mily:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px 0px;=
border-style:solid;border-color:rgb(0,153,57);padding-top:2px;margin-top:2p=
x">=C2=A0<a href=3D"mailto:andrewmcgr@google.com" target=3D"_blank">andrewm=
cgr@google.com</a>=C2=A0|</span><span style=3D"color:rgb(85,85,85);font-fam=
ily:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px 0px;b=
order-style:solid;border-color:rgb(238,178,17);padding-top:2px;margin-top:2=
px">=C2=A0+61 4 1071 2221</span><br></div></div>
</div>

--001a1147ba8c9e59190513939d48--


From nobody Mon Apr 13 07:38:10 2015
Return-Path: <gettysjim@gmail.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 177111AC432 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 07:38:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.382
X-Spam-Level: **
X-Spam-Status: No, score=2.382 tagged_above=-999 required=5 tests=[BAYES_50=0.8, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, FM_FORGED_GMAIL=0.622, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, SPF_PASS=-0.001, URI_TRY_3LD=0.959] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id gLDi4d4MOUH0 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 07:38:00 -0700 (PDT)
Received: from mail-wi0-x22c.google.com (mail-wi0-x22c.google.com [IPv6:2a00:1450:400c:c05::22c]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 84B8E1AC436 for <aqm@ietf.org>; Mon, 13 Apr 2015 07:37:54 -0700 (PDT)
Received: by wizk4 with SMTP id k4so75447769wiz.1 for <aqm@ietf.org>; Mon, 13 Apr 2015 07:37:52 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:sender:date:message-id:subject:from:to:content-type;  bh=UuEDmlJ2/6xslgA7d37Lf6b3A5DRsTjHde6B2pr05Rk=; b=IqD167s4UDLwPOjacIqSIscvzUFkHHSQ7Edypil9Q62Z6PJkFBR/jJwrluskgmkOy1 p64vvsRdzWhDkbh43PWlWp5mzDI0yvQXP4jLCrgixEtzO6GaQcfiaJcW//v1GDaWzlVa Oixlt+JHeeRrUQd/LJHHy/KlHwuPTiq6HnQbxEqKuzg2clgU5VYDMwtWuxVOqvL3UqnB VoocktBIZh0b+92F2imZJoqAcdhS4cR+ieJg1j+tU9MORXVi01DfrTv9Dk8WhkF6f9Sv tkGcg1Z4nSIwkojxjsWnFfMUgH/BrQgPMpG4wm/nWBLQf5wj26+9ZohJfJjK9L18w6ep XClQ==
MIME-Version: 1.0
X-Received: by 10.180.90.236 with SMTP id bz12mr21864773wib.33.1428935872442;  Mon, 13 Apr 2015 07:37:52 -0700 (PDT)
Sender: gettysjim@gmail.com
Received: by 10.194.118.166 with HTTP; Mon, 13 Apr 2015 07:37:52 -0700 (PDT)
Date: Mon, 13 Apr 2015 10:37:52 -0400
X-Google-Sender-Auth: oNdifEo6gcK7SlllOTrNN6QoQ_4
Message-ID: <CAGhGL2Cj44U94UL6W-9OeoyKTm4TE=JWVozCzAp-wr5eN9UETg@mail.gmail.com>
From: Jim Gettys <jg@freedesktop.org>
To: "aqm@ietf.org" <aqm@ietf.org>
Content-Type: multipart/alternative; boundary=f46d043be09c1577de05139c0fa1
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/HUmxuDzNxiBtqQ0aSiMrkeQDcGY>
Subject: [aqm] Review of draft-ietf-aqm-eval-guidelines
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 14:38:09 -0000

--f46d043be09c1577de05139c0fa1
Content-Type: text/plain; charset=UTF-8

Sorry this review is later than I would have liked: I came down with a cold
after the IETF and this review was half done. Hope this helps.

                                           - Jim


Required behavior, and running code
============================

My immediate, overall reaction is that this document prescribes a large set
of tests to be performed, using RFC 2119 MUST/SHOULD/MAY terminology.

However, I doubt if any AQM algorithm has been evaluated to all of those
tests; and it is unrealistic without "running code" for such tests that
anyone will actually run them; instead, this seems to be an "exercise to
the AQM implementer" of large magnitude, possibly so large as that no one
will actually perform even a fraction of them.

Will that mean that future AQM algorithms get stalled in working groups
because the tests are not done as prescribed? And how will we intercompare
tests between algorithms, with the test code not available?

And as there is little consequence that I can see to not performing the
tests, it seems unlikely that anyone will actually perform them.
I think if one cataloged all the tests requested by this document that the
number of tests is very large. I wonder if a small number of more
aggressive tests might not be easier.

It seems to me that if, as a community, we really believe in doing such
evaluations, we need a common set of tools for doing them, so that we can
actually easily inter compare results.

We really need *running code* for tests to evaluate AQM's/packet scheduling
algorithms. I fear that this will be an academic exercise without such a
shared code base for testing.

My reaction to this draft is that without such a set of tests, this
document is an academic exercise that will have little impact. I'd much
prefer to see a (smaller) concrete set of tests devised (probably an
elaboration of the data we get from Toke's netperf-wrapper tool).


Packet scheduling
==============

This draft is essentially bereft of any discussion of packet
scheduling/flow isolation, when, in many ways, it makes a bigger difference
than the AQM (mark/drop algorithm only, in the definition of this document).

Section 12 purports to discuss packet scheduling, but it is 5 sentences (<
1/2 page) total out of a document of >20 pages (of actual discussion, not
boiler plate)!

Somehow I suspect it is worth a bit more discussion ;-).

Yet packet scheduling/AQM do interact at some level; I don't think
that it can/should be handled in a separate draft.

For example, one of the wins of flow isolation is that TCP's ack clocking
starts behaving normally again, and ack compression is much less
problematic. One of the surprises of fq_codel was finding that we were able
to use the full bandwidth of the link in both directions simultaneously
(increasing net aggregate bandwidth on a link).

A second example: once you have flow isolation (e.g. fq_codel), it is still
important to understand if the AQM (mark/drop algorithm) is effective, as
bufferbloat can still kill single application performance if the mark/drop
algorithm is ineffective. Knowing what will happen to that that single
flow's behavior still matters.

A third example: ironically, TCP RTT unfairness gets worse when you control
the RTT with an effective AQM.  Flow queuing helps with this issue, which

Time variance
===========

The draft is silent on the topic of time variance behavior of AQM's.

Those of us facing the problems at the edge of the network are amazingly
aware of how rapidly available bandwidth may change, both because radio is
inherently a shared medium, and because of radio propagation.

Other media are also highly variable, if not as commonly as WiFi/cellular;
even switched Ethernet is very variable these days (vlan's have hidden the
actual bandwidth of a cable from view).
I talked to one VOIP provider who tracked their problems down to what vlans
were doing behind his back.

Cable and other systems are also shared/variable media (e.g. PowerBoost on
Comcast, or even day/night diurnal variation.

*The* most exciting day for me while CoDel was being developed was when
Kathy Nichols shared with me some data from a simulation showing CoDel's
behavior when faced with a dramatic (instantaneous, in this case) drop in
bandwidth (10x), and watching it recover to a new operating point.

It seems to me that some sort of test(s) for what happens when bandwidth
changes (in particular, high to low) are in order to evaluate algorithms.

Unfortunately, the last I knew some simulators were not very good at
simulating this sort of behavior (IIRC, Kathy was using ns2, and she had to
kludge something to get this behavior).

Most important (particularly once you have flow queuing), may be whether
the algorithm may become unstable/oscillate in the face of such changes or
not.  This is still an area I worry about for WiFi, where the RTT of many
paths is roughly comparable to the kind of variance in throughput seen.
Time will tell, I guess; but right now, we have no tools to explore this is
any sort of testing.


Specific comments
===============

Section 1. Introduction
------------------------------

Para 1.
It is a bit strange there is no reference to fq_codel at this date,
particularly since everyone involved in CoDel much prefers fq_codel and we
cannot anticipate anyone using CoDel in preference to fq_codel; it's in
Linux primarily to enable easy testing, rather than expecting anyone to
*actually* use it.

 Flow queuing is a huge win. And fq_codel is shipping/deployed on millions
of devices at this date.  There is an ID in the working group defining
fq_codel.... Having a concrete thing to talk about you that does packet
scheduling will help you when you get to redoing section 12.

Para 3:
The first sentence is misleading and untrue, as far as I can tell. SLA's
have had little to do with it, at least in the consumer edge where we hurt
the most today.

Bufferbloat has happened as 1) memory became really cheap, 2) Worshiping at
the altar of the religion that dropping a packet must be bad 2) to do a
single stream TCP performance test has meant that the minimum size of a
drop tail queue has been at least the BDP.

o the maximum bandwidth the device might ever go
o 100ms, which is roughly continental delay.

Net result is *minimum* buffering that is often 5-10 or more times sane
buffering.  Then people, thinking that more must be better, used any left
over memory. Note that this is already much too much for many applications
(music, VOIP; 100ms is *not* an acceptable amount of latency for a
connection to impose)....

Compounding this has been accounting in packets, rather than bytes or
time.  So people built operating systems to saturate a high speed link with
tiny UDP packets for RPC, without realizing that the buffering was then
another factor of 20 higher for TCP...

Then these numbers (for Ethernet) were copied into the WiFi drivers without
thought. Broadband head ends and modems aren't much better.

As trying to explain the history of how things went so badly wrong is not
helpful, I'd just strike the first sentence starting "In order to meet" and
just note that buffering has often been ten-100 or more times larger than
needed.

The Interplanetary record at the moment is held by GoGo Inflight, with 760
seconds of buffering measured (Mars is closer for part of its orbit).


Section 2. End to end metrics.
----------------------------------------

There is no discussion of any "Flow startup time".

For many/most "ants", the key time for good performance is the time to
*start* or *restart an idle* flow.

Not only are these operations such as DNS lookups, TCP opens, SSL
handshakes, dhcp handshakes, etc, but for web pages, the first packet or so
contains the size information required to layout a page, which is key to
interactive "speed", all needing any formal traffic classification.

The biggest single win of flow queuing is minimizing this lost time;
fq_codel goes even a step further by its "sparse flow" optimization, where
the first packet(s) of a new or idle flow are scheduled before flows that
have built a queue. So DNS lookups, TCP opens, the first data packet in an
image, SSL handshakes, all fly through the router.

For web surfing, the flow completion time is relatively irrelevant; what
the user cares about is how long until he can read the screen without it
reflowing.

Section 3.2 Buffer Size
------------------------------

The first sentence of this section is pretty useless.  What is the
bandwidth?  What is the delay?  later tests talk about testing at different
RTT's...  And the whole point of AQM's is to get away from static buffer
sizes....

Instead, when static buffer sizes are set, there needs to be justification
and explanation given, so that other implementers/testers can understand
how/when/if it should be changed for deployment or other testing.

These buffer default sizes are mentioned in the fq_codel draft since home
routers have limited RAM, so having some upper limit on buffer sizes is a
practical implementation limit (and one we'd be happy to do do away with
entirely; we just haven't gotten around to it as yet, and the draft
describes the current implementation, rather than mythology).

The buffer size constants on fq_codel are really on the "to fix someday"
list, rather than anything we otherwise care about in the algorithm.  With
more code complexity, we could remove both the # flows and buffer size
limits in fq_codel and make either/both fully dynamic.  So far, it hasn't
seemed worth the effort to make these automatic, particularly before we
understand more about the dynamic behavior in wireless networks, ***and
have better feedback/integration with the device driver layer so that we
might have insight into the information required to make these dynamic***.
Right now, too much information is hidden from the queue management layer
(at least in Linux).

Section 4.3 Unresponsive transport sender
---------------------------------------------------------

You might note that flow queuing systems help protect against (some) such
unresponsive senders.

Why are you specifying a New Reno flow?  One could argue, possibly
correctly, that Cubic is now the most common CC algorithm.  Some rational
is in order.  And Cubic is more likely to drive badly bloated connections
to the point of insanity.

Section 5.1 Motivation
-----------------------------

Some mention of flow queuing is in order...

5.2 Required Tests
--------------------------

It seems to me that there is a third test: a short RTT (maybe 20MS), which
is comparable to what you might find given a CDN is nearby.

Section 6. Burst absorption
-------------------------------------

There is a converse situation to burst absorption: that is momentary
bandwidth drops; these can occur both in VLAN situations, but also in
wireless systems, where you can easily have temporary drops in available
bandwidth.

Section 6.2: Required tests
-------------------------------------

Unfortunately, HTTP/1.1 Web traffic in the current world is much worse than
the proposed tests: typically it will be N essentially simultaneous
connections, each with an IW10; and the size of most objects are small, and
fit into that IW10 window.  Bursts of hundreds of packets are commonplace.

I cover some of this from Google data in my blog:

https://gettys.wordpress.com/2013/07/10/low-latency-requires-smart-queuing-traditional-aqm-is-not-enough/

So the web test is not specified decently: is it 100 packets back to back?
or 10 IW10 connections?

Also, flow completion, while interesting, is *not* the only
interesting/useful metric; getting the first packet(s) of each connection
allows page layout to occur without reflows and is equally
interesting/useful.

If you look in the above blog entry, you'll see a bar chart of visiting a
web page N times; one of the wonderful fq_codel results was the lack of
variance in the completion time, since the wrong packets weren't getting
badly delayed/dropped. (The web page completion time bar chart).

7.2.1 Definition of the congestion level
---------------------------------------------------

I am nervous about defining congestion in terms of loss rate; when I
measured loss rates in today's bloated edge network, the loss rates were
quite high (several percent) even with single flows (probably because Cubic
was being driven insane by the lack of losses on links with large amounts
of buffering.  So encouraging readers to believe that loss rates always
correlates with congestion level seems fraught with problems to me, as most
will jump to the incorrect conclusion that measuring one gives you the
other.

7.3 Parameter Sensitivity
---------------------------------

How would one show stability?  Is there a reference you can give to an
example of such stability analysis?

8.1 Traffic Mix
-------------------

The test is ill defined...  What is "6 webs" for a test?  Why 5 TCP flows?
are the flows bi-directional, or unidirectional?

8.2 Bi-directional traffic
------------------------------

It isn't clear that bi-directional traffic is actually required; the second
paragraph seems to contradict the third paragraph.

10. Operator control knobs...
---------------------------------------

Isn't really this a case of wanting to understand the degenerate behavior
of an algorithm, so that we understand the "worst case" behavior (and try
to see if an algorithm "first, does no harm", even in cases where said
algorithm should have been tuned (or autotuned itself)?

For example, the current # of hash buckets default in fq_codel (that
someday, we may make fully auto-tuning), if applied in a situation with
many flows so there are frequent hash collisions causes the algorithm to
lose its sparse flow optimization, but it is still going to apply CoDel to
any elephant flows and behave roughly as DRR would; it should not
"misbehave" and cause bad behavior: just not the better optimal behavior if
the # of hash buckets had been set to something reasonable.

This sort of analysis is worth including in an evaluation....

12. Interaction with scheduling
----------------------------------------

Somehow this section is currently pretty useless, particularly since good
scheduling can be of greater benefit that the details of mark/drop
algorithm.

I think we can/should do better here.

13.3 Comparing AQM schemes
------------------------------------------

I think there is a different issue left unmentioned at the moment, which is
the "applicability" of an algorithm to a given situation.

For example, setting the CoDel target in fq_codel below the media access
latency will hurt its behavior.  So twisting that knob is in fact necessary
under some circumstances, and a document should discuss when that may be
necessary.

Similarly the target of 5ms is pretty arbitrary and CoDel is in general not
very sensitive to what the target or interval are (except in this case of
setting an unacheivable target. The values were chosen after a lot of
simulation by Kathy Nichols. Helping people understand these cases is
important.

Section 13.4 Packet Sizes...
--------------------------------------

I think there is a simple rule that needs addition to this section: an
AQM/scheduling system "should guarantee progress", or particular flows can
fail entirely under some circumstances. This "liveness" constraint is a
property that all algorithms should provide.

--f46d043be09c1577de05139c0fa1
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_default" style=3D"font-size:small">Sor=
ry this review is later than I would have liked: I came down with a cold af=
ter the IETF and this review was half done. Hope this helps.</div><div clas=
s=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gmail=
_default" style=3D"font-size:small">=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0- Jim</div><div class=3D"gmail_def=
ault" style=3D"font-size:small"><br></div><div class=3D"gmail_default" styl=
e=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D"font-=
size:small">Required behavior, and running code</div><div class=3D"gmail_de=
fault" style=3D"font-size:small">=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</div><div class=3D"gmail_default=
" style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D=
"font-size:small">My immediate, overall reaction is that this document pres=
cribes a large set of tests to be performed, using RFC 2119 MUST/SHOULD/MAY=
 terminology.</div><div class=3D"gmail_default" style=3D"font-size:small"><=
br></div><div class=3D"gmail_default" style=3D"font-size:small">However, I =
doubt if any AQM algorithm has been evaluated to all of those tests; and it=
 is unrealistic without &quot;running code&quot; for such tests that anyone=
 will actually run them; instead, this seems to be an &quot;exercise to the=
 AQM implementer&quot; of large magnitude, possibly so large as that no one=
 will actually perform even a fraction of them.</div><div class=3D"gmail_de=
fault" style=3D"font-size:small"><br></div><div class=3D"gmail_default" sty=
le=3D"font-size:small">Will that mean that future AQM algorithms get stalle=
d in working groups because the tests are not done as prescribed? And how w=
ill we intercompare tests between algorithms, with the test code not availa=
ble?<br></div><div class=3D"gmail_default" style=3D"font-size:small"><br></=
div><div class=3D"gmail_default" style=3D"font-size:small"><div class=3D"gm=
ail_default">And as there is little consequence that I can see to not perfo=
rming the tests, it seems unlikely that anyone will actually perform them.<=
/div><div class=3D"gmail_default">I think if one cataloged all the tests re=
quested by this document that the number of tests is very large. I wonder i=
f a small number of more aggressive tests might not be easier.</div></div><=
div class=3D"gmail_default" style=3D"font-size:small"><br></div><div class=
=3D"gmail_default" style=3D"font-size:small">It seems to me that if, as a c=
ommunity, we really believe in doing such evaluations, we need a common set=
 of tools for doing them, so that we can actually easily inter compare resu=
lts.</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div>=
<div class=3D"gmail_default" style=3D"font-size:small">We really need *runn=
ing code* for tests to evaluate AQM&#39;s/packet scheduling algorithms. I f=
ear that this will be an academic exercise without such a shared code base =
for testing.=C2=A0</div><div class=3D"gmail_default" style=3D"font-size:sma=
ll"><br></div><div class=3D"gmail_default" style=3D"font-size:small">My rea=
ction to this draft is that without such a set of tests, this document is a=
n academic exercise that will have little impact. I&#39;d much prefer to se=
e a (smaller) concrete set of tests devised (probably an elaboration of the=
 data we get from Toke&#39;s netperf-wrapper tool).</div><div class=3D"gmai=
l_default" style=3D"font-size:small">=C2=A0</div><div class=3D"gmail_defaul=
t" style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=
=3D"font-size:small">Packet scheduling</div><div class=3D"gmail_default" st=
yle=3D"font-size:small">=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</div><di=
v class=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D=
"gmail_default" style=3D"font-size:small">This draft is essentially bereft =
of any discussion of packet scheduling/flow isolation, when, in many ways, =
it makes a bigger difference than the AQM (mark/drop algorithm only, in the=
 definition of this document).</div><div class=3D"gmail_default" style=3D"f=
ont-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:s=
mall">Section 12 purports to discuss packet scheduling, but it is 5 sentenc=
es (&lt; 1/2 page) total out of a document of &gt;20 pages (of actual discu=
ssion, not boiler plate)!</div><div class=3D"gmail_default" style=3D"font-s=
ize:small"><br></div><div class=3D"gmail_default" style=3D"font-size:small"=
>Somehow I suspect it is worth a bit more discussion ;-).</div><div class=
=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gmail_=
default" style=3D"font-size:small"><div class=3D"gmail_default">Yet packet =
scheduling/AQM do interact at some level; I don&#39;t think=C2=A0</div><div=
>that it can/should be handled in a separate draft.=C2=A0</div><div><br></d=
iv><div>For example, one of the wins of flow isolation is that TCP&#39;s ac=
k clocking starts behaving normally again, and ack compression is much less=
 problematic. One of the surprises of fq_codel was finding that we were abl=
e to use the full bandwidth of the link in both directions simultaneously (=
increasing net aggregate bandwidth on a link).</div></div><div class=3D"gma=
il_default" style=3D"font-size:small"><br></div><div class=3D"gmail_default=
" style=3D"font-size:small">A second example: once you have flow isolation =
(e.g. fq_codel), it is still important to understand if the AQM (mark/drop =
algorithm) is effective, as bufferbloat can still kill single application p=
erformance if the mark/drop algorithm is ineffective. Knowing what will hap=
pen to that that single flow&#39;s behavior still matters.</div><div class=
=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gmail_=
default" style=3D"font-size:small">A third example: ironically, TCP RTT unf=
airness gets worse when you control the RTT with an effective AQM.=C2=A0 Fl=
ow queuing helps with this issue, which=C2=A0</div><div class=3D"gmail_defa=
ult" style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=
=3D"font-size:small">Time variance</div><div class=3D"gmail_default" style=
=3D"font-size:small">=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</div><div class=3D"g=
mail_default" style=3D"font-size:small"><br></div><div class=3D"gmail_defau=
lt" style=3D"font-size:small">The draft is silent on the topic of time vari=
ance behavior of AQM&#39;s.</div><div class=3D"gmail_default" style=3D"font=
-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:smal=
l">Those of us facing the problems at the edge of the network are amazingly=
 aware of how rapidly available bandwidth may change, both because radio is=
 inherently a shared medium, and because of radio propagation.</div><div cl=
ass=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gma=
il_default" style=3D"font-size:small">Other media are also highly variable,=
 if not as commonly as WiFi/cellular; even switched Ethernet is very variab=
le these days (vlan&#39;s have hidden the actual bandwidth of a cable from =
view).=C2=A0</div><div class=3D"gmail_default" style=3D"font-size:small">I =
talked to one VOIP provider who tracked their problems down to what vlans w=
ere doing behind his back.</div><div class=3D"gmail_default" style=3D"font-=
size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:small=
">Cable and other systems are also shared/variable media (e.g. PowerBoost o=
n Comcast, or even day/night diurnal variation.</div><div class=3D"gmail_de=
fault" style=3D"font-size:small"><br></div><div class=3D"gmail_default" sty=
le=3D"font-size:small">*The* most exciting day for me while CoDel was being=
 developed was when Kathy Nichols shared with me some data from a simulatio=
n showing CoDel&#39;s behavior when faced with a dramatic (instantaneous, i=
n this case) drop in bandwidth (10x), and watching it recover to a new oper=
ating point.</div><div class=3D"gmail_default" style=3D"font-size:small"><b=
r></div><div class=3D"gmail_default" style=3D"font-size:small">It seems to =
me that some sort of test(s) for what happens when bandwidth changes (in pa=
rticular, high to low) are in order to evaluate algorithms.</div><div class=
=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gmail_=
default" style=3D"font-size:small">Unfortunately, the last I knew some simu=
lators were not very good at simulating this sort of behavior (IIRC, Kathy =
was using ns2, and she had to kludge something to get this behavior).</div>=
<div class=3D"gmail_default" style=3D"font-size:small"><br></div><div class=
=3D"gmail_default" style=3D"font-size:small">Most important (particularly o=
nce you have flow queuing), may be whether the algorithm may become unstabl=
e/oscillate in the face of such changes or not.=C2=A0 This is still an area=
 I worry about for WiFi, where the RTT of many paths is roughly comparable =
to the kind of variance in throughput seen.=C2=A0 Time will tell, I guess; =
but right now, we have no tools to explore this is any sort of testing.</di=
v><div class=3D"gmail_default" style=3D"font-size:small"><br></div><div cla=
ss=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gmai=
l_default" style=3D"font-size:small">Specific comments</div><div class=3D"g=
mail_default" style=3D"font-size:small">=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D</div><div class=3D"gmail_default" style=3D"font-size:small"><b=
r></div><div class=3D"gmail_default" style=3D"font-size:small">Section 1. I=
ntroduction</div><div class=3D"gmail_default" style=3D"font-size:small">---=
---------------------------</div><div class=3D"gmail_default" style=3D"font=
-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:smal=
l">Para 1.</div><div class=3D"gmail_default" style=3D"font-size:small">It i=
s a bit strange there is no reference to fq_codel at this date, particularl=
y since everyone involved in CoDel much prefers fq_codel and we cannot anti=
cipate anyone using CoDel in preference to fq_codel; it&#39;s in Linux prim=
arily to enable easy testing, rather than expecting anyone to *actually* us=
e it.</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div=
><div class=3D"gmail_default" style=3D"font-size:small">=C2=A0Flow queuing =
is a huge win. And fq_codel is shipping/deployed on millions of devices at =
this date.=C2=A0 There is an ID in the working group defining fq_codel.... =
Having a concrete thing to talk about you that does packet scheduling will =
help you when you get to redoing section 12.</div><div class=3D"gmail_defau=
lt" style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=
=3D"font-size:small">Para 3:</div><div class=3D"gmail_default" style=3D"fon=
t-size:small">The first sentence is misleading and untrue, as far as I can =
tell. SLA&#39;s have had little to do with it, at least in the consumer edg=
e where we hurt the most today.</div><div class=3D"gmail_default" style=3D"=
font-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:=
small">Bufferbloat has happened as 1) memory became really cheap, 2) Worshi=
ping at the altar of the religion that dropping a packet must be bad 2) to =
do a single stream TCP performance test has meant that the minimum size of =
a drop tail queue has been at least the BDP.</div><div class=3D"gmail_defau=
lt" style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=
=3D"font-size:small">o the maximum bandwidth the device might ever go</div>=
<div class=3D"gmail_default" style=3D"font-size:small">o 100ms, which is ro=
ughly continental delay.</div><div class=3D"gmail_default" style=3D"font-si=
ze:small"><br></div><div class=3D"gmail_default" style=3D"font-size:small">=
Net result is *minimum* buffering that is often 5-10 or more times sane buf=
fering.=C2=A0 Then people, thinking that more must be better, used any left=
 over memory. Note that this is already much too much for many applications=
 (music, VOIP; 100ms is *not* an acceptable amount of latency for a connect=
ion to impose)....</div><div class=3D"gmail_default" style=3D"font-size:sma=
ll"><br></div><div class=3D"gmail_default" style=3D"font-size:small">Compou=
nding this has been accounting in packets, rather than bytes or time.=C2=A0=
 So people built operating systems to saturate a high speed link with tiny =
UDP packets for RPC, without realizing that the buffering was then another =
factor of 20 higher for TCP...</div><div class=3D"gmail_default" style=3D"f=
ont-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:s=
mall">Then these numbers (for Ethernet) were copied into the WiFi drivers w=
ithout thought. Broadband head ends and modems aren&#39;t much better. =C2=
=A0 =C2=A0 =C2=A0 =C2=A0=C2=A0</div><div class=3D"gmail_default" style=3D"f=
ont-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:s=
mall">As trying to explain the history of how things went so badly wrong is=
 not helpful, I&#39;d just strike the first sentence starting &quot;In orde=
r to meet&quot; and just note that buffering has often been ten-100 or more=
 times larger than needed.</div><div class=3D"gmail_default" style=3D"font-=
size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:small=
">The Interplanetary record at the moment is held by GoGo Inflight, with 76=
0 seconds of buffering measured (Mars is closer for part of its orbit).</di=
v><div class=3D"gmail_default" style=3D"font-size:small"><br></div><div cla=
ss=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D"gmai=
l_default" style=3D"font-size:small">Section 2. End to end metrics.</div><d=
iv class=3D"gmail_default" style=3D"font-size:small">----------------------=
------------------</div><div class=3D"gmail_default" style=3D"font-size:sma=
ll"><br></div><div class=3D"gmail_default" style=3D"font-size:small">There =
is no discussion of any &quot;Flow startup time&quot;.</div><div class=3D"g=
mail_default" style=3D"font-size:small"><br></div><div class=3D"gmail_defau=
lt" style=3D"font-size:small">For many/most &quot;ants&quot;, the key time =
for good performance is the time to *start* or *restart an idle* flow.</div=
><div class=3D"gmail_default" style=3D"font-size:small"><br></div><div clas=
s=3D"gmail_default" style=3D"font-size:small">Not only are these operations=
 such as DNS lookups, TCP opens, SSL handshakes, dhcp handshakes, etc, but =
for web pages, the first packet or so contains the size information require=
d to layout a page, which is key to interactive &quot;speed&quot;, all need=
ing any formal traffic classification.</div><div class=3D"gmail_default" st=
yle=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D"fon=
t-size:small">The biggest single win of flow queuing is minimizing this los=
t time; fq_codel goes even a step further by its &quot;sparse flow&quot; op=
timization, where the first packet(s) of a new or idle flow are scheduled b=
efore flows that have built a queue. So DNS lookups, TCP opens, the first d=
ata packet in an image, SSL handshakes, all fly through the router.</div><d=
iv class=3D"gmail_default" style=3D"font-size:small"><br></div><div class=
=3D"gmail_default" style=3D"font-size:small">For web surfing, the flow comp=
letion time is relatively irrelevant; what the user cares about is how long=
 until he can read the screen without it reflowing.</div><div class=3D"gmai=
l_default" style=3D"font-size:small"><br></div><div class=3D"gmail_default"=
 style=3D"font-size:small">Section 3.2 Buffer Size</div><div class=3D"gmail=
_default" style=3D"font-size:small">------------------------------</div><di=
v class=3D"gmail_default" style=3D"font-size:small"><br></div><div class=3D=
"gmail_default" style=3D"font-size:small">The first sentence of this sectio=
n is pretty useless.=C2=A0 What is the bandwidth?=C2=A0 What is the delay? =
=C2=A0later tests talk about testing at different RTT&#39;s...=C2=A0 And th=
e whole point of AQM&#39;s is to get away from static buffer sizes....</div=
><div class=3D"gmail_default" style=3D"font-size:small"><br></div><div clas=
s=3D"gmail_default" style=3D"font-size:small">Instead, when static buffer s=
izes are set, there needs to be justification and explanation given, so tha=
t other implementers/testers can understand how/when/if it should be change=
d for deployment or other testing. =C2=A0</div><div class=3D"gmail_default"=
 style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D"=
font-size:small">These buffer default sizes are mentioned in the fq_codel d=
raft since home routers have limited RAM, so having some upper limit on buf=
fer sizes is a practical implementation limit (and one we&#39;d be happy to=
 do do away with entirely; we just haven&#39;t gotten around to it as yet, =
and the draft describes the current implementation, rather than mythology).=
=C2=A0</div><div class=3D"gmail_default" style=3D"font-size:small"><br></di=
v><div class=3D"gmail_default" style=3D"font-size:small">The buffer size co=
nstants on fq_codel are really on the &quot;to fix someday&quot; list, rath=
er than anything we otherwise care about in the algorithm.=C2=A0 With more =
code complexity, we could remove both the # flows and buffer size limits in=
 fq_codel and make either/both fully dynamic.=C2=A0 So far, it hasn&#39;t s=
eemed worth the effort to make these automatic, particularly before we unde=
rstand more about the dynamic behavior in wireless networks, ***and have be=
tter feedback/integration with the device driver layer so that we might hav=
e insight into the information required to make these dynamic***.=C2=A0 Rig=
ht now, too much information is hidden from the queue management layer (at =
least in Linux).</div><div class=3D"gmail_default" style=3D"font-size:small=
"><br></div><div class=3D"gmail_default" style=3D"font-size:small">Section =
4.3 Unresponsive transport sender</div><div class=3D"gmail_default" style=
=3D"font-size:small">------------------------------------------------------=
---</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div><=
div class=3D"gmail_default" style=3D"font-size:small">You might note that f=
low queuing systems help protect against (some) such unresponsive senders.<=
/div><div class=3D"gmail_default" style=3D"font-size:small"><br></div><div =
class=3D"gmail_default" style=3D"font-size:small">Why are you specifying a =
New Reno flow?=C2=A0 One could argue, possibly correctly, that Cubic is now=
 the most common CC algorithm.=C2=A0 Some rational is in order.=C2=A0 And C=
ubic is more likely to drive badly bloated connections to the point of insa=
nity.</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div=
><div class=3D"gmail_default" style=3D"font-size:small">Section 5.1 Motivat=
ion</div><div class=3D"gmail_default" style=3D"font-size:small">-----------=
------------------</div><div class=3D"gmail_default" style=3D"font-size:sma=
ll"><br></div><div class=3D"gmail_default" style=3D"font-size:small">Some m=
ention of flow queuing is in order...</div><div class=3D"gmail_default" sty=
le=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D"font=
-size:small">5.2 Required Tests</div><div class=3D"gmail_default" style=3D"=
font-size:small">--------------------------</div><div class=3D"gmail_defaul=
t" style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=
=3D"font-size:small">It seems to me that there is a third test: a short RTT=
 (maybe 20MS), which is comparable to what you might find given a CDN is ne=
arby.</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div=
><div class=3D"gmail_default" style=3D"font-size:small">Section 6. Burst ab=
sorption</div><div class=3D"gmail_default" style=3D"font-size:small">------=
-------------------------------</div><div class=3D"gmail_default" style=3D"=
font-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:=
small">There is a converse situation to burst absorption: that is momentary=
 bandwidth drops; these can occur both in VLAN situations, but also in wire=
less systems, where you can easily have temporary drops in available bandwi=
dth.</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div>=
<div class=3D"gmail_default" style=3D"font-size:small">Section 6.2: Require=
d tests</div><div class=3D"gmail_default" style=3D"font-size:small">-------=
------------------------------</div><div class=3D"gmail_default" style=3D"f=
ont-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:s=
mall">Unfortunately, HTTP/1.1 Web traffic in the current world is much wors=
e than the proposed tests: typically it will be N essentially simultaneous =
connections, each with an IW10; and the size of most objects are small, and=
 fit into that IW10 window.=C2=A0 Bursts of hundreds of packets are commonp=
lace.</div><div class=3D"gmail_default" style=3D"font-size:small"><br></div=
><div class=3D"gmail_default" style=3D"font-size:small">I cover some of thi=
s from Google data in my blog:</div><div class=3D"gmail_default" style=3D"f=
ont-size:small"><br></div><div class=3D"gmail_default" style><a href=3D"htt=
ps://gettys.wordpress.com/2013/07/10/low-latency-requires-smart-queuing-tra=
ditional-aqm-is-not-enough/">https://gettys.wordpress.com/2013/07/10/low-la=
tency-requires-smart-queuing-traditional-aqm-is-not-enough/</a><br></div><d=
iv class=3D"gmail_default" style><br></div><div class=3D"gmail_default" sty=
le>So the web test is not specified decently: is it 100 packets back to bac=
k? or 10 IW10 connections?</div><div class=3D"gmail_default" style><br></di=
v><div class=3D"gmail_default" style>Also, flow completion, while interesti=
ng, is *not* the only interesting/useful metric; getting the first packet(s=
) of each connection allows page layout to occur without reflows and is equ=
ally interesting/useful.</div><div class=3D"gmail_default" style><br></div>=
<div class=3D"gmail_default" style>If you look in the above blog entry, you=
&#39;ll see a bar chart of visiting a web page N times; one of the wonderfu=
l fq_codel results was the lack of variance in the completion time, since t=
he wrong packets weren&#39;t getting badly delayed/dropped. (The web page c=
ompletion time bar chart).</div><div class=3D"gmail_default" style><br></di=
v><div class=3D"gmail_default" style>7.2.1 Definition of the congestion lev=
el</div><div class=3D"gmail_default" style>--------------------------------=
-------------------</div><div class=3D"gmail_default" style><br></div><div =
class=3D"gmail_default" style>I am nervous about defining congestion in ter=
ms of loss rate; when I measured loss rates in today&#39;s bloated edge net=
work, the loss rates were quite high (several percent) even with single flo=
ws (probably because Cubic was being driven insane by the lack of losses on=
 links with large amounts of buffering.=C2=A0 So encouraging readers to bel=
ieve that loss rates always correlates with congestion level seems fraught =
with problems to me, as most will jump to the incorrect conclusion that mea=
suring one gives you the other.</div><div class=3D"gmail_default" style><br=
></div><div class=3D"gmail_default" style>7.3 Parameter Sensitivity</div><d=
iv class=3D"gmail_default" style>---------------------------------</div><di=
v class=3D"gmail_default" style><br></div><div class=3D"gmail_default" styl=
e>How would one show stability?=C2=A0 Is there a reference you can give to =
an example of such stability analysis?</div><div class=3D"gmail_default" st=
yle><br></div><div class=3D"gmail_default" style>8.1 Traffic Mix</div><div =
class=3D"gmail_default" style>-------------------</div><div class=3D"gmail_=
default" style><br></div><div class=3D"gmail_default" style>The test is ill=
 defined...=C2=A0 What is &quot;6 webs&quot; for a test?=C2=A0 Why 5 TCP fl=
ows? are the flows bi-directional, or unidirectional?</div><div class=3D"gm=
ail_default" style><br></div><div class=3D"gmail_default" style>8.2 Bi-dire=
ctional traffic</div><div class=3D"gmail_default" style>-------------------=
-----------</div><div class=3D"gmail_default" style><br></div><div class=3D=
"gmail_default" style>It isn&#39;t clear that bi-directional traffic is act=
ually required; the second paragraph seems to contradict the third paragrap=
h.</div><div class=3D"gmail_default" style><br></div><div class=3D"gmail_de=
fault" style>10. Operator control knobs...</div><div class=3D"gmail_default=
" style>---------------------------------------</div><div class=3D"gmail_de=
fault" style><br></div><div class=3D"gmail_default" style>Isn&#39;t really =
this a case of wanting to understand the degenerate behavior of an algorith=
m, so that we understand the &quot;worst case&quot; behavior (and try to se=
e if an algorithm &quot;first, does no harm&quot;, even in cases where said=
 algorithm should have been tuned (or autotuned itself)? =C2=A0</div><div c=
lass=3D"gmail_default" style><br></div><div class=3D"gmail_default" style>F=
or example, the current # of hash buckets default in fq_codel (that someday=
, we may make fully auto-tuning), if applied in a situation with many flows=
 so there are frequent hash collisions causes the algorithm to lose its spa=
rse flow optimization, but it is still going to apply CoDel to any elephant=
 flows and behave roughly as DRR would; it should not &quot;misbehave&quot;=
 and cause bad behavior: just not the better optimal behavior if the # of h=
ash buckets had been set to something reasonable.</div><div class=3D"gmail_=
default" style><br></div><div class=3D"gmail_default" style>This sort of an=
alysis is worth including in an evaluation....</div><div class=3D"gmail_def=
ault" style><br></div><div class=3D"gmail_default" style>12. Interaction wi=
th scheduling</div><div class=3D"gmail_default" style>---------------------=
-------------------</div><div class=3D"gmail_default" style><br></div><div =
class=3D"gmail_default" style>Somehow this section is currently pretty usel=
ess, particularly since good scheduling can be of greater benefit that the =
details of mark/drop algorithm.</div><div class=3D"gmail_default" style><br=
></div><div class=3D"gmail_default" style>I think we can/should do better h=
ere.</div><div class=3D"gmail_default" style><br></div><div class=3D"gmail_=
default" style>13.3 Comparing AQM schemes</div><div class=3D"gmail_default"=
 style>------------------------------------------</div><div class=3D"gmail_=
default" style><br></div><div class=3D"gmail_default" style>I think there i=
s a different issue left unmentioned at the moment, which is the &quot;appl=
icability&quot; of an algorithm to a given situation.</div><div class=3D"gm=
ail_default" style><br></div><div class=3D"gmail_default" style>For example=
, setting the CoDel target in fq_codel below the media access latency will =
hurt its behavior.=C2=A0 So twisting that knob is in fact necessary under s=
ome circumstances, and a document should discuss when that may be necessary=
.=C2=A0</div><div class=3D"gmail_default" style><br></div><div class=3D"gma=
il_default" style>Similarly the target of 5ms is pretty arbitrary and CoDel=
 is in general not very sensitive to what the target or interval are (excep=
t in this case of setting an unacheivable target. The values were chosen af=
ter a lot of simulation by Kathy Nichols. Helping people understand these c=
ases is important.</div><div class=3D"gmail_default" style><br></div><div c=
lass=3D"gmail_default" style>Section 13.4 Packet Sizes...</div><div class=
=3D"gmail_default" style>--------------------------------------</div><div c=
lass=3D"gmail_default" style><br></div><div class=3D"gmail_default" style>I=
 think there is a simple rule that needs addition to this section: an AQM/s=
cheduling system &quot;should guarantee progress&quot;, or particular flows=
 can fail entirely under some circumstances. This &quot;liveness&quot; cons=
traint is a property that all algorithms should provide.</div><div class=3D=
"gmail_default" style><br></div><div class=3D"gmail_default" style><br></di=
v><div class=3D"gmail_default" style><br></div><div class=3D"gmail_default"=
 style><br></div><div class=3D"gmail_default" style><br></div><div class=3D=
"gmail_default" style><br></div><div class=3D"gmail_default" style><br></di=
v><div class=3D"gmail_default" style><br></div><div class=3D"gmail_default"=
 style=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D"=
font-size:small"><br></div><div class=3D"gmail_default" style=3D"font-size:=
small"><br></div><div class=3D"gmail_default" style=3D"font-size:small"><br=
></div></div>

--f46d043be09c1577de05139c0fa1--


From nobody Mon Apr 13 08:11:35 2015
Return-Path: <bob.briscoe@bt.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 331391ACCF4 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 08:11:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.089
X-Spam-Level: 
X-Spam-Status: No, score=0.089 tagged_above=-999 required=5 tests=[BAYES_50=0.8, RCVD_IN_DNSWL_LOW=-0.7, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id MeXCb9PXCRld for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 08:11:27 -0700 (PDT)
Received: from hubrelay-rd.bt.com (hubrelay-rd.bt.com [62.239.224.98]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 01FE71AC43A for <aqm@ietf.org>; Mon, 13 Apr 2015 08:11:24 -0700 (PDT)
Received: from EVMHR71-UKRD.domain1.systemhost.net (10.36.3.109) by EVMHR66-UKRD.bt.com (10.187.101.21) with Microsoft SMTP Server (TLS) id 14.3.195.1; Mon, 13 Apr 2015 16:11:20 +0100
Received: from EPHR02-UKIP.domain1.systemhost.net (147.149.100.81) by EVMHR71-UKRD.domain1.systemhost.net (10.36.3.109) with Microsoft SMTP Server (TLS) id 8.3.348.2; Mon, 13 Apr 2015 16:11:22 +0100
Received: from bagheera.jungle.bt.co.uk (132.146.168.158) by EPHR02-UKIP.domain1.systemhost.net (147.149.100.81) with Microsoft SMTP Server id 14.3.181.6; Mon, 13 Apr 2015 16:11:21 +0100
Received: from BTP075694.jungle.bt.co.uk ([10.109.34.80])	by bagheera.jungle.bt.co.uk (8.13.5/8.12.8) with ESMTP id t3DFBG3R002270;	Mon, 13 Apr 2015 16:11:17 +0100
Message-ID: <201504131511.t3DFBG3R002270@bagheera.jungle.bt.co.uk>
X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9
Date: Mon, 13 Apr 2015 16:11:15 +0100
To: David Lang <david@lang.hm>
From: Bob Briscoe <bob.briscoe@bt.com>
In-Reply-To: <alpine.DEB.2.02.1503291923300.26044@nftneq.ynat.uz>
References: <23AFEFE3-4D93-4DD9-A22B-952C63DB9FE3@cisco.com> <BF6B00CC65FD2D45A326E74492B2C19FB75BAA82@FR711WXCHMBA05.zeu.alcatel-lucent.com> <72EE366B-05E6-454C-9E53-5054E6F9E3E3@ifi.uio.no> <55146DB9.7050501@rogers.com> <08C34E4A-DFB7-4816-92AE-2ED161799488@ifi.uio.no> <BF6B00CC65FD2D45A326E74492B2C19FB75BAFA0@FR711WXCHMBA05.zeu.alcatel-lucent.com> <alpine.DEB.2.02.1503271024550.2416@nftneq.ynat.uz> <5d58d2e21400449280173aa63069bf7a@hioexcmbx05-prd.hq.netapp.com> <20150327183659.GI39886@verdi> <72C12F6B-9DDE-4483-81F2-2D9A0F2D3A48@cs.columbia.edu> <alpine.DEB.2.02.1503271211200.19390@nftneq.ynat.uz> <D13AFCE7.46BC%kk@cs.ucr.edu> <alpine.DEB.2.02.1503271257230.19390@nftneq.ynat.uz> <AE342093-DE05-4D93-96DA-EB07E221F1D9@netapp.com> <alpine.DEB.2.02.1503291923300.26044@nftneq.ynat.uz>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-Scanned-By: MIMEDefang 2.56 on 132.146.168.158
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/kcP-lH7_QyrzwFwdH9mrBZNR9Is>
Cc: "Scheffenegger, Richard" <rs@netapp.com>, Vishal Misra <misra@cs.columbia.edu>, KK <kk@cs.ucr.edu>, John Leslie <john@jlc.net>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] think once to mark, think twice to drop: draft-ietf-aqm-ecn-benefits-02
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 15:11:31 -0000

David,

Returning from a fortnight offlist...

I think your conception of how ECN works is incorrect. You describe 
ECN as if the AQM marks one packet when it drops another packet. You 
say that the ECN-mark speeds up the retransmission of the dropped 
packet. On the contrary, the idea of classic ECN [RFC3168] is that 
the ECN marks replace the drops. In all known testing (except 
pathological cases), classic ECN effectively eliminates drops for all 
ECN-capable packets.

Nonetheless, I do agree with your sentiment that the perfect is the 
enemy of the good. We can remove most of the really bloat-induced 
latency without ECN. So the message must be clear: Deploy AQM now. No 
need to wait for ECN. But implementations SHOULD allow ECN packets to 
be classifed into a separately configurable instance of the AQM algo. {Note 1}

Similarly, this WG made sure we did not deprecate RED in the AQM 
recommendations. Because, in existing equipment, even a poorly tuned 
RED is usually much better than a bloated buffer with no AQM.

Just as the WG mustn't confuse messages, you mustn't get confused by 
a discussion about the potential for more ambitious reductions in 
latency. Koen started the thread with reference to our presentation 
in the ICCRG in the IRTF, where 'R' in both cases stands for Research.

And I believe it is valid for the ECN benefits draft (in the IETF AQM 
WG) to point to the potential of ECN, by using insights from research 
in progress.

There are a number of different contributions to unnecessary latency, 
not just the one popularised as bufferbloat:

[In the following, I will use the term 'short message(s)' as 
shorthand for either a short interactive flow or a long flow 
consisting of short interactive messages (like in a game), or video 
frames or voice datagrams or anything where the perceived latency 
depends on the latency of each 'message', especially a string of 
messages with serial dependency as is typical in Web.]

#1 The well-known bufferbloat problem, where a long-running flow 
fills a bloated buffer, delays short messages.
   - AQM and/or flow queuing can remove this delay, without needing ECN.

#2 A loss causes head-of-line blocking for short message(s) while 
waiting for the retransmission.
   - AQM without ECN cannot remove this delay.
   - Flow queuing cannot remove this delay, if the losses are self-induced.
   - FEC can remove this delay without needing ECN, but the increased 
redundancy is equivalent to poorer utilisation (altho only the short 
flows need the redundancy).
   - ECN can remove this delay.

#3 A loss near the end of a short flow can lead to multi-RTT delay.
   - AQM without ECN cannot remove this delay.
   - Techniques like tail-loss probe and RTO-restart can mitigate 
this delay, without ECN, but not remove it.
   - ECN can remove this delay.

#4 The Reno/Cubic sawtooth causes variation in delay between 1 and 2 
base RTTs. This can affect short messages.
   - AQM without ECN cannot remove this delay unless configured to 
sacrifice utilisation.
   - Flow queuing removes this delay variation if caused by a 
separate long-running flow.{Note 2}
   - A change to the TCP algo (e.g. DCTCP) can remove this delay 
variation. Smaller sawteeth imply a much higher signalling rate, 
which in turn requires ECN, otherwise drop probability would be 
excessive. (This was the main point Koen was making.)
   - Therefore, ECN can remove this delay.

#5 Slow-starts{Note 3} cause spikes in delay.
   - AQM without ECN cannot remove this delay, and typically AQM is 
designed to allow such bursts of delay in the hope they will 
disappear of their own accord.
   - Flow queuing can remove the effect of these delay bursts on 
other flows, but only if it gives all flows a separate queue from the 
start.{Note 2}
   - Delay-based softening of slow-start, such as Hybrid Slow-Start 
in Linux, can mitigate these variations, but with increased risk of 
coming out of SS early, causing significantly longer completion time.
   - ECN with AQM based on the instantaneous queue limits this delay, 
without the risk of longer completion time.

#6 Slow-starts{Note 3} can cause runs of losses, which in turn cause delays.
   - AQM without ECN cannot remove these delays.
   - Flow queuing cannot remove these losses, if self-induced.
   - Delay-based SS like HSS can mitigate these losses, with 
increased risk of longer completion time.
   - ECN can remove these losses, and the consequent delays.

Summary:
* AQM alone solves the main problem
* Flow queuing solves or mitigates most of the remaining secondary problems.
* ECN has the potential to solve all the remaining secondary problems 
(pending further research to prove some of them).

Whether flow queuing is applicable depends on the scale. The work I'm 
doing with Koen is to reduce the cost of the queuing mechanisms on 
our BNGs (broadband network gateways). We're trying to reduce the 
cost of per-customer queuing at scale, so per-flow queuing is simply 
out of the question. Whereas ECN requires no more processing than drop.

ECN has potential cheating problems, but we have per-customer queues anyway.
Using flow as the unit of allocation also has its own problems, with 
no proposed solutions.


Bob

{Note 1}: Your general point is that the perfect can be the enemy of 
the good. Here's the presentation I gave in TSVAREA straight after VJ 
presented CoDel in 2012 entitled "DCTCP & CoDel; the Best is the 
Friend of the Good."
<http://www.bobbriscoe.net/presents/1207ietf/1207-tsvarea-dctcp.pdf>

{Note 2}:  A lone flow can cause this delay variation to itself, but 
that's irrelevant because, if the delay were not in the network it 
would be at the sender.

{Note 3}: Delay and loss spikes can equivalently be caused when 
Cubic's window rises to seek out newly available capacity after 
another flow finishes or the link rate varies.

At 05:16 30/03/2015, David Lang wrote:
>On Sat, 28 Mar 2015, Scheffenegger, Richard wrote:
>
>>David,
>>
>>Perhaps you would care to provide some text to address the 
>>misconception that you pointed out? (To wait for a 100% fix as a 
>>90% fix appears much less appealing, while the current state of art is at 0%)
>
>Ok, you put me on the spot :-) Here goes.
>
>>If you think that aqm-recommendations is not strogly enough worded. 
>>I think this particular discussion (to aqm or not) really belongs 
>>there. The other document (ecn benefits) has a different target in 
>>arguing for going those last 10%...
>
>so here is my "elevator pitch" on the problem. Feel free to take 
>anything I say here for any purpose, and I'm sure I'll get corrected 
>for anything I am wong on
>
>Problem statement: Transmit buffers are needed to keep the network 
>layer fully utilized, but excessive buffers result in poor latency 
>for all traffic. This latency is frequently bad enough to cause some 
>types of traffic to fail entirely.
>
>   <link to more background goes here, including how separate 
> benchmarks for throughput and latency have mislead people, "packet 
> loss considered evil", cheaper memory encouraging larger buffers, 
> etc. Include tests like netperf-wrapper and ping latency while 
> under load, etc. Include examples where buffers have resulted in 
> latencies so long that packets are retransmitted before the first 
> copy gets to the destination>
>
>Traditionally, transmit buffers have been sized to handle a fixed 
>number of packets. Due to teh variation in packet sizes, it is 
>impossible to tune this value to both keep the link fully utilized 
>when small packets dominate the trafific without having the queue 
>size be large enough to cause latency problems when large packets 
>dominate the traffic.
>
>Shifting to Byte Queue Lengths where queues are allowed to hold a 
>variable number of packets depending on how large they are makes it 
>possible to manually tune the transmit buffer size to get good 
>latency under all traffic conditions at a given speed. However, this 
>step forward revealed two additional problems.
>
>1. whenever the data rate changes, this value needs to be manually 
>changed (multi-link paths loose a link, noise degrades max 
>throughput on a link, etc)
>
>2. high volume flows (i.e. bulk downloads) can starve other flows 
>(DNS lookups, VoIP, Gaming, etc). this happens because space in tue 
>queue is on a first-com-first-served basis, so the high-volume 
>traffic fills the queue (at which point it starts to be dropped), 
>but all other traffic that tries to arrive is also dropped. It turns 
>out that these light flows tend to have a larger effect on the user 
>experience than heavier flows, because things tend to be serialized 
>behind the lighter flows (DNS lookup before doing a large download, 
>retrieving a small HTML page to find what additional resources need 
>to be fetched to display a page), or the user experience is directly 
>effected by light flows (gaming lag, VoIP drops, etc)
>
>Active Queue Management addresses these problems by adapting the 
>amount of data that is buffered to match the data transmission 
>capacity, and prevents high volume flows from starving low-volume 
>flows without the need to implement QoS classifications.
>
>   <insert link about how you can't trust QoS tags that are made by 
> other organizations, ways that it can be abused, etc>
>
>This is possible because AQM algoithms don't have to drop the new 
>packet that arrives, the algorithm can decide to drop the packet for 
>one of the heavy flows rather than for one of the lightweight flows.
>
><insert references to currently favored AQM options here, PIE, 
>fq_codel, cake, ???. Also links to failed approaches>
>
>Turning on aqm on every bottleneck link makes the Internet usable 
>for everyone, no matter what sort of application they are using.
>
><insert link on how to deal with equipment you can't configure by 
>throttling bandwidth before the bottleneck oand/or doing ingress 
>shaping of traffic>
>
>
>While AQM makes the network usable, there is still additional room 
>for improvement. While dropping packets does result in the TCP 
>senders slowing down,and eventually stabilizing at around the right 
>speed to keep the link fully utilized, the only way that senders 
>have been able to detect problems is to discover that they have not 
>received an ack for the traffic within the allowed time. This causes 
>a 'bubble' in the flow as teh dropped packet must be retransmitted 
>(and sometimes a significant amount of data after the dropped packet 
>that did make it to the destination, but could not be acked because 
>fo the missing packet).
>
>This "bubble" in the data flow can be greatly compressed by 
>configuring the AQM algorithm to send an ECN packet to the sender 
>when it drops a packet in a flow. The sender can then adapt faster, 
>slowing down it's new data, and re-sending the dropped packet 
>without having to wait for the timeout. This has two major effects 
>by allowing the sender to retransmit the packet sooner the dealy on 
>the dropped data is not as long, and because the replacement data 
>can arrive before the timeout of the following packets, they may not 
>need to be re-sent. by configuring the AQM algorithm to send the ECN 
>notification to the sender only when the packet is being dropped, 
>the effect of failure of the ECN packet to get through to the sender 
>(the notification packet runs into congestion and gets dropped, some 
>network device blocks it, etc) is that the ECN enabled case devolves 
>to match the non-ECN case in that the sender will still detect the 
>dropped packet via the timeout waiting for the ack as if ENCN was not enabled.
>
><insert link to possible problems that can happen here, including 
>the potential for an app to 'game' things if packets are marked at a 
>different level than when they are dropped.>
>
>
>So a very strong recommendation to enable Active Queue Management, 
>while the different algorithms have different advantages and levels 
>of testing, even the 'worst' of the set results in a night-and-day 
>improvement for usability compared to unmanaged buffers.
>
>Enabling ECN at the same point as dropping packets as part of 
>enabling any AQM algorithm results in a noticable improvement over 
>the base algorithm without ECN. When compared to the baseline, the 
>improvement added by ECN is tiny compared to the improvement from enabling AQM.
>
>   Is it fair to say that plain aqm vs aqm+ecn variation is on the 
> same order of difference as the differences between the different 
> AQM algorithms?
>
>
>
>Future research items (which others here may already have done, and 
>would not be part of my 'elevator pitch')
>
>I believe that currently ECn triggers the exact same slowdown that a 
>missed packet does, and it may be appropriate to have the sender do 
>a less drastic slowdown.
>
>It would be very interesing to provide soem way for the application 
>sending the traffic to detect dropped packets and ECN responses. For 
>example, a streaming media source (especially an interactive one 
>like video conferencing) could adjust the bitrate that it's sending.
>
>David Lang
>
>_______________________________________________
>aqm mailing list
>aqm@ietf.org
>https://www.ietf.org/mailman/listinfo/aqm

________________________________________________________________
Bob Briscoe,                                                  BT 


From nobody Mon Apr 13 08:40:52 2015
Return-Path: <andrea.francini@alcatel-lucent.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 540801ACD3C for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 08:40:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.909
X-Spam-Level: 
X-Spam-Status: No, score=-6.909 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_HI=-5, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id mxU1vP9t3a3x for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 08:40:46 -0700 (PDT)
Received: from smtp-fr.alcatel-lucent.com (fr-hpida-esg-01.alcatel-lucent.com [135.245.210.20]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 840911ACC8C for <aqm@ietf.org>; Mon, 13 Apr 2015 08:40:46 -0700 (PDT)
Received: from us70tusmtp2.zam.alcatel-lucent.com (unknown [135.5.2.64]) by Websense Email Security Gateway with ESMTPS id 6EFCFD8388649 for <aqm@ietf.org>; Mon, 13 Apr 2015 15:40:42 +0000 (GMT)
Received: from US70UWXCHHUB02.zam.alcatel-lucent.com (us70uwxchhub02.zam.alcatel-lucent.com [135.5.2.49]) by us70tusmtp2.zam.alcatel-lucent.com (GMO) with ESMTP id t3DFegVf003671 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL) for <aqm@ietf.org>; Mon, 13 Apr 2015 11:40:43 -0400
Received: from US70TWXCHMBA11.zam.alcatel-lucent.com ([169.254.5.179]) by US70UWXCHHUB02.zam.alcatel-lucent.com ([135.5.2.49]) with mapi id 14.03.0195.001; Mon, 13 Apr 2015 11:40:42 -0400
From: "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com>
To: "aqm@ietf.org" <aqm@ietf.org>
Thread-Topic: References on AQM test results and fq_nocodel
Thread-Index: AdB2ACpSZak1wModQSWAzPrkxbVugA==
Date: Mon, 13 Apr 2015 15:40:41 +0000
Message-ID: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [135.5.27.18]
Content-Type: multipart/alternative; boundary="_000_1BFAC0A1D7955144A2444E902CB628F85E092996US70TWXCHMBA11z_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/s1zekNl6oeQ_TKYMZXVwhxPcBJ0>
Subject: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 15:40:50 -0000

--_000_1BFAC0A1D7955144A2444E902CB628F85E092996US70TWXCHMBA11z_
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

I need help finding a couple of references:


1.       Besides the results in Toke's ICCRG presentation at IETF 91 in Nov=
ember 2014 (http://www.ietf.org/proceedings/91/slides/slides-91-iccrg-4.pdf=
), where can I find  other comprehensive comparisons of AQM implementations=
 for Linux (or any other non-simulated system that handles real packets)?

2.       In Toke's slides, the fq_nocodel scheme appears to be never worse =
than fq_codel (or any other AQM in those tests), except for the VoIP delay =
plot of slide 12. Besides the Linux code, is there anywhere a description o=
f the fq_nocodel scheme?

Thanks!

Andrea

--_000_1BFAC0A1D7955144A2444E902CB628F85E092996US70TWXCHMBA11z_
Content-Type: text/html; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable

<html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-micr=
osoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:word" =
xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml" xmlns=3D"http:=
//www.w3.org/TR/REC-html40">
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-ascii"=
>
<meta name=3D"Generator" content=3D"Microsoft Word 12 (filtered medium)">
<style><!--
/* Font Definitions */
@font-face
	{font-family:"Cambria Math";
	panose-1:2 4 5 3 5 4 6 3 2 4;}
@font-face
	{font-family:Calibri;
	panose-1:2 15 5 2 2 2 4 3 2 4;}
/* Style Definitions */
p.MsoNormal, li.MsoNormal, div.MsoNormal
	{margin:0in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
a:link, span.MsoHyperlink
	{mso-style-priority:99;
	color:blue;
	text-decoration:underline;}
a:visited, span.MsoHyperlinkFollowed
	{mso-style-priority:99;
	color:purple;
	text-decoration:underline;}
p.MsoListParagraph, li.MsoListParagraph, div.MsoListParagraph
	{mso-style-priority:34;
	margin-top:0in;
	margin-right:0in;
	margin-bottom:0in;
	margin-left:.5in;
	margin-bottom:.0001pt;
	font-size:11.0pt;
	font-family:"Calibri","sans-serif";}
span.EmailStyle17
	{mso-style-type:personal-compose;
	font-family:"Calibri","sans-serif";
	color:windowtext;}
.MsoChpDefault
	{mso-style-type:export-only;}
@page WordSection1
	{size:8.5in 11.0in;
	margin:1.0in 1.0in 1.0in 1.0in;}
div.WordSection1
	{page:WordSection1;}
/* List Definitions */
@list l0
	{mso-list-id:1092166906;
	mso-list-type:hybrid;
	mso-list-template-ids:-1788021404 67698703 67698713 67698715 67698703 6769=
8713 67698715 67698703 67698713 67698715;}
@list l0:level1
	{mso-level-tab-stop:none;
	mso-level-number-position:left;
	text-indent:-.25in;}
ol
	{margin-bottom:0in;}
ul
	{margin-bottom:0in;}
--></style><!--[if gte mso 9]><xml>
<o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext=3D"edit">
<o:idmap v:ext=3D"edit" data=3D"1" />
</o:shapelayout></xml><![endif]-->
</head>
<body lang=3D"EN-US" link=3D"blue" vlink=3D"purple">
<div class=3D"WordSection1">
<p class=3D"MsoNormal">I need help finding a couple of references:<o:p></o:=
p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoListParagraph" style=3D"text-indent:-.25in;mso-list:l0 level=
1 lfo1"><![if !supportLists]><span style=3D"mso-list:Ignore">1.<span style=
=3D"font:7.0pt &quot;Times New Roman&quot;">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&=
nbsp;
</span></span><![endif]>Besides the results in Toke&#8217;s ICCRG presentat=
ion at IETF 91 in November 2014 (http://www.ietf.org/proceedings/91/slides/=
slides-91-iccrg-4.pdf), where can I find &nbsp;other comprehensive comparis=
ons of AQM implementations for Linux (or any
 other non-simulated system that handles real packets)? <o:p></o:p></p>
<p class=3D"MsoListParagraph" style=3D"text-indent:-.25in;mso-list:l0 level=
1 lfo1"><![if !supportLists]><span style=3D"mso-list:Ignore">2.<span style=
=3D"font:7.0pt &quot;Times New Roman&quot;">&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&=
nbsp;
</span></span><![endif]>In Toke&#8217;s slides, the fq_nocodel scheme appea=
rs to be never worse than fq_codel (or any other AQM in those tests), excep=
t for the VoIP delay plot of slide 12. Besides the Linux code, is there any=
where a description of the fq_nocodel
 scheme?<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Thanks!<o:p></o:p></p>
<p class=3D"MsoNormal"><o:p>&nbsp;</o:p></p>
<p class=3D"MsoNormal">Andrea<o:p></o:p></p>
</div>
</body>
</html>

--_000_1BFAC0A1D7955144A2444E902CB628F85E092996US70TWXCHMBA11z_--


From nobody Mon Apr 13 08:48:28 2015
Return-Path: <michawe@ifi.uio.no>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 49CF21ACDE6 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 08:48:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.91
X-Spam-Level: 
X-Spam-Status: No, score=-1.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id s2HtEk9jT3v8 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 08:48:25 -0700 (PDT)
Received: from mail-out5.uio.no (mail-out5.uio.no [IPv6:2001:700:100:10::17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 119301ACDE5 for <aqm@ietf.org>; Mon, 13 Apr 2015 08:48:25 -0700 (PDT)
Received: from mail-mx1.uio.no ([129.240.10.29]) by mail-out5.uio.no with esmtp (Exim 4.80.1) (envelope-from <michawe@ifi.uio.no>) id 1Yhgb1-0000cC-4P; Mon, 13 Apr 2015 17:48:23 +0200
Received: from mail-ex01.exprod.uio.no ([129.240.52.4]) by mail-mx1.uio.no with esmtps (TLSv1:AES256-SHA:256) (Exim 4.80) (envelope-from <michawe@ifi.uio.no>) id 1Yhgb0-0000GR-Nn; Mon, 13 Apr 2015 17:48:23 +0200
Received: from mail-ex01.exprod.uio.no (2001:700:100:52::4) by mail-ex01.exprod.uio.no (2001:700:100:52::4) with Microsoft SMTP Server (TLS) id 15.0.1044.25; Mon, 13 Apr 2015 17:48:20 +0200
Received: from mail-ex01.exprod.uio.no ([fe80::5c5b:6cfb:485b:6990]) by mail-ex01.exprod.uio.no ([fe80::5c5b:6cfb:485b:6990%20]) with mapi id 15.00.1044.021; Mon, 13 Apr 2015 17:48:20 +0200
From: Michael Welzl <michawe@ifi.uio.no>
To: "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com>
Thread-Topic: [aqm] References on AQM test results and fq_nocodel
Thread-Index: AQHQdgFDhleuDFqyOkSpsR3EBRpFtA==
Date: Mon, 13 Apr 2015 15:48:19 +0000
Message-ID: <CD9871D7-0C3C-4397-AB3C-27A6F56E2B0B@ifi.uio.no>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com>
In-Reply-To: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com>
Accept-Language: en-GB, nb-NO, en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-ms-exchange-transport-fromentityheader: Hosted
x-originating-ip: [129.240.169.59]
Content-Type: text/plain; charset="utf-8"
Content-ID: <1C5BCACBD36EFA4FA7D0B1F92BF10563@mail.uio.no>
Content-Transfer-Encoding: base64
MIME-Version: 1.0
X-UiO-SPF-Received: 
X-UiO-Ratelimit-Test: rcpts/h 11 msgs/h 4 sum rcpts/h 15 sum msgs/h 7 total rcpts 27743 max rcpts/h 44 ratelimit 0
X-UiO-Spam-info: not spam, SpamAssassin (score=-5.0, required=5.0, autolearn=disabled, T_RP_MATCHES_RCVD=-0.01, UIO_MAIL_IS_INTERNAL=-5, uiobl=NO, uiouri=NO)
X-UiO-Scanned: CCDEE93CD853C2B81A478CF1E1092232D2CB1E37
X-UiO-SPAM-Test: remote_host: 129.240.52.4 spam_score: -49 maxlevel 80 minaction 2 bait 0 mail/h: 96 total 690983 max/h 480 blacklist 0 greylist 0 ratelimit 0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/4Xxd1RnvfH9kTzDvNj2t63z9RTE>
Cc: "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 15:48:27 -0000
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From nobody Mon Apr 13 09:35:11 2015
Return-Path: <toke@toke.dk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 53B8B1ACEB6 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 09:35:09 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.206
X-Spam-Level: *
X-Spam-Status: No, score=1.206 tagged_above=-999 required=5 tests=[BAYES_40=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_DK=1.009, MIME_8BIT_HEADER=0.3, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id POQC1p90nb1U for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 09:35:08 -0700 (PDT)
Received: from mail2.tohojo.dk (mail2.tohojo.dk [77.235.48.147]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id E39341ACEB0 for <aqm@ietf.org>; Mon, 13 Apr 2015 09:35:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at mail2.tohojo.dk
Sender: toke@toke.dk
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toke.dk; s=201310; t=1428940437; bh=VwTs8yCXlUi6OpYcsFQpSEUsvrlhw2/Rqtj6ZJrmCE8=; h=From:To:Cc:Subject:References:Date:In-Reply-To; b=Ww50qvZUB+4SozSN1QAMaSKG1V4ZWTji1rTLg+DnsPwivE1NFie4sdsKv1TZlCMnD RUMGVe5T3lFcM0vx1cohV4YZG1cjTpOpinTvOWvn6bJiBKAAFfL930AeSyUrhyq3Qy enISf236ilEQ1Ve1vK4P5CnVhA9C56j57K89XQCw=
Received: by alrua-kau.kau.toke.dk (Postfix, from userid 1000) id 73A38C40445; Mon, 13 Apr 2015 17:53:56 +0200 (CEST)
From: =?utf-8?Q?Toke_H=C3=B8iland-J=C3=B8rgensen?= <toke@toke.dk>
To: "Francini\, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com>
Date: Mon, 13 Apr 2015 17:53:56 +0200
In-Reply-To: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> (Andrea Francini's message of "Mon, 13 Apr 2015 15:40:41 +0000")
X-Clacks-Overhead: GNU Terry Pratchett
Message-ID: <87vbh010tn.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/Rm6_wKKav7oeUOrIHmqRxGly8Hs>
Cc: "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 16:35:09 -0000

Hi Andrea

> 1. Besides the results in Toke=E2=80=99s ICCRG presentation at IETF 91 in
> November 2014
> (http://www.ietf.org/proceedings/91/slides/slides-91-iccrg-4.pdf),
> where can I find other comprehensive comparisons of AQM
> implementations for Linux (or any other non-simulated system that
> handles real packets)?

The data underlying my experiments are available from here:
http://www.cs.kau.se/tohojo/modern-aqms/ -- that page also has links to
the test scripts used to run the experiments. The paper is still under
submission, so can't link that, sadly, but I can share a copy with you
privately if you are interested.

> 2. In Toke=E2=80=99s slides, the fq_nocodel scheme appears to be never wo=
rse
> than fq_codel (or any other AQM in those tests), except for the VoIP
> delay plot of slide 12. Besides the Linux code, is there anywhere a
> description of the fq_nocodel scheme?

Well, fq_nocodel is basically the term I use for fq_codel configured so
as to disable the CoDel algorithm (i.e. the target and/or interval is
set to 100 seconds so the algorithm never triggers). As such, it doesn't
have a description outside of my paper, but you can read the fq_codel
draft and ignore everything it says about CoDel, I suppose :)

Also, note that configuring fq_codel in this way in Linux is not optimal
in terms of CPU use: it deliberate triggers the overflow drop mechanism,
which does a linear search through all configured queues, eating up 4k
of cache on every drop. If your test router is sufficiently
CPU-over-provisioned (a modern x86 counts as this for <=3D1Gbit speeds)
this doesn't affect performance, but I wouldn't recommend it on a tiny
MIPS processor, for instance. (Oh, and this limitation is specific to
the current implementation, and could easily be fixed if anyone wanted
to; the overflow mechanism was meant as a backup, and so wasn't
optimised like, say, the SFQ was).

-Toke


From nobody Mon Apr 13 10:24:41 2015
Return-Path: <andrea.francini@alcatel-lucent.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 85E2B1ACF1C for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 10:24:39 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -6.91
X-Spam-Level: 
X-Spam-Status: No, score=-6.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_HI=-5, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GSqgkCY8aGIX for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 10:24:37 -0700 (PDT)
Received: from smtp-fr.alcatel-lucent.com (fr-hpida-esg-01.alcatel-lucent.com [135.245.210.20]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 1A6C81ACF1D for <aqm@ietf.org>; Mon, 13 Apr 2015 10:24:36 -0700 (PDT)
Received: from us70tusmtp2.zam.alcatel-lucent.com (unknown [135.5.2.64]) by Websense Email Security Gateway with ESMTPS id 8F2D83ABBD86B; Mon, 13 Apr 2015 17:24:27 +0000 (GMT)
Received: from US70TWXCHHUB04.zam.alcatel-lucent.com (us70twxchhub04.zam.alcatel-lucent.com [135.5.2.36]) by us70tusmtp2.zam.alcatel-lucent.com (GMO) with ESMTP id t3DHORJA015399 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Mon, 13 Apr 2015 13:24:27 -0400
Received: from US70TWXCHMBA11.zam.alcatel-lucent.com ([169.254.5.179]) by US70TWXCHHUB04.zam.alcatel-lucent.com ([135.5.2.36]) with mapi id 14.03.0195.001; Mon, 13 Apr 2015 13:24:27 -0400
From: "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com>
To: "toke@toke.dk" <toke@toke.dk>
Thread-Topic: [aqm] References on AQM test results and fq_nocodel
Thread-Index: AdB2ACpSZak1wModQSWAzPrkxbVugAAB6VUXAACscRA=
Date: Mon, 13 Apr 2015 17:24:27 +0000
Message-ID: <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk>
In-Reply-To: <87vbh010tn.fsf@toke.dk>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [135.5.27.18]
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/7d5fmmyecaaFDQg6tShFI-e7Q9w>
Cc: "aqm@ietf.org" <aqm@ietf.org>, Michael Welzl <michawe@ifi.uio.no>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 17:24:39 -0000
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From nobody Mon Apr 13 10:41:29 2015
Return-Path: <toke@toke.dk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 0B11F1AD060 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 10:41:28 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.206
X-Spam-Level: *
X-Spam-Status: No, score=1.206 tagged_above=-999 required=5 tests=[BAYES_20=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_DK=1.009, MIME_8BIT_HEADER=0.3, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id nHfnVDEExlGR for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 10:41:26 -0700 (PDT)
Received: from mail2.tohojo.dk (mail2.tohojo.dk [77.235.48.147]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id A66D91AD072 for <aqm@ietf.org>; Mon, 13 Apr 2015 10:41:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at mail2.tohojo.dk
Sender: toke@toke.dk
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toke.dk; s=201310; t=1428946875; bh=RQbUjUPWLYvyFTbMADDs0iI9yfHpakEG5uegV3EcUM4=; h=From:To:Cc:Subject:References:Date:In-Reply-To; b=QMYZ4nBGWbH5Cuah8XPjeiaEFL2rTeUOOzfu5gvYtk/urIfA86bc02MKXeFtsNaum qzTS5xK4MLepbVdfzzXaFZJj3OY7Fn0BqQcsHUacAgNFmRsJPpl7z+bLco/6PPkb9i QCL2yi13odHdH/Gzcz71ikyA/YYRIycHZnCmg65g=
Received: by alrua-karlstad.karlstad.toke.dk (Postfix, from userid 1000) id 4635130DB1B; Mon, 13 Apr 2015 19:41:14 +0200 (CEST)
From: =?utf-8?Q?Toke_H=C3=B8iland-J=C3=B8rgensen?= <toke@toke.dk>
To: "Francini\, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com>
Date: Mon, 13 Apr 2015 19:41:14 +0200
In-Reply-To: <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> (Andrea Francini's message of "Mon, 13 Apr 2015 17:24:27 +0000")
X-Clacks-Overhead: GNU Terry Pratchett
Message-ID: <87r3rolydh.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/Hl5AIM8GllrfaSUoOul6HowochM>
Cc: "aqm@ietf.org" <aqm@ietf.org>, Michael Welzl <michawe@ifi.uio.no>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 17:41:28 -0000

"Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com> writes:

> CoDel's effort to limit the queuing delay appears well motivated when
> there is only one queue (e.g., to shield VoIP from the delay induced
> by bulk-transfer or video traffic), but when multiple queues are
> available the net gain of the effort is unclear. I suspect the same
> may happen with the flow-queue version of PIE.

Yep, this has been my conclusion as well, and the test results on those
slides are what convinced me of it: flow queueing is much more important
than AQM, for most applications. However, note that in the tests I have
run, the latency is measured by a separate flow; the latency of the
packets making up the bulk flows is not measured. For most common
traffic, this is fine, but there can be exceptions where you have a flow
that has enough traffic to induce a queue, but which also requires low
latency throughout. Adaptive video codecs for real-time media is a
notable example. For those, having the AQM is nice.

Also, having the AQM in place guards you against hash collisions. With
the AQM in place, you degrade to its performance, whereas without,
you're basically back to a large, dumb FIFO queue. fq_codel as it exists
now is unfortunately not immune to those; we're working (in the
Bufferbloat community) on an improved queueing discipline that solves
this in most cases. See http://www.bufferbloat.net/projects/codel/wiki/Cake

> Is this just because the set of tests in the slides does not include
> one that clearly exposes the superiority of fq_codel over fq_nocodel?
> And what would that test look like?

Well, see above. Adaptive real-time video in particular is under-tested,
in part due to a lack of good tools to generate traffic. Also, if you
want to see fq_* break, try sticking all your flows in an encrypted VPN :)

-Toke


From nobody Mon Apr 13 11:42:39 2015
Return-Path: <gettysjim@gmail.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id A53A11B2A36 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 11:42:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.723
X-Spam-Level: *
X-Spam-Status: No, score=1.723 tagged_above=-999 required=5 tests=[BAYES_50=0.8, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, FM_FORGED_GMAIL=0.622, FREEMAIL_FROM=0.001, HTML_MESSAGE=0.001, MIME_8BIT_HEADER=0.3, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id wox4AJzso1pR for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 11:42:34 -0700 (PDT)
Received: from mail-wi0-x22c.google.com (mail-wi0-x22c.google.com [IPv6:2a00:1450:400c:c05::22c]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 7815C1AD26E for <aqm@ietf.org>; Mon, 13 Apr 2015 11:42:33 -0700 (PDT)
Received: by widdi4 with SMTP id di4so84551444wid.0 for <aqm@ietf.org>; Mon, 13 Apr 2015 11:42:32 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113;  h=mime-version:sender:in-reply-to:references:date:message-id:subject :from:to:cc:content-type; bh=ADrt31CEgtZgT+8lZsjx9bocxzXxRs745c4lRDGyj18=; b=wWjaA73Og5KE0td51Mt4xuymRgDlZKCL0CCTTJ+TD7eRQHquN48p79N1O7T+xGaEaE qqF8nsOgswaqgXfhH/JSotyIXd7lrCl+uSvE3oHk4DggHsZDB/rtAADnyDHuVbIrItx/ sDB2ePkGfW8YgTFt2j7uAgXQlyPfju6YKv8QpICGM9nSSqGXQlvItARQV8Sf1gqjz7rB faS0WEokW3nYFY9TGG9R2gChO409AbwqXBMsxj1wvhB+LEjtDzLYNRyOc9HcM9+zCKdz jv5adGatwAB0WnCJf0eeeaWYRl1+DN7Yf8r35wxOVBnFqS4W9jjxpUar75/eiZ6T7XG4 cCmg==
MIME-Version: 1.0
X-Received: by 10.180.78.65 with SMTP id z1mr24151360wiw.14.1428950552188; Mon, 13 Apr 2015 11:42:32 -0700 (PDT)
Sender: gettysjim@gmail.com
Received: by 10.194.118.166 with HTTP; Mon, 13 Apr 2015 11:42:32 -0700 (PDT)
In-Reply-To: <87r3rolydh.fsf@toke.dk>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk>
Date: Mon, 13 Apr 2015 14:42:32 -0400
X-Google-Sender-Auth: h4IfX43bZzEW_xeW7A2W45i3TNs
Message-ID: <CAGhGL2Cn4CyBe0ZpKK6Mfca-bHnYmSC3BaTx5FZX5=hD5qhGVw@mail.gmail.com>
From: Jim Gettys <jg@freedesktop.org>
To: =?UTF-8?B?VG9rZSBIw7hpbGFuZC1Kw7hyZ2Vuc2Vu?= <toke@toke.dk>
Content-Type: multipart/alternative; boundary=f46d043be13c109c3305139f7add
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/7zWEX2j3K4FJH9Dqu6wwNLAjEmU>
Cc: Michael Welzl <michawe@ifi.uio.no>, "aqm@ietf.org" <aqm@ietf.org>, "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 18:42:36 -0000

--f46d043be13c109c3305139f7add
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Note that having no AQM at all, and only flow queuing (e.g. fq_nocodel),
leaves some applications in dire straights, being both interactive, and
using a single TCP stream (for example, you really don't want to run my
favorite window system, X, over a bloated link without some form of
mark/drop algorithm to keep TCP behaving sanely).  And with RTT's of those
applications growing without bounds, they become really insane charging
elephants.  Ironically, I used to be able to run X over intercontinental
links (e.g. 150ms RTT's); trying that on today's bloated edge puts you in a
world of hurt.

So both a mark/drop algorithm and flow queuing are needed to really solve
the problem, though as Toke says, fq makes the biggest difference if you
only pick one. But this is really more of an artifact of the web's abuse of
TCP; tons of short, transient connections...
                                            - Jim



On Mon, Apr 13, 2015 at 1:41 PM, Toke H=C3=B8iland-J=C3=B8rgensen <toke@tok=
e.dk>
wrote:

> "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com> writes:
>
> > CoDel's effort to limit the queuing delay appears well motivated when
> > there is only one queue (e.g., to shield VoIP from the delay induced
> > by bulk-transfer or video traffic), but when multiple queues are
> > available the net gain of the effort is unclear. I suspect the same
> > may happen with the flow-queue version of PIE.
>
> Yep, this has been my conclusion as well, and the test results on those
> slides are what convinced me of it: flow queueing is much more important
> than AQM, for most applications. However, note that in the tests I have
> run, the latency is measured by a separate flow; the latency of the
> packets making up the bulk flows is not measured. For most common
> traffic, this is fine, but there can be exceptions where you have a flow
> that has enough traffic to induce a queue, but which also requires low
> latency throughout. Adaptive video codecs for real-time media is a
> notable example. For those, having the AQM is nice.
>
> Also, having the AQM in place guards you against hash collisions. With
> the AQM in place, you degrade to its performance, whereas without,
> you're basically back to a large, dumb FIFO queue. fq_codel as it exists
> now is unfortunately not immune to those; we're working (in the
> Bufferbloat community) on an improved queueing discipline that solves
> this in most cases. See
> http://www.bufferbloat.net/projects/codel/wiki/Cake
>
> > Is this just because the set of tests in the slides does not include
> > one that clearly exposes the superiority of fq_codel over fq_nocodel?
> > And what would that test look like?
>
> Well, see above. Adaptive real-time video in particular is under-tested,
> in part due to a lack of good tools to generate traffic. Also, if you
> want to see fq_* break, try sticking all your flows in an encrypted VPN :=
)
>
> -Toke
>
> _______________________________________________
> aqm mailing list
> aqm@ietf.org
> https://www.ietf.org/mailman/listinfo/aqm
>

--f46d043be13c109c3305139f7add
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_default" style=3D"font-size:small">Not=
e that having no AQM at all, and only flow queuing (e.g. fq_nocodel), leave=
s some applications in dire straights, being both interactive, and using a =
single TCP stream (for example, you really don&#39;t want to run my favorit=
e window system, X, over a bloated link without some form of mark/drop algo=
rithm to keep TCP behaving sanely).=C2=A0 And with RTT&#39;s of those appli=
cations growing without bounds, they become really insane charging elephant=
s.=C2=A0 Ironically, I used to be able to run X over intercontinental links=
 (e.g. 150ms RTT&#39;s); trying that on today&#39;s bloated edge puts you i=
n a world of hurt.</div><div class=3D"gmail_default" style=3D"font-size:sma=
ll"><br></div><div class=3D"gmail_default" style=3D"font-size:small">So bot=
h a mark/drop algorithm and flow queuing are needed to really solve the pro=
blem, though as Toke says, fq makes the biggest difference if you only pick=
 one. But this is really more of an artifact of the web&#39;s abuse of TCP;=
 tons of short, transient connections...</div><div class=3D"gmail_default" =
style=3D"font-size:small">=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 - Jim</div><div class=3D"gmail_default" sty=
le=3D"font-size:small"><br></div><div class=3D"gmail_default" style=3D"font=
-size:small"><br></div></div><div class=3D"gmail_extra"><br><div class=3D"g=
mail_quote">On Mon, Apr 13, 2015 at 1:41 PM, Toke H=C3=B8iland-J=C3=B8rgens=
en <span dir=3D"ltr">&lt;<a href=3D"mailto:toke@toke.dk" target=3D"_blank">=
toke@toke.dk</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" sty=
le=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span =
class=3D"">&quot;Francini, Andrea (Andrea)&quot; &lt;<a href=3D"mailto:andr=
ea.francini@alcatel-lucent.com">andrea.francini@alcatel-lucent.com</a>&gt; =
writes:<br>
<br>
&gt; CoDel&#39;s effort to limit the queuing delay appears well motivated w=
hen<br>
&gt; there is only one queue (e.g., to shield VoIP from the delay induced<b=
r>
&gt; by bulk-transfer or video traffic), but when multiple queues are<br>
&gt; available the net gain of the effort is unclear. I suspect the same<br=
>
&gt; may happen with the flow-queue version of PIE.<br>
<br>
</span>Yep, this has been my conclusion as well, and the test results on th=
ose<br>
slides are what convinced me of it: flow queueing is much more important<br=
>
than AQM, for most applications. However, note that in the tests I have<br>
run, the latency is measured by a separate flow; the latency of the<br>
packets making up the bulk flows is not measured. For most common<br>
traffic, this is fine, but there can be exceptions where you have a flow<br=
>
that has enough traffic to induce a queue, but which also requires low<br>
latency throughout. Adaptive video codecs for real-time media is a<br>
notable example. For those, having the AQM is nice.<br>
<br>
Also, having the AQM in place guards you against hash collisions. With<br>
the AQM in place, you degrade to its performance, whereas without,<br>
you&#39;re basically back to a large, dumb FIFO queue. fq_codel as it exist=
s<br>
now is unfortunately not immune to those; we&#39;re working (in the<br>
Bufferbloat community) on an improved queueing discipline that solves<br>
this in most cases. See <a href=3D"http://www.bufferbloat.net/projects/code=
l/wiki/Cake" target=3D"_blank">http://www.bufferbloat.net/projects/codel/wi=
ki/Cake</a><br>
<span class=3D""><br>
&gt; Is this just because the set of tests in the slides does not include<b=
r>
&gt; one that clearly exposes the superiority of fq_codel over fq_nocodel?<=
br>
&gt; And what would that test look like?<br>
<br>
</span>Well, see above. Adaptive real-time video in particular is under-tes=
ted,<br>
in part due to a lack of good tools to generate traffic. Also, if you<br>
want to see fq_* break, try sticking all your flows in an encrypted VPN :)<=
br>
<div class=3D"HOEnZb"><div class=3D"h5"><br>
-Toke<br>
<br>
_______________________________________________<br>
aqm mailing list<br>
<a href=3D"mailto:aqm@ietf.org">aqm@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/aqm" target=3D"_blank">htt=
ps://www.ietf.org/mailman/listinfo/aqm</a><br>
</div></div></blockquote></div><br></div>

--f46d043be13c109c3305139f7add--


From nobody Mon Apr 13 14:47:15 2015
Return-Path: <david@lang.hm>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 9615E1A897D for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 14:47:14 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.91
X-Spam-Level: 
X-Spam-Status: No, score=-1.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 0f4vN0DjBA5v for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 14:47:11 -0700 (PDT)
Received: from bifrost.lang.hm (mail.lang.hm [64.81.33.126]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id CF8051A8938 for <aqm@ietf.org>; Mon, 13 Apr 2015 14:47:10 -0700 (PDT)
Received: from asgard.lang.hm (asgard.lang.hm [10.0.0.100]) by bifrost.lang.hm (8.13.4/8.13.4/Debian-3) with ESMTP id t3DLkgkw006770; Mon, 13 Apr 2015 14:46:42 -0700
Date: Mon, 13 Apr 2015 14:46:42 -0700 (PDT)
From: David Lang <david@lang.hm>
X-X-Sender: dlang@asgard.lang.hm
To: Bob Briscoe <bob.briscoe@bt.com>
In-Reply-To: <201504131511.t3DFBG3R002270@bagheera.jungle.bt.co.uk>
Message-ID: <alpine.DEB.2.02.1504131441190.11469@nftneq.ynat.uz>
References: <23AFEFE3-4D93-4DD9-A22B-952C63DB9FE3@cisco.com> <BF6B00CC65FD2D45A326E74492B2C19FB75BAA82@FR711WXCHMBA05.zeu.alcatel-lucent.com> <72EE366B-05E6-454C-9E53-5054E6F9E3E3@ifi.uio.no> <55146DB9.7050501@rogers.com> <08C34E4A-DFB7-4816-92AE-2ED161799488@ifi.uio.no> <BF6B00CC65FD2D45A326E74492B2C19FB75BAFA0@FR711WXCHMBA05.zeu.alcatel-lucent.com> <alpine.DEB.2.02.1503271024550.2416@nftneq.ynat.uz> <5d58d2e21400449280173aa63069bf7a@hioexcmbx05-prd.hq.netapp.com> <20150327183659.GI39886@verdi> <72C12F6B-9DDE-4483-81F2-2D9A0F2D3A48@cs.columbia.edu> <alpine.DEB.2.02.1503271211200.19390@nftneq.ynat.uz> <D13AFCE7.46BC%kk@cs.ucr.edu> <alpine.DEB.2.02.1503271257230.19390@nftneq.ynat.uz> <AE342093-DE05-4D93-96DA-EB07E221F1D9@netapp.com> <alpine.DEB.2.02.1503291923300.26044@nftneq.ynat.uz> <201504131511.t3DFBG3R002270@bagheera.jungle.bt.co.uk>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/Bx6JkablSwLPHx4gW9MdC6YTzfc>
Cc: "Scheffenegger, Richard" <rs@netapp.com>, Vishal Misra <misra@cs.columbia.edu>, KK <kk@cs.ucr.edu>, John Leslie <john@jlc.net>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] think once to mark, think twice to drop:  draft-ietf-aqm-ecn-benefits-02
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 13 Apr 2015 21:47:14 -0000

On Mon, 13 Apr 2015, Bob Briscoe wrote:

> David,
>
> Returning from a fortnight offlist...
>
> I think your conception of how ECN works is incorrect. You describe ECN as if 
> the AQM marks one packet when it drops another packet. You say that the 
> ECN-mark speeds up the retransmission of the dropped packet. On the contrary, 
> the idea of classic ECN [RFC3168] is that the ECN marks replace the drops. In 
> all known testing (except pathological cases), classic ECN effectively 
> eliminates drops for all ECN-capable packets.

That's what I thought, and if that was the case, then marking packets as 
ECN-capable would mean that they would have an advantage over non-ECN packets 
(by not getting dropped, so getting a higher share of bandwidth)

that's what the gaming ECN thread was about, and if I understood the responses, 
I was being told that marking packets as ECN-capable, but not slowing down 
(actually responding to ECN) would not let an application get any advantage 
because the packets would just end up getting dropped anyway, since marking and 
dropping happen at the same level, even on ECN-capable flows.

If the packets are just marked, but not dropped, then the ECN-capable flows will 
occupy a disproportinate share of the available buffer space, since they just 
get marked instead of dropped.

David Lang

> Nonetheless, I do agree with your sentiment that the perfect is the enemy of 
> the good. We can remove most of the really bloat-induced latency without ECN. 
> So the message must be clear: Deploy AQM now. No need to wait for ECN. But 
> implementations SHOULD allow ECN packets to be classifed into a separately 
> configurable instance of the AQM algo. {Note 1}
>
> Similarly, this WG made sure we did not deprecate RED in the AQM 
> recommendations. Because, in existing equipment, even a poorly tuned RED is 
> usually much better than a bloated buffer with no AQM.
>
> Just as the WG mustn't confuse messages, you mustn't get confused by a 
> discussion about the potential for more ambitious reductions in latency. Koen 
> started the thread with reference to our presentation in the ICCRG in the 
> IRTF, where 'R' in both cases stands for Research.
>
> And I believe it is valid for the ECN benefits draft (in the IETF AQM WG) to 
> point to the potential of ECN, by using insights from research in progress.
>
> There are a number of different contributions to unnecessary latency, not 
> just the one popularised as bufferbloat:
>
> [In the following, I will use the term 'short message(s)' as shorthand for 
> either a short interactive flow or a long flow consisting of short 
> interactive messages (like in a game), or video frames or voice datagrams or 
> anything where the perceived latency depends on the latency of each 
> 'message', especially a string of messages with serial dependency as is 
> typical in Web.]
>
> #1 The well-known bufferbloat problem, where a long-running flow fills a 
> bloated buffer, delays short messages.
>  - AQM and/or flow queuing can remove this delay, without needing ECN.
>
> #2 A loss causes head-of-line blocking for short message(s) while waiting for 
> the retransmission.
>  - AQM without ECN cannot remove this delay.
>  - Flow queuing cannot remove this delay, if the losses are self-induced.
>  - FEC can remove this delay without needing ECN, but the increased 
> redundancy is equivalent to poorer utilisation (altho only the short flows 
> need the redundancy).
>  - ECN can remove this delay.
>
> #3 A loss near the end of a short flow can lead to multi-RTT delay.
>  - AQM without ECN cannot remove this delay.
>  - Techniques like tail-loss probe and RTO-restart can mitigate this delay, 
> without ECN, but not remove it.
>  - ECN can remove this delay.
>
> #4 The Reno/Cubic sawtooth causes variation in delay between 1 and 2 base 
> RTTs. This can affect short messages.
>  - AQM without ECN cannot remove this delay unless configured to sacrifice 
> utilisation.
>  - Flow queuing removes this delay variation if caused by a separate 
> long-running flow.{Note 2}
>  - A change to the TCP algo (e.g. DCTCP) can remove this delay variation. 
> Smaller sawteeth imply a much higher signalling rate, which in turn requires 
> ECN, otherwise drop probability would be excessive. (This was the main point 
> Koen was making.)
>  - Therefore, ECN can remove this delay.
>
> #5 Slow-starts{Note 3} cause spikes in delay.
>  - AQM without ECN cannot remove this delay, and typically AQM is designed 
> to allow such bursts of delay in the hope they will disappear of their own 
> accord.
>  - Flow queuing can remove the effect of these delay bursts on other flows, 
> but only if it gives all flows a separate queue from the start.{Note 2}
>  - Delay-based softening of slow-start, such as Hybrid Slow-Start in Linux, 
> can mitigate these variations, but with increased risk of coming out of SS 
> early, causing significantly longer completion time.
>  - ECN with AQM based on the instantaneous queue limits this delay, without 
> the risk of longer completion time.
>
> #6 Slow-starts{Note 3} can cause runs of losses, which in turn cause delays.
>  - AQM without ECN cannot remove these delays.
>  - Flow queuing cannot remove these losses, if self-induced.
>  - Delay-based SS like HSS can mitigate these losses, with increased risk of 
> longer completion time.
>  - ECN can remove these losses, and the consequent delays.
>
> Summary:
> * AQM alone solves the main problem
> * Flow queuing solves or mitigates most of the remaining secondary problems.
> * ECN has the potential to solve all the remaining secondary problems 
> (pending further research to prove some of them).
>
> Whether flow queuing is applicable depends on the scale. The work I'm doing 
> with Koen is to reduce the cost of the queuing mechanisms on our BNGs 
> (broadband network gateways). We're trying to reduce the cost of per-customer 
> queuing at scale, so per-flow queuing is simply out of the question. Whereas 
> ECN requires no more processing than drop.
>
> ECN has potential cheating problems, but we have per-customer queues anyway.
> Using flow as the unit of allocation also has its own problems, with no 
> proposed solutions.
>
>
> Bob
>
> {Note 1}: Your general point is that the perfect can be the enemy of the 
> good. Here's the presentation I gave in TSVAREA straight after VJ presented 
> CoDel in 2012 entitled "DCTCP & CoDel; the Best is the Friend of the Good."
> <http://www.bobbriscoe.net/presents/1207ietf/1207-tsvarea-dctcp.pdf>
>
> {Note 2}:  A lone flow can cause this delay variation to itself, but that's 
> irrelevant because, if the delay were not in the network it would be at the 
> sender.
>
> {Note 3}: Delay and loss spikes can equivalently be caused when Cubic's 
> window rises to seek out newly available capacity after another flow finishes 
> or the link rate varies.
>
> At 05:16 30/03/2015, David Lang wrote:
>> On Sat, 28 Mar 2015, Scheffenegger, Richard wrote:
>> 
>>> David,
>>> 
>>> Perhaps you would care to provide some text to address the misconception 
>>> that you pointed out? (To wait for a 100% fix as a 90% fix appears much 
>>> less appealing, while the current state of art is at 0%)
>> 
>> Ok, you put me on the spot :-) Here goes.
>> 
>>> If you think that aqm-recommendations is not strogly enough worded. I 
>>> think this particular discussion (to aqm or not) really belongs there. The 
>>> other document (ecn benefits) has a different target in arguing for going 
>>> those last 10%...
>> 
>> so here is my "elevator pitch" on the problem. Feel free to take anything I 
>> say here for any purpose, and I'm sure I'll get corrected for anything I am 
>> wong on
>> 
>> Problem statement: Transmit buffers are needed to keep the network layer 
>> fully utilized, but excessive buffers result in poor latency for all 
>> traffic. This latency is frequently bad enough to cause some types of 
>> traffic to fail entirely.
>>
>>   <link to more background goes here, including how separate benchmarks for 
>> throughput and latency have mislead people, "packet loss considered evil", 
>> cheaper memory encouraging larger buffers, etc. Include tests like 
>> netperf-wrapper and ping latency while under load, etc. Include examples 
>> where buffers have resulted in latencies so long that packets are 
>> retransmitted before the first copy gets to the destination>
>> 
>> Traditionally, transmit buffers have been sized to handle a fixed number of 
>> packets. Due to teh variation in packet sizes, it is impossible to tune 
>> this value to both keep the link fully utilized when small packets dominate 
>> the trafific without having the queue size be large enough to cause latency 
>> problems when large packets dominate the traffic.
>> 
>> Shifting to Byte Queue Lengths where queues are allowed to hold a variable 
>> number of packets depending on how large they are makes it possible to 
>> manually tune the transmit buffer size to get good latency under all 
>> traffic conditions at a given speed. However, this step forward revealed 
>> two additional problems.
>> 
>> 1. whenever the data rate changes, this value needs to be manually changed 
>> (multi-link paths loose a link, noise degrades max throughput on a link, 
>> etc)
>> 
>> 2. high volume flows (i.e. bulk downloads) can starve other flows (DNS 
>> lookups, VoIP, Gaming, etc). this happens because space in tue queue is on 
>> a first-com-first-served basis, so the high-volume traffic fills the queue 
>> (at which point it starts to be dropped), but all other traffic that tries 
>> to arrive is also dropped. It turns out that these light flows tend to have 
>> a larger effect on the user experience than heavier flows, because things 
>> tend to be serialized behind the lighter flows (DNS lookup before doing a 
>> large download, retrieving a small HTML page to find what additional 
>> resources need to be fetched to display a page), or the user experience is 
>> directly effected by light flows (gaming lag, VoIP drops, etc)
>> 
>> Active Queue Management addresses these problems by adapting the amount of 
>> data that is buffered to match the data transmission capacity, and prevents 
>> high volume flows from starving low-volume flows without the need to 
>> implement QoS classifications.
>>
>>   <insert link about how you can't trust QoS tags that are made by other 
>> organizations, ways that it can be abused, etc>
>> 
>> This is possible because AQM algoithms don't have to drop the new packet 
>> that arrives, the algorithm can decide to drop the packet for one of the 
>> heavy flows rather than for one of the lightweight flows.
>> 
>> <insert references to currently favored AQM options here, PIE, fq_codel, 
>> cake, ???. Also links to failed approaches>
>> 
>> Turning on aqm on every bottleneck link makes the Internet usable for 
>> everyone, no matter what sort of application they are using.
>> 
>> <insert link on how to deal with equipment you can't configure by 
>> throttling bandwidth before the bottleneck oand/or doing ingress shaping of 
>> traffic>
>> 
>> 
>> While AQM makes the network usable, there is still additional room for 
>> improvement. While dropping packets does result in the TCP senders slowing 
>> down,and eventually stabilizing at around the right speed to keep the link 
>> fully utilized, the only way that senders have been able to detect problems 
>> is to discover that they have not received an ack for the traffic within 
>> the allowed time. This causes a 'bubble' in the flow as teh dropped packet 
>> must be retransmitted (and sometimes a significant amount of data after the 
>> dropped packet that did make it to the destination, but could not be acked 
>> because fo the missing packet).
>> 
>> This "bubble" in the data flow can be greatly compressed by configuring the 
>> AQM algorithm to send an ECN packet to the sender when it drops a packet in 
>> a flow. The sender can then adapt faster, slowing down it's new data, and 
>> re-sending the dropped packet without having to wait for the timeout. This 
>> has two major effects by allowing the sender to retransmit the packet 
>> sooner the dealy on the dropped data is not as long, and because the 
>> replacement data can arrive before the timeout of the following packets, 
>> they may not need to be re-sent. by configuring the AQM algorithm to send 
>> the ECN notification to the sender only when the packet is being dropped, 
>> the effect of failure of the ECN packet to get through to the sender (the 
>> notification packet runs into congestion and gets dropped, some network 
>> device blocks it, etc) is that the ECN enabled case devolves to match the 
>> non-ECN case in that the sender will still detect the dropped packet via 
>> the timeout waiting for the ack as if ENCN was not enabled.
>> 
>> <insert link to possible problems that can happen here, including the 
>> potential for an app to 'game' things if packets are marked at a different 
>> level than when they are dropped.>
>> 
>> 
>> So a very strong recommendation to enable Active Queue Management, while 
>> the different algorithms have different advantages and levels of testing, 
>> even the 'worst' of the set results in a night-and-day improvement for 
>> usability compared to unmanaged buffers.
>> 
>> Enabling ECN at the same point as dropping packets as part of enabling any 
>> AQM algorithm results in a noticable improvement over the base algorithm 
>> without ECN. When compared to the baseline, the improvement added by ECN is 
>> tiny compared to the improvement from enabling AQM.
>>
>>   Is it fair to say that plain aqm vs aqm+ecn variation is on the same 
>> order of difference as the differences between the different AQM 
>> algorithms?
>> 
>> 
>> 
>> Future research items (which others here may already have done, and would 
>> not be part of my 'elevator pitch')
>> 
>> I believe that currently ECn triggers the exact same slowdown that a missed 
>> packet does, and it may be appropriate to have the sender do a less drastic 
>> slowdown.
>> 
>> It would be very interesing to provide soem way for the application sending 
>> the traffic to detect dropped packets and ECN responses. For example, a 
>> streaming media source (especially an interactive one like video 
>> conferencing) could adjust the bitrate that it's sending.
>> 
>> David Lang
>> 
>> _______________________________________________
>> aqm mailing list
>> aqm@ietf.org
>> https://www.ietf.org/mailman/listinfo/aqm
>
> ________________________________________________________________
> Bob Briscoe,                                                  BT 
>


From nobody Mon Apr 13 17:25:33 2015
Return-Path: <john@jlc.net>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 158241B2B6B for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 17:25:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.51
X-Spam-Level: 
X-Spam-Status: No, score=-1.51 tagged_above=-999 required=5 tests=[BAYES_50=0.8, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 4XTQfRzsFUJ6 for <aqm@ietfa.amsl.com>; Mon, 13 Apr 2015 17:25:29 -0700 (PDT)
Received: from mailhost.jlc.net (mailhost.jlc.net [199.201.159.4]) by ietfa.amsl.com (Postfix) with ESMTP id A74F41B2B66 for <aqm@ietf.org>; Mon, 13 Apr 2015 17:25:29 -0700 (PDT)
Received: by mailhost.jlc.net (Postfix, from userid 104) id 0B1A1C94BF; Mon, 13 Apr 2015 20:25:16 -0400 (EDT)
Date: Mon, 13 Apr 2015 20:25:16 -0400
From: John Leslie <john@jlc.net>
To: David Lang <david@lang.hm>
Message-ID: <20150414002515.GD63465@verdi>
References: <5d58d2e21400449280173aa63069bf7a@hioexcmbx05-prd.hq.netapp.com> <20150327183659.GI39886@verdi> <72C12F6B-9DDE-4483-81F2-2D9A0F2D3A48@cs.columbia.edu> <alpine.DEB.2.02.1503271211200.19390@nftneq.ynat.uz> <D13AFCE7.46BC%kk@cs.ucr.edu> <alpine.DEB.2.02.1503271257230.19390@nftneq.ynat.uz> <AE342093-DE05-4D93-96DA-EB07E221F1D9@netapp.com> <alpine.DEB.2.02.1503291923300.26044@nftneq.ynat.uz> <201504131511.t3DFBG3R002270@bagheera.jungle.bt.co.uk> <alpine.DEB.2.02.1504131441190.11469@nftneq.ynat.uz>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <alpine.DEB.2.02.1504131441190.11469@nftneq.ynat.uz>
User-Agent: Mutt/1.4.1i
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/JYiVdFvuqR0DXfvwKcR0hiEixo0>
Cc: aqm@ietf.org
Subject: Re: [aqm] think once to mark, think twice to drop:  draft-ietf-aqm-ecn-benefits-02
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 00:25:32 -0000

David Lang <david@lang.hm> wrote:
> On Mon, 13 Apr 2015, Bob Briscoe wrote:
> 
>> I think your conception of how ECN works is incorrect. You describe ECN as 
>> if the AQM marks one packet when it drops another packet. You say that the 
>> ECN-mark speeds up the retransmission of the dropped packet. On the 
>> contrary, the idea of classic ECN [RFC3168] is that the ECN marks replace 
>> the drops. In all known testing (except pathological cases), classic ECN 
>> effectively eliminates drops for all ECN-capable packets.
> 
> That's what I thought, and if that was the case,

   What does "that" refer to -- I honestly can't tell from your text.

   (We have an awful lot of talking past each other when we talk ECN:
it would really help to be as clear as possible.)

> then marking packets as ECN-capable would mean that they would have an
> advantage over non-ECN packets (by not getting dropped, so getting a
> higher share of bandwidth)

   I wish I could go back in time an excise every phrase saying "same as
a packet drop" from the ECN RFCs.

> that's what the gaming ECN thread was about, and if I understood the 
> responses, I was being told that marking packets as ECN-capable, but not 
> slowing down (actually responding to ECN) would not let an application get 
> any advantage because the packets would just end up getting dropped anyway, 
> since marking and dropping happen at the same level, even on ECN-capable 
> flows.

   No, I don't think any of us were saying that.

   A few of us _were_ saying that abuse like that is a small percentage
of the abuse where the sender simply doesn't reduce the sending rate if
packet loss is detected.

   And several folks _did_ say that it was important to have the marking
and dropping threshholds the same _and_ the reaction to marking the same
as the reaction to dropping be the same "so that neither would suffer in
comparison to the other".

   (They are wrong, BTW, but most of us just let this canard die down.)

> If the packets are just marked, but not dropped, then the ECN-capable flows 
> will occupy a disproportinate share of the available buffer space, since 
> they just get marked instead of dropped.

   True.

   But so what?

   What do you want to do about it? (That's my trademark question, BTW.)

   If you are the node experiencing congestion, and your buffer is full,
you absolutely should drop the marked packet. No disagreement there.

   (This can happen, BTW, in PIE, which marks packets as the enter the
buffer; and the buffer may fill before they can be sent.)

   But if your buffer isn't even close to full, what's wrong with ECN
capable packets using a larger share of the buffer? If they go to a
responsive flow, the reaction will be quicker than if they were dropped.

   Of course, if they go to a non-responsive flow, they would get a
slight advantage, but why is that _your_ concern? And there's really
nothing you can do to counter the advantage that doesn't break the ECN
paradigm for responsive flows.

   If instead, you are a node later on the path, there's nothing wrong
with dropping ECN packets with Congestion-Experienced marks _when_ you
need to drop packets. But there's no reason to "punish" them for having
occupied a "disproportionate share" of somebody else's buffer.

   It would be much better if we could think in terms of ECN-marking
_well_before_ there is any need to drop packets because the buffer is
full; and treating any "advantage" this might give them an an incentive
for more flows to enable ECN.

   One of the things I'm quite sure Bob Briscoe is thinking -- but not
writing -- is that he believes ECN can solve a lot of the problems he
named _if_ the ECN marking occurs at a lower threshhold than packet drop.
We are _not_ chartered to "fix ECN" in this WG: so that discussion doesn't
really belong in this thread; but you and Bob are unlikely to communicate
fully if you ignore that possibility.

   Hope this helps...

--
John Leslie <john@jlc.net>


From nobody Tue Apr 14 08:37:32 2015
Return-Path: <simon@superduper.net>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 2969B1B2D35 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 08:37:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.09
X-Spam-Level: *
X-Spam-Status: No, score=1.09 tagged_above=-999 required=5 tests=[BAYES_50=0.8, MIME_8BIT_HEADER=0.3, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kE0c34AHAcd0 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 08:37:29 -0700 (PDT)
Received: from masada.superduper.net (masada.superduper.net [85.119.82.91]) (using TLSv1.2 with cipher DHE-RSA-AES128-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id D321F1B2CEA for <aqm@ietf.org>; Tue, 14 Apr 2015 08:36:57 -0700 (PDT)
Received: from 199-116-72-167.public.monkeybrains.net ([199.116.72.167] helo=[192.168.0.12]) by masada.superduper.net with esmtpsa (TLS1.0:RSA_ARCFOUR_MD5:128) (Exim 4.80) (envelope-from <simon@superduper.net>) id 1Yi2tN-00089g-Ap; Tue, 14 Apr 2015 16:36:51 +0100
From: Simon Barber <simon@superduper.net>
To: =?UTF-8?B?VG9rZSBIw7hpbGFuZC1Kw7hyZ2Vuc2Vu?= <toke@toke.dk>, "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com>
Date: Tue, 14 Apr 2015 08:36:45 -0700
Message-ID: <14cb8934af8.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net>
In-Reply-To: <87r3rolydh.fsf@toke.dk>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0 AquaMail/1.5.0.19 (build: 2100846)
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/TKSsq_z9hynf4TqA4KGe2Iqjki8>
Cc: aqm@ietf.org, Michael Welzl <michawe@ifi.uio.no>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 15:37:31 -0000

One problem with fair queueing is that it can be gamed. By opening multiple 
flows you achieve unfair priority. Part of Codel or PIE's beauty is that 
they are blind to the traffic, only reacting to the externally visible 
characteristics. This stems from the question 'what is a flow?'. There is 
no easy answer to this question, so Codel and PIE intentionally avoid the 
question.

Of course in many practical situations some simple assumptions about what a 
flow is do work, and in those situations fair queueing performs very well. 
It's important to keep in mind the limitations though. Fair queueing is not 
a panacea.

Simon

Sent with AquaMail for Android
http://www.aqua-mail.com


On April 13, 2015 10:41:37 AM Toke H�¸iland-J�¸rgensen <toke@toke.dk> wrote:

> "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com> writes:
>
> > CoDel's effort to limit the queuing delay appears well motivated when
> > there is only one queue (e.g., to shield VoIP from the delay induced
> > by bulk-transfer or video traffic), but when multiple queues are
> > available the net gain of the effort is unclear. I suspect the same
> > may happen with the flow-queue version of PIE.
>
> Yep, this has been my conclusion as well, and the test results on those
> slides are what convinced me of it: flow queueing is much more important
> than AQM, for most applications. However, note that in the tests I have
> run, the latency is measured by a separate flow; the latency of the
> packets making up the bulk flows is not measured. For most common
> traffic, this is fine, but there can be exceptions where you have a flow
> that has enough traffic to induce a queue, but which also requires low
> latency throughout. Adaptive video codecs for real-time media is a
> notable example. For those, having the AQM is nice.
>
> Also, having the AQM in place guards you against hash collisions. With
> the AQM in place, you degrade to its performance, whereas without,
> you're basically back to a large, dumb FIFO queue. fq_codel as it exists
> now is unfortunately not immune to those; we're working (in the
> Bufferbloat community) on an improved queueing discipline that solves
> this in most cases. See http://www.bufferbloat.net/projects/codel/wiki/Cake
>
> > Is this just because the set of tests in the slides does not include
> > one that clearly exposes the superiority of fq_codel over fq_nocodel?
> > And what would that test look like?
>
> Well, see above. Adaptive real-time video in particular is under-tested,
> in part due to a lack of good tools to generate traffic. Also, if you
> want to see fq_* break, try sticking all your flows in an encrypted VPN :)
>
> -Toke
>
> _______________________________________________
> aqm mailing list
> aqm@ietf.org
> https://www.ietf.org/mailman/listinfo/aqm



From nobody Tue Apr 14 08:40:08 2015
Return-Path: <koen.de_schepper@alcatel-lucent.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 9DADA1B2D26 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 08:40:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.909
X-Spam-Level: 
X-Spam-Status: No, score=-3.909 tagged_above=-999 required=5 tests=[BAYES_50=0.8, HTML_MESSAGE=0.001, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_HI=-5, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id P_GAVt6W0FGo for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 08:40:03 -0700 (PDT)
Received: from smtp-fr.alcatel-lucent.com (fr-hpida-esg-02.alcatel-lucent.com [135.245.210.21]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 94EC01B2CEA for <aqm@ietf.org>; Tue, 14 Apr 2015 08:40:02 -0700 (PDT)
Received: from fr712usmtp2.zeu.alcatel-lucent.com (unknown [135.239.2.42]) by Websense Email Security Gateway with ESMTPS id 8D6D2AACF286D; Tue, 14 Apr 2015 15:39:54 +0000 (GMT)
Received: from FR712WXCHHUB03.zeu.alcatel-lucent.com (fr712wxchhub03.zeu.alcatel-lucent.com [135.239.2.74]) by fr712usmtp2.zeu.alcatel-lucent.com (GMO) with ESMTP id t3EFdu1U008999 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Tue, 14 Apr 2015 17:39:56 +0200
Received: from FR711WXCHMBA05.zeu.alcatel-lucent.com ([169.254.1.213]) by FR712WXCHHUB03.zeu.alcatel-lucent.com ([135.239.2.74]) with mapi id 14.03.0195.001; Tue, 14 Apr 2015 17:39:56 +0200
From: "De Schepper, Koen (Koen)" <koen.de_schepper@alcatel-lucent.com>
To: Jim Gettys <jg@freedesktop.org>, =?utf-8?B?VG9rZSBIw7hpbGFuZC1Kw7hyZ2Vuc2Vu?= <toke@toke.dk>
Thread-Topic: [aqm] References on AQM test results and fq_nocodel
Thread-Index: AdB2ACpSZak1wModQSWAzPrkxbVugAAB6jHt///sNYCAACZRTf//74AA//6HN/A=
Date: Tue, 14 Apr 2015 15:39:56 +0000
Message-ID: <BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6@FR711WXCHMBA05.zeu.alcatel-lucent.com>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk> <CAGhGL2Cn4CyBe0ZpKK6Mfca-bHnYmSC3BaTx5FZX5=hD5qhGVw@mail.gmail.com>
In-Reply-To: <CAGhGL2Cn4CyBe0ZpKK6Mfca-bHnYmSC3BaTx5FZX5=hD5qhGVw@mail.gmail.com>
Accept-Language: nl-BE, en-US
Content-Language: en-US
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
x-originating-ip: [135.239.27.41]
Content-Type: multipart/alternative; boundary="_000_BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6FR711WXCHMBA05z_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/l-z7IZ0Me8eAwQVXSFA9AYcC_hg>
Cc: "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>, "aqm@ietf.org" <aqm@ietf.org>, Michael Welzl <michawe@ifi.uio.no>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 15:40:07 -0000

--_000_BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6FR711WXCHMBA05z_
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: base64
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==

--_000_BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6FR711WXCHMBA05z_
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: base64
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--_000_BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6FR711WXCHMBA05z_--


From nobody Tue Apr 14 08:57:24 2015
Return-Path: <toke@toke.dk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 5AE111B2D45 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 08:57:23 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.206
X-Spam-Level: *
X-Spam-Status: No, score=1.206 tagged_above=-999 required=5 tests=[BAYES_40=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_DK=1.009, MIME_8BIT_HEADER=0.3, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 61oF-R2lwUYF for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 08:57:22 -0700 (PDT)
Received: from mail2.tohojo.dk (mail2.tohojo.dk [77.235.48.147]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id CB3681B2D7E for <aqm@ietf.org>; Tue, 14 Apr 2015 08:57:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at mail2.tohojo.dk
Sender: toke@toke.dk
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toke.dk; s=201310; t=1429027033; bh=L0CkdFf6OJbTIJKTvox/iv/nj4LTZyij+ZnQGTUI28I=; h=From:To:Cc:Subject:References:Date:In-Reply-To; b=Y4SmcdvTsAS9tJJOfMGUeiwTYItvcWS9yG8FpKQePnWkHAwlK7LOlniBNxe2/WFZg VSi+8xR5lchqNCBdNWLGZkkhpNmqlH4aNwqTXsyw4u/wCE6FOl/djBhooxGKXDcEHH jdsL721vre40SXmUASU6w4ddfm39geotciV5YXkk=
Received: by alrua-karlstad.karlstad.toke.dk (Postfix, from userid 1000) id F183631AC9F; Tue, 14 Apr 2015 17:57:12 +0200 (CEST)
From: =?utf-8?Q?Toke_H=C3=B8iland-J=C3=B8rgensen?= <toke@toke.dk>
To: "De Schepper\, Koen \(Koen\)" <koen.de_schepper@alcatel-lucent.com>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk> <CAGhGL2Cn4CyBe0ZpKK6Mfca-bHnYmSC3BaTx5FZX5=hD5qhGVw@mail.gmail.com> <BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6@FR711WXCHMBA05.zeu.alcatel-lucent.com>
Date: Tue, 14 Apr 2015 17:57:12 +0200
In-Reply-To: <BF6B00CC65FD2D45A326E74492B2C19FB75C1ED6@FR711WXCHMBA05.zeu.alcatel-lucent.com> (De Schepper's message of "Tue, 14 Apr 2015 15:39:56 +0000")
X-Clacks-Overhead: GNU Terry Pratchett
Message-ID: <87egnmn1nr.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/i5e2K_k_Mj9XdmfU_w70m2Dr8Ow>
Cc: "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>, Jim Gettys <jg@freedesktop.org>, Michael Welzl <michawe@ifi.uio.no>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 15:57:23 -0000

"De Schepper, Koen (Koen)" <koen.de_schepper@alcatel-lucent.com> writes:

> Reporting the self inflicted delay on greedy flows might be a good
> measurement to additionally report in the test suites, so its impact
> on greedy interactive applications can be estimated.

Some sort of greedy real-time media or equivalent is definitely missing
from my suite of tests. However, that is partly due to the lack of a
good tool to do it with; preferably one that actually matches real
deployed applications (webrtc for instance), and that doesn't require
either synchronised clocks, packet dumps, or inspecting the router
queues while running.

On the page I linked to there are packet dumps available from my tests,
btw, in case anyone does want to go extract some of these figures from
them...

> I expect also that the advantage of FQ in general will becomes less
> relevant if the queue of all flows in the network are very small and
> even most of the time empty. This is what you see already when a low
> latency configured AQM is combined with FQ. The lower the latency in
> the queues the less the impact of FQ. FQ might still have an advantage
> in case of dynamics where a queue build up cannot be avoided.

Well, I haven't yet seen an AQM that can keep the queue "very small and
even most of the time empty" in the presence of (e.g.) the RRUL test.
Especially not during the period where the flows start up (or, roughly
equivalent, if the bandwidth changes very rapidly). FQ helps a lot here,
especially for the transients.

-Toke


From nobody Tue Apr 14 09:04:24 2015
Return-Path: <toke@toke.dk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id F41731B2DE4 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 09:04:22 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.206
X-Spam-Level: *
X-Spam-Status: No, score=1.206 tagged_above=-999 required=5 tests=[BAYES_40=-0.001, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_DK=1.009, MIME_8BIT_HEADER=0.3, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id EAPEHhzK-gX2 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 09:04:19 -0700 (PDT)
Received: from mail2.tohojo.dk (mail2.tohojo.dk [77.235.48.147]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 67D161B2E09 for <aqm@ietf.org>; Tue, 14 Apr 2015 09:01:51 -0700 (PDT)
X-Virus-Scanned: amavisd-new at mail2.tohojo.dk
Sender: toke@toke.dk
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toke.dk; s=201310; t=1429027304; bh=ayZf254jrlI+0SAFETJ+ZenpueP8dIX0UIh0nIqfmQg=; h=From:To:Cc:Subject:References:Date:In-Reply-To; b=oFK9rXFjnN9teqS+f48vh3xp1zJjA6ZnckzxIO4YM30nFfSEbetEXrdVGjGoaHUKd m60wI1FWm5kaG9KR3YIrTfG9uApz9vRN3HNm+4GmN05m5rFc+7k+cdXpN6/8Uzo/jA MQSHkxMBH4O4Jzd0fLHn2xCu8+uK+S1LhP7MMeMw=
Received: by alrua-karlstad.karlstad.toke.dk (Postfix, from userid 1000) id D05E231ACAF; Tue, 14 Apr 2015 18:01:43 +0200 (CEST)
From: =?utf-8?Q?Toke_H=C3=B8iland-J=C3=B8rgensen?= <toke@toke.dk>
To: Simon Barber <simon@superduper.net>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk> <14cb8934af8.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net>
Date: Tue, 14 Apr 2015 18:01:43 +0200
In-Reply-To: <14cb8934af8.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net> (Simon Barber's message of "Tue, 14 Apr 2015 08:36:45 -0700")
X-Clacks-Overhead: GNU Terry Pratchett
Message-ID: <87a8yan1g8.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/XkNqW1xisa7UcOrSZfgFa18K_6U>
Cc: Michael Welzl <michawe@ifi.uio.no>, aqm@ietf.org, "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 16:04:23 -0000

Simon Barber <simon@superduper.net> writes:

> One problem with fair queueing is that it can be gamed. By opening
> multiple flows you achieve unfair priority. Part of Codel or PIE's
> beauty is that they are blind to the traffic, only reacting to the
> externally visible characteristics. This stems from the question 'what
> is a flow?'. There is no easy answer to this question, so Codel and
> PIE intentionally avoid the question.

Sure, if we knew what the One True Queue Management Scheme that always
did the right thing was, we probably wouldn't be having this
conversation. :)

> Of course in many practical situations some simple assumptions about
> what a flow is do work, and in those situations fair queueing performs
> very well. It's important to keep in mind the limitations though. Fair
> queueing is not a panacea.

Well, my focus has primarily been "why does my internet suck so much and
what can I do to fix it". And for that (e.g. home type networks)
fq_codel is very close to a panacea.

I am well aware that there probably exists situations in which the
fq_codel assumptions do not hold up (and we do point out a couple in the
draft). However, I don't think I've ever seen someone actually
*demonstrate* (you know, with data) a scenario where fq_codel performs
worse than either Codel or PIE. If someone can point me to such a
demonstration I'll be happy to be proved wrong... :)

-Toke


From nobody Tue Apr 14 09:20:29 2015
Return-Path: <simon@superduper.net>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 65D2B1B2E3A for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 09:20:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 1.09
X-Spam-Level: *
X-Spam-Status: No, score=1.09 tagged_above=-999 required=5 tests=[BAYES_50=0.8, MIME_8BIT_HEADER=0.3, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Hghp8sqsBVO5 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 09:20:25 -0700 (PDT)
Received: from masada.superduper.net (masada.superduper.net [85.119.82.91]) (using TLSv1.2 with cipher DHE-RSA-AES128-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 872A51B2E37 for <aqm@ietf.org>; Tue, 14 Apr 2015 09:20:25 -0700 (PDT)
Received: from 199-116-72-167.public.monkeybrains.net ([199.116.72.167] helo=[192.168.0.12]) by masada.superduper.net with esmtpsa (TLS1.0:RSA_ARCFOUR_MD5:128) (Exim 4.80) (envelope-from <simon@superduper.net>) id 1Yi3ZK-0008Hb-Sp; Tue, 14 Apr 2015 17:20:17 +0100
From: Simon Barber <simon@superduper.net>
To: =?UTF-8?B?VG9rZSBIw7hpbGFuZC1Kw7hyZ2Vuc2Vu?= <toke@toke.dk>
Date: Tue, 14 Apr 2015 09:20:06 -0700
Message-ID: <14cb8baf738.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net>
In-Reply-To: <87a8yan1g8.fsf@toke.dk>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk> <14cb8934af8.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net> <87a8yan1g8.fsf@toke.dk>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.6.0 AquaMail/1.5.0.19 (build: 2100846)
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/3Z81L_RtfXdHodaduoDH915tXHg>
Cc: Michael Welzl <michawe@ifi.uio.no>, aqm@ietf.org, "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 16:20:27 -0000

Indeed fq_codel will work very well for many home users - especially where 
the home user can be expected to be aware of apps that might be gaming the 
system. If multiple homes are sharing a link it might not always work so well.

To construct an example where fair queueing works poorly, place a flow that 
is trying to game the system (e.g. a download performed by a download 
accelerator, launching multiple TCP streams for the single download) next 
to a bunch of flows that are opaque (e.g. Several VoIP calls through a VPN).

Simon

Sent with AquaMail for Android
http://www.aqua-mail.com


On April 14, 2015 9:01:55 AM Toke H�¸iland-J�¸rgensen <toke@toke.dk> wrote:

> Simon Barber <simon@superduper.net> writes:
>
> > One problem with fair queueing is that it can be gamed. By opening
> > multiple flows you achieve unfair priority. Part of Codel or PIE's
> > beauty is that they are blind to the traffic, only reacting to the
> > externally visible characteristics. This stems from the question 'what
> > is a flow?'. There is no easy answer to this question, so Codel and
> > PIE intentionally avoid the question.
>
> Sure, if we knew what the One True Queue Management Scheme that always
> did the right thing was, we probably wouldn't be having this
> conversation. :)
>
> > Of course in many practical situations some simple assumptions about
> > what a flow is do work, and in those situations fair queueing performs
> > very well. It's important to keep in mind the limitations though. Fair
> > queueing is not a panacea.
>
> Well, my focus has primarily been "why does my internet suck so much and
> what can I do to fix it". And for that (e.g. home type networks)
> fq_codel is very close to a panacea.
>
> I am well aware that there probably exists situations in which the
> fq_codel assumptions do not hold up (and we do point out a couple in the
> draft). However, I don't think I've ever seen someone actually
> *demonstrate* (you know, with data) a scenario where fq_codel performs
> worse than either Codel or PIE. If someone can point me to such a
> demonstration I'll be happy to be proved wrong... :)
>
> -Toke



From nobody Tue Apr 14 12:57:38 2015
Return-Path: <david@lang.hm>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id A7B101A1A58 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 12:57:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.61
X-Spam-Level: 
X-Spam-Status: No, score=-1.61 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id zX8VrI84neft for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 12:57:35 -0700 (PDT)
Received: from bifrost.lang.hm (mail.lang.hm [64.81.33.126]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 23C3A1A1A1D for <aqm@ietf.org>; Tue, 14 Apr 2015 12:57:34 -0700 (PDT)
Received: from asgard.lang.hm (asgard.lang.hm [10.0.0.100]) by bifrost.lang.hm (8.13.4/8.13.4/Debian-3) with ESMTP id t3EJvPtm012844; Tue, 14 Apr 2015 12:57:25 -0700
Date: Tue, 14 Apr 2015 12:57:25 -0700 (PDT)
From: David Lang <david@lang.hm>
X-X-Sender: dlang@asgard.lang.hm
To: =?ISO-8859-15?Q?Toke_H=F8iland-J=F8rgensen?= <toke@toke.dk>
In-Reply-To: <87r3rolydh.fsf@toke.dk>
Message-ID: <alpine.DEB.2.02.1504141254290.4306@nftneq.ynat.uz>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="680960-1636622003-1429041451=:4306"
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/qOIxcmhF9wD7Zq97t7W9qaP8MIo>
Cc: Michael Welzl <michawe@ifi.uio.no>, "aqm@ietf.org" <aqm@ietf.org>, "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 19:57:36 -0000

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--680960-1636622003-1429041451=:4306
Content-Type: TEXT/PLAIN; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 8BIT

On Mon, 13 Apr 2015, Toke Hĝiland-Jĝrgensen wrote:

> "Francini, Andrea (Andrea)" <andrea.francini@alcatel-lucent.com> writes:
>
>> CoDel's effort to limit the queuing delay appears well motivated when
>> there is only one queue (e.g., to shield VoIP from the delay induced
>> by bulk-transfer or video traffic), but when multiple queues are
>> available the net gain of the effort is unclear. I suspect the same
>> may happen with the flow-queue version of PIE.
>
> Yep, this has been my conclusion as well, and the test results on those
> slides are what convinced me of it: flow queueing is much more important
> than AQM, for most applications. However, note that in the tests I have
> run, the latency is measured by a separate flow; the latency of the
> packets making up the bulk flows is not measured. For most common
> traffic, this is fine, but there can be exceptions where you have a flow
> that has enough traffic to induce a queue, but which also requires low
> latency throughout. Adaptive video codecs for real-time media is a
> notable example. For those, having the AQM is nice.
>
> Also, having the AQM in place guards you against hash collisions. With
> the AQM in place, you degrade to its performance, whereas without,
> you're basically back to a large, dumb FIFO queue. fq_codel as it exists
> now is unfortunately not immune to those; we're working (in the
> Bufferbloat community) on an improved queueing discipline that solves
> this in most cases. See http://www.bufferbloat.net/projects/codel/wiki/Cake

do your finding still stand if you have enough different flows through the 
system that you end up with lots of flows (some of which may be bulk data) in 
each queue?

>> Is this just because the set of tests in the slides does not include
>> one that clearly exposes the superiority of fq_codel over fq_nocodel?
>> And what would that test look like?
>
> Well, see above. Adaptive real-time video in particular is under-tested,
> in part due to a lack of good tools to generate traffic. Also, if you
> want to see fq_* break, try sticking all your flows in an encrypted VPN :)

In this case, doesn't the router just see everything as a single flow? In which 
case there is nothing that it can do to affect priorities within that flow.

David Lang
--680960-1636622003-1429041451=:4306--


From nobody Tue Apr 14 13:20:37 2015
Return-Path: <david@lang.hm>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 98D9B1AD218 for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 13:20:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.289
X-Spam-Level: 
X-Spam-Status: No, score=0.289 tagged_above=-999 required=5 tests=[BAYES_40=-0.001, MIME_8BIT_HEADER=0.3, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ZlkziSATHK6s for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 13:20:32 -0700 (PDT)
Received: from bifrost.lang.hm (mail.lang.hm [64.81.33.126]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 497341AD272 for <aqm@ietf.org>; Tue, 14 Apr 2015 13:20:28 -0700 (PDT)
Received: from asgard.lang.hm (asgard.lang.hm [10.0.0.100]) by bifrost.lang.hm (8.13.4/8.13.4/Debian-3) with ESMTP id t3EKKQlp013006; Tue, 14 Apr 2015 13:20:26 -0700
Date: Tue, 14 Apr 2015 13:20:26 -0700 (PDT)
From: David Lang <david@lang.hm>
X-X-Sender: dlang@asgard.lang.hm
To: =?ISO-8859-15?Q?Toke_H=F8iland-J=F8rgensen?= <toke@toke.dk>
In-Reply-To: <87a8yan1g8.fsf@toke.dk>
Message-ID: <alpine.DEB.2.02.1504141313400.4306@nftneq.ynat.uz>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk> <14cb8934af8.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net> <87a8yan1g8.fsf@toke.dk>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="680960-1656474942-1429042826=:4306"
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/o8B_N9coijECmNJ_mMUqEPKADNo>
Cc: Simon Barber <simon@superduper.net>, aqm@ietf.org, Michael Welzl <michawe@ifi.uio.no>, "Francini, Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 14 Apr 2015 20:20:34 -0000

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--680960-1656474942-1429042826=:4306
Content-Type: TEXT/PLAIN; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 8BIT

On Tue, 14 Apr 2015, Toke Hĝiland-Jĝrgensen wrote:

> Simon Barber <simon@superduper.net> writes:
>
>> One problem with fair queueing is that it can be gamed. By opening
>> multiple flows you achieve unfair priority. Part of Codel or PIE's
>> beauty is that they are blind to the traffic, only reacting to the
>> externally visible characteristics. This stems from the question 'what
>> is a flow?'. There is no easy answer to this question, so Codel and
>> PIE intentionally avoid the question.
>
> Sure, if we knew what the One True Queue Management Scheme that always
> did the right thing was, we probably wouldn't be having this
> conversation. :)

if the fq portion is being gamed, how severe can the imbalance be? Is it a 
matter that if there are N flows without gaming the system, and each is getting 
1/N bandwith, then if a cheater uses M flows the cheater gets M/(N+M) of the 
bandwidth?

more importantly, is there any way short of a massive number of flows 
(essentially a DDOS), that this gaming of the system can hurt the latency of the 
other flows? or is it just the relative bandwidth of the different apps that 
suffers, and if an app is low bandwidth, but latency sensitive (i.e. VoIP), it 
won't be affected until it's "share" of bandwidth drops below the minimum it 
needs?

>> Of course in many practical situations some simple assumptions about
>> what a flow is do work, and in those situations fair queueing performs
>> very well. It's important to keep in mind the limitations though. Fair
>> queueing is not a panacea.
>
> Well, my focus has primarily been "why does my internet suck so much and
> what can I do to fix it". And for that (e.g. home type networks)
> fq_codel is very close to a panacea.
>
> I am well aware that there probably exists situations in which the
> fq_codel assumptions do not hold up (and we do point out a couple in the
> draft). However, I don't think I've ever seen someone actually
> *demonstrate* (you know, with data) a scenario where fq_codel performs
> worse than either Codel or PIE. If someone can point me to such a
> demonstration I'll be happy to be proved wrong... :)

Looping back to the start of the topic, the slides apparently are being read 
that fq without codel is just as good, or better than fq_codel, so codel can be 
dropped, and only fq is needed.

Is this what you are concluding from this? Or is that a misreading of something 
(the data, your work, or me misreading this thread)?

David Lang
--680960-1656474942-1429042826=:4306--


From nobody Tue Apr 14 21:16:35 2015
Return-Path: <garmitage@swin.edu.au>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id C3EC41B2BBB for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 21:16:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.806
X-Spam-Level: 
X-Spam-Status: No, score=-0.806 tagged_above=-999 required=5 tests=[BAYES_50=0.8, HELO_EQ_AU=0.377, HOST_EQ_AU=0.327, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id GhXVi18BE3YJ for <aqm@ietfa.amsl.com>; Tue, 14 Apr 2015 21:16:32 -0700 (PDT)
Received: from gpo1.cc.swin.edu.au (gpo1.cc.swin.edu.au [136.186.1.30]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 502101B2BBA for <aqm@ietf.org>; Tue, 14 Apr 2015 21:16:32 -0700 (PDT)
Received: from [136.186.229.37] (garmitage.caia.swin.edu.au [136.186.229.37]) by gpo1.cc.swin.edu.au (8.14.3/8.14.3) with ESMTP id t3F4GT5R023510 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO);  Wed, 15 Apr 2015 14:16:30 +1000
Message-ID: <552DE61D.5020308@swin.edu.au>
Date: Wed, 15 Apr 2015 14:16:29 +1000
From: grenville armitage <garmitage@swin.edu.au>
User-Agent: Mozilla/5.0 (X11; FreeBSD amd64; rv:16.0) Gecko/20121107 Thunderbird/16.0.2
MIME-Version: 1.0
To: "aqm@ietf.org" <aqm@ietf.org>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/OvRFhYdT8VmacxKeUD3sn0q0AMU>
Subject: [aqm] TEACUP -- automating TCP experiments on small testbeds
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2015 04:16:34 -0000

All,

Hard to know who might find this useful, but people doing TCP experiments might find the following tool useful:

http://caia.swin.edu.au/tools/teacup

We built TEACUP (TCP Experiment Automation Controlled Using Python) to automate many aspects of running TCP performance experiments in our small, specially-constructed physical testbed. TEACUP enables repeatable testing of different TCP algorithms over a range of emulated network path conditions, bottleneck rate limits and bottleneck queuing disciplines.

A key caveat: TEACUP assumes your physical testbed is a multi-host/single-bottleneck dumbbell-like topology with suitably configured end hosts and Linux-based bottleneck router. TEACUP does not try to run experiments over arbitrary network paths or the wider Internet. This has satisfied our use-cases, but YMMV :-)

We've released TEACUP v0.9 publicly in case it may be useful to other researchers who already have (or are interested in setting up) similar network testbeds.

cheers,
gja


From nobody Wed Apr 15 02:23:19 2015
Return-Path: <bob.briscoe@bt.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id F14EF1B33BF for <aqm@ietfa.amsl.com>; Wed, 15 Apr 2015 02:23:17 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.611
X-Spam-Level: 
X-Spam-Status: No, score=-0.611 tagged_above=-999 required=5 tests=[BAYES_05=-0.5, J_CHICKENPOX_44=0.6, RCVD_IN_DNSWL_LOW=-0.7, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id X-a7Jgdy4Hru for <aqm@ietfa.amsl.com>; Wed, 15 Apr 2015 02:23:13 -0700 (PDT)
Received: from hubrelay-by-04.bt.com (hubrelay-by-04.bt.com [62.7.242.140]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 638AC1B33BC for <aqm@ietf.org>; Wed, 15 Apr 2015 02:23:12 -0700 (PDT)
Received: from EVMHR71-UKRD.domain1.systemhost.net (10.36.3.109) by EVMHR04-UKBR.bt.com (10.216.161.36) with Microsoft SMTP Server (TLS) id 8.3.348.2; Wed, 15 Apr 2015 10:23:09 +0100
Received: from EPHR02-UKIP.domain1.systemhost.net (147.149.100.81) by EVMHR71-UKRD.domain1.systemhost.net (10.36.3.109) with Microsoft SMTP Server (TLS) id 8.3.348.2; Wed, 15 Apr 2015 10:23:09 +0100
Received: from bagheera.jungle.bt.co.uk (132.146.168.158) by EPHR02-UKIP.domain1.systemhost.net (147.149.100.81) with Microsoft SMTP Server id 14.3.181.6; Wed, 15 Apr 2015 10:23:10 +0100
Received: from BTP075694.jungle.bt.co.uk ([10.109.147.62])	by bagheera.jungle.bt.co.uk (8.13.5/8.12.8) with ESMTP id t3F9N4Ns019432;	Wed, 15 Apr 2015 10:23:05 +0100
Message-ID: <201504150923.t3F9N4Ns019432@bagheera.jungle.bt.co.uk>
X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9
Date: Wed, 15 Apr 2015 10:23:04 +0100
To: David Lang <david@lang.hm>
From: Bob Briscoe <bob.briscoe@bt.com>
In-Reply-To: <alpine.DEB.2.02.1504131441190.11469@nftneq.ynat.uz>
References: <23AFEFE3-4D93-4DD9-A22B-952C63DB9FE3@cisco.com> <BF6B00CC65FD2D45A326E74492B2C19FB75BAA82@FR711WXCHMBA05.zeu.alcatel-lucent.com> <72EE366B-05E6-454C-9E53-5054E6F9E3E3@ifi.uio.no> <55146DB9.7050501@rogers.com> <08C34E4A-DFB7-4816-92AE-2ED161799488@ifi.uio.no> <BF6B00CC65FD2D45A326E74492B2C19FB75BAFA0@FR711WXCHMBA05.zeu.alcatel-lucent.com> <alpine.DEB.2.02.1503271024550.2416@nftneq.ynat.uz> <5d58d2e21400449280173aa63069bf7a@hioexcmbx05-prd.hq.netapp.com> <20150327183659.GI39886@verdi> <72C12F6B-9DDE-4483-81F2-2D9A0F2D3A48@cs.columbia.edu> <alpine.DEB.2.02.1503271211200.19390@nftneq.ynat.uz> <D13AFCE7.46BC%kk@cs.ucr.edu> <alpine.DEB.2.02.1503271257230.19390@nftneq.ynat.uz> <AE342093-DE05-4D93-96DA-EB07E221F1D9@netapp.com> <alpine.DEB.2.02.1503291923300.26044@nftneq.ynat.uz> <201504131511.t3DFBG3R002270@bagheera.jungle.bt.co.uk> <alpine.DEB.2.02.1504131441190.11469@nftneq.ynat.uz>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-Scanned-By: MIMEDefang 2.56 on 132.146.168.158
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/66txslu9-8LxRnrUD7NQErVYr5Y>
Cc: "Scheffenegger, Richard" <rs@netapp.com>, Vishal Misra <misra@cs.columbia.edu>, KK <kk@cs.ucr.edu>, John Leslie <john@jlc.net>, "aqm@ietf.org" <aqm@ietf.org>
Subject: [aqm] Gaming ECN (again) (was:  think once to mark, think twice to drop:  draft-ietf-aqm-ecn-benefits-02)
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2015 09:23:18 -0000

David,

At 22:46 13/04/2015, David Lang wrote:
>On Mon, 13 Apr 2015, Bob Briscoe wrote:
>
>>David,
>>
>>Returning from a fortnight offlist...
>>
>>I think your conception of how ECN works is incorrect. You describe 
>>ECN as if the AQM marks one packet when it drops another packet. 
>>You say that the ECN-mark speeds up the retransmission of the 
>>dropped packet. On the contrary, the idea of classic ECN [RFC3168] 
>>is that the ECN marks replace the drops. In all known testing 
>>(except pathological cases), classic ECN effectively eliminates 
>>drops for all ECN-capable packets.
>
>That's what I thought, and if that was the case, then marking 
>packets as ECN-capable would mean that they would have an advantage 
>over non-ECN packets (by not getting dropped, so getting a higher 
>share of bandwidth)

This is a common fallacy. An ECN-capable TCP achieves the same 
throughput as an otherwise identical non-ECN TCP. The fallacy comes 
from people who think that the network caps the rate by removing 
packets. But the source determines the rate by how many drop or mark 
signals it sees. For today's ('classic') ECN, the source's rate 
reduction in response to either is the same; both voluntary as well.

There will be a tiny difference in goodput, because of the 
retransmissions. However, the loss (or marking) probability that TCP 
uses to determine its rate is a fraction of a percent, so this 
difference is in the noise.


>that's what the gaming ECN thread was about, and if I understood the 
>responses, I was being told that marking packets as ECN-capable, but 
>not slowing down (actually responding to ECN) would not let an 
>application get any advantage because the packets would just end up 
>getting dropped anyway, since marking and dropping happen at the 
>same level, even on ECN-capable flows.

Nope.

I recall that thread. People came up with a number of complicated 
arguments for why it is hard to game ECN, but none were solid. Below 
I have given a simple argument that I think is solid on its own. I 
thought about intevening at the time, but this stuff needs care and 
time that I didn't have then.

1) For the most simple and complete argument, all you need to know 
about the ECN behaviour of AQMs is: under normal load conditions, an 
AQM decides it's time to send a congestion signal irrespective of 
whether the next packet is ECN-capable or not. Then, if the next 
packet is ECN-capable, it marks it, else it drops it. This is from 
RFC3168, which also requires the source to respond equally to either 
a loss or a mark. I call this 'classic' ECN.{Note 1}

2) I will try to correct your misunderstanding about "marking and 
dropping at the same level even on on ECN-capable flows". However, to 
determine whether ECN can be gamed, there's no need to go there. So 
I'll come back to that as a post-script{Note 2}.

3) I will prove that it is as easy to game loss as it is to game ECN, 
first considering sender cheating, then receiver cheating:

3a) Sender Cheating
 From the sender's point of view, the only difference between a loss 
and an ECN mark is that it has to retransmit a loss. But that has 
nothing to do with the rate it can go at. If it has been programmed 
to ignore congestion feedback (and instead to go at a constant 
unresponsive rate{Note 2}), it is as easy for it to ignore loss 
feedback as ECN feedback. See {Note 3} for an example.

3b) Receiver Cheating
* An ECN receiver can best fool an ECN-capable TCP sender into going 
faster by only feeding back a small fraction of ECN marks.{Note 4}
* A non-ECN receiver could fool a non-ECN TCP sender into going 
faster by only revealing a small fraction of the losses. However, it 
would have to ACK undelivered bytes, and most TCP-based apps won't 
work unless all bytes are delivered.{Note 5}

So it seems that it's easier for a receiver to game ECN than loss. However:
* returning to the ECN case, the sender can validate the receiver by 
randomly setting an ECN mark itself on a very small proportion of 
packets (probably only on unusually high rate connections). Then if 
it doesn't see ECN feedback on the ACK of any one of its 
self-inserted marks, it can close the connection.

In summary,
* a sender can't game ECN any more easily than it can game loss.
* a receiver can only game ECN if the sender doesn't take measures to 
prevent it.{Note 6}

>If the packets are just marked, but not dropped, then the 
>ECN-capable flows will occupy a disproportinate share of the 
>available buffer space, since they just get marked instead of dropped.

Nope.

The arrival rates will be the same, whether or not ECN is used (see 
earlier). And recall  that TCP drives the marking or loss probability 
at very small fractions in all normal conditions.

Example: if there are 10 flows in a 100Mb/s link, 5 ECN and 5 
non-ECN, they will all arrive at the buffer at 10Mb/s (all other 
factors being equal). Then, if the loss or marking probability is 
0.5%, the AQM will be marking but not dropping 1 in 200 packets in 
the ECN flows whereas it would drop 1 in 200 from the non-ECN flows.

So, assuming tail drop, if there were 399 packets in this buffer, on 
average 200 would be ECN-capable (20 in each flow) with one marked; 
and 199 would be non-ECN-capable (20 in each flow except one with 
19). And one of those 199 would be a retransimssion from an earlier loss.

[Of course, we would hope that there would be 4 packets in the 
buffer, not 400. The proportions would still be the same on average. 
I merely used 399 to avoid fractions of packets for the averages.]


===Footnotes===

{Note 1} To be concrete, I've assumed classic ECN [RFC3168]. The 
argument is similar for research approaches like "think once to mark, 
twice to drop", but let's not make it more complicated than it needs to be.

{Note 2} RFC3168 (and draft-ietf-aqm-recommendation) require that, 
whenever the AQM decides it is time to signal congestion on the next 
packet, if the queue has been persistently long, the AQM must only 
use drop as a congestion signal, irrespective of whether the next 
packet is ECN-capable or not.

So, if a source naively just continues to increase its window until 
it drives the queue into overload, then it would cause the AQM to 
turn off ECN and consequently not be able to game ECN. But the simple 
strategy of sending at a high but constant rate avoids driving the 
queue into overload. So that's the strategy I described for gaming 
ECN. Because strategies that don't work are irrelevant if there's a 
strategy that does work.

{Note 3} Examples to show source cheating is as easy with loss as ECN:
* An ECN source sends at a constant unresponsive 90Mb/s through a 
100Mb/s bottleneck. In parallel some other responsive flows (say 10 
non-ECN TCP flows) squeeze themselves into the remaining 10Mb/s. They 
will cause themselves (say) 0.5% loss probability, while the 
unresponsive flow will experience 0.5% marking and zero loss.
* A non-ECN source can just as easily send unresponsively at 90.5Mb/s 
as 90Mb/s. The other flows will still drive loss to about 0.5%, which 
the unresponsive flow will now experience as well. Nonetheless, after 
it retransmits the 0.5% loss it still achieves goodput of about 90Mb/s.

{Note 4} Again, feeding back no marks at all would be naive, because 
it would drive the bottleneck into overload, causing it to turn off 
ECN (and driving the loss-rate over a cliff). A better strategy is to 
feedback only a small proportion. Because TCP's rate depends on the 
square root of the congestion probability, to download N times 
faster, the receiver should feed back only about 1 in N^2 of the 
marks or losses. E.g. to go 90 times faster, feed back 1 in 8100 
marks (or losses).

{Note 5} There are two classes of apps that use TCP but can get away 
without reliable delivery:
i) Some streaming media apps are designed with a loss-tolerant 
encoding, so they can use TCP but play out the media even if some 
retransmissions haven't arrived yet (e.g. using a raw socket at the receiver).
ii) In the specific case of HTTP, a hacked receiver can open another 
connection to the same server and download the byte-ranges it needs 
to repair the holes in the other connection.

{Note 6} ConEx (congestion exposure [RFC6789]) provides a 
comprehensive framework for the network to prevent senders and 
receivers from cheating. However, for this argument, we don't need to 
go there either.

Cheers



Bob



>David Lang

[snip]



>>ECN has potential cheating problems, but we have per-customer queues anyway.
>>Using flow as the unit of allocation also has its own problems, 
>>with no proposed solutions.

[snip]


>>Bob
>>
>>At 05:16 30/03/2015, David Lang wrote:

[snip]

>>>While AQM makes the network usable, there is still additional room 
>>>for improvement. While dropping packets does result in the TCP 
>>>senders slowing down,and eventually stabilizing at around the 
>>>right speed to keep the link fully utilized, the only way that 
>>>senders have been able to detect problems is to discover that they 
>>>have not received an ack for the traffic within the allowed time. 
>>>This causes a 'bubble' in the flow as teh dropped packet must be 
>>>retransmitted (and sometimes a significant amount of data after 
>>>the dropped packet that did make it to the destination, but could 
>>>not be acked because fo the missing packet).
>>>This "bubble" in the data flow can be greatly compressed by 
>>>configuring the AQM algorithm to send an ECN packet to the sender 
>>>when it drops a packet in a flow. The sender can then adapt 
>>>faster, slowing down it's new data, and re-sending the dropped 
>>>packet without having to wait for the timeout. This has two major 
>>>effects by allowing the sender to retransmit the packet sooner the 
>>>dealy on the dropped data is not as long, and because the 
>>>replacement data can arrive before the timeout of the following 
>>>packets, they may not need to be re-sent. by configuring the AQM 
>>>algorithm to send the ECN notification to the sender only when the 
>>>packet is being dropped, the effect of failure of the ECN packet 
>>>to get through to the sender (the notification packet runs into 
>>>congestion and gets dropped, some network device blocks it, etc) 
>>>is that the ECN enabled case devolves to match the non-ECN case in 
>>>that the sender will still detect the dropped packet via the 
>>>timeout waiting for the ack as if ENCN was not enabled.
>>><insert link to possible problems that can happen here, including 
>>>the potential for an app to 'game' things if packets are marked at 
>>>a different level than when they are dropped.>
>>>
>>>So a very strong recommendation to enable Active Queue Management, 
>>>while the different algorithms have different advantages and 
>>>levels of testing, even the 'worst' of the set results in a 
>>>night-and-day improvement for usability compared to unmanaged buffers.
>>>Enabling ECN at the same point as dropping packets as part of 
>>>enabling any AQM algorithm results in a noticable improvement over 
>>>the base algorithm without ECN. When compared to the baseline, the 
>>>improvement added by ECN is tiny compared to the improvement from enabling AQM.
>>>
>>>   Is it fair to say that plain aqm vs aqm+ecn variation is on the 
>>> same order of difference as the differences between the different 
>>> AQM algorithms?
>>>
>>>Future research items (which others here may already have done, 
>>>and would not be part of my 'elevator pitch')
>>>I believe that currently ECn triggers the exact same slowdown that 
>>>a missed packet does, and it may be appropriate to have the sender 
>>>do a less drastic slowdown.
>>>It would be very interesing to provide soem way for the 
>>>application sending the traffic to detect dropped packets and ECN 
>>>responses. For example, a streaming media source (especially an 
>>>interactive one like video conferencing) could adjust the bitrate 
>>>that it's sending.
>>>David Lang
>>>_______________________________________________
>>>aqm mailing list
>>>aqm@ietf.org
>>>https://www.ietf.org/mailman/listinfo/aqm
>>
>>________________________________________________________________
>>Bob Briscoe,                                                  BT
>
>________________________________________________________________
>Bob Briscoe,                                                  BT 


From nobody Wed Apr 15 13:16:33 2015
Return-Path: <david@lang.hm>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 291211A1EFD for <aqm@ietfa.amsl.com>; Wed, 15 Apr 2015 13:16:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.31
X-Spam-Level: 
X-Spam-Status: No, score=-1.31 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, J_CHICKENPOX_44=0.6, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id eNIyexIXc2Zb for <aqm@ietfa.amsl.com>; Wed, 15 Apr 2015 13:16:30 -0700 (PDT)
Received: from bifrost.lang.hm (mail.lang.hm [64.81.33.126]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 2B0AB1A2130 for <aqm@ietf.org>; Wed, 15 Apr 2015 13:16:26 -0700 (PDT)
Received: from asgard.lang.hm (asgard.lang.hm [10.0.0.100]) by bifrost.lang.hm (8.13.4/8.13.4/Debian-3) with ESMTP id t3FKFLci020556; Wed, 15 Apr 2015 13:15:21 -0700
Date: Wed, 15 Apr 2015 13:15:21 -0700 (PDT)
From: David Lang <david@lang.hm>
X-X-Sender: dlang@asgard.lang.hm
To: Bob Briscoe <bob.briscoe@bt.com>
In-Reply-To: <201504150923.t3F9N4Ns019432@bagheera.jungle.bt.co.uk>
Message-ID: <alpine.DEB.2.02.1504151259500.26320@nftneq.ynat.uz>
References: <23AFEFE3-4D93-4DD9-A22B-952C63DB9FE3@cisco.com> <BF6B00CC65FD2D45A326E74492B2C19FB75BAA82@FR711WXCHMBA05.zeu.alcatel-lucent.com> <72EE366B-05E6-454C-9E53-5054E6F9E3E3@ifi.uio.no> <55146DB9.7050501@rogers.com> <08C34E4A-DFB7-4816-92AE-2ED161799488@ifi.uio.no> <BF6B00CC65FD2D45A326E74492B2C19FB75BAFA0@FR711WXCHMBA05.zeu.alcatel-lucent.com> <alpine.DEB.2.02.1503271024550.2416@nftneq.ynat.uz> <5d58d2e21400449280173aa63069bf7a@hioexcmbx05-prd.hq.netapp.com> <20150327183659.GI39886@verdi> <72C12F6B-9DDE-4483-81F2-2D9A0F2D3A48@cs.columbia.edu> <alpine.DEB.2.02.1503271211200.19390@nftneq.ynat.uz> <D13AFCE7.46BC%kk@cs.ucr.edu> <alpine.DEB.2.02.1503271257230.19390@nftneq.ynat.uz> <AE342093-DE05-4D93-96DA-EB07E221F1D9@netapp.com> <alpine.DEB.2.02.1503291923300.26044@nftneq.ynat.uz> <201504131511.t3DFBG3R002270@bagheera.jungle.bt.co.uk> <alpine.DEB.2.02.1504131441190.11469@nftneq.ynat.uz> <201504150923.t3F9N4Ns019432@bagheera.jungle.bt.co.uk>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/GDFYUMXBGJZuJbZVNqQzccnSmYw>
Cc: "Scheffenegger, Richard" <rs@netapp.com>, Vishal Misra <misra@cs.columbia.edu>, KK <kk@cs.ucr.edu>, John Leslie <john@jlc.net>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] Gaming ECN (again) (was:  think once to mark, think twice to drop:  draft-ietf-aqm-ecn-benefits-02)
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2015 20:16:32 -0000

On Wed, 15 Apr 2015, Bob Briscoe wrote:

> 3) I will prove that it is as easy to game loss as it is to game ECN, first 
> considering sender cheating, then receiver cheating:

I think there is a key difference

cheating non-ECN requires changes to the TCP stack code

cheating ECN requires an iptables rule to zero out the ECN bit on received 
packets before they hit the TCP stack.

I've seen enough people doing cargo-cult network configs to 'improve 
gaming' (MTU 576 as one example) that I expect that if there is any noticable 
advantage for doing so, there will be a significant population doing so.

> 3a) Sender Cheating
> From the sender's point of view, the only difference between a loss and an 
> ECN mark is that it has to retransmit a loss. But that has nothing to do with 
> the rate it can go at. If it has been programmed to ignore congestion 
> feedback (and instead to go at a constant unresponsive rate{Note 2}), it is 
> as easy for it to ignore loss feedback as ECN feedback. See {Note 3} for an 
> example.
>
> 3b) Receiver Cheating
> * An ECN receiver can best fool an ECN-capable TCP sender into going faster 
> by only feeding back a small fraction of ECN marks.{Note 4}
> * A non-ECN receiver could fool a non-ECN TCP sender into going faster by 
> only revealing a small fraction of the losses. However, it would have to ACK 
> undelivered bytes, and most TCP-based apps won't work unless all bytes are 
> delivered.{Note 5}
>
> So it seems that it's easier for a receiver to game ECN than loss. However:
> * returning to the ECN case, the sender can validate the receiver by randomly 
> setting an ECN mark itself on a very small proportion of packets (probably 
> only on unusually high rate connections). Then if it doesn't see ECN feedback 
> on the ACK of any one of its self-inserted marks, it can close the 
> connection.

do any TCP stacks do this?

won't this cause the datarate to slow down, something that will seem 'bad' for 
the people who would need to code this (it will show up on benchmarks as random 
slowdowns, right?)

> In summary,
> * a sender can't game ECN any more easily than it can game loss.

I think it's more "there is a way to detect gaming ECN"

> * a receiver can only game ECN if the sender doesn't take measures to prevent 
> it.{Note 6}
>
>> If the packets are just marked, but not dropped, then the ECN-capable flows 
>> will occupy a disproportinate share of the available buffer space, since 
>> they just get marked instead of dropped.
>
> Nope.
>
> The arrival rates will be the same, whether or not ECN is used (see earlier). 
> And recall  that TCP drives the marking or loss probability at very small 
> fractions in all normal conditions.
>
> Example: if there are 10 flows in a 100Mb/s link, 5 ECN and 5 non-ECN, they 
> will all arrive at the buffer at 10Mb/s (all other factors being equal). 
> Then, if the loss or marking probability is 0.5%, the AQM will be marking but 
> not dropping 1 in 200 packets in the ECN flows whereas it would drop 1 in 200 
> from the non-ECN flows.
>
> So, assuming tail drop, if there were 399 packets in this buffer, on average 
> 200 would be ECN-capable (20 in each flow) with one marked; and 199 would be 
> non-ECN-capable (20 in each flow except one with 19). And one of those 199 
> would be a retransimssion from an earlier loss.
>
> [Of course, we would hope that there would be 4 packets in the buffer, not 
> 400. The proportions would still be the same on average. I merely used 399 to 
> avoid fractions of packets for the averages.]
>
>
> ===Footnotes===
>
> {Note 3} Examples to show source cheating is as easy with loss as ECN:
> * An ECN source sends at a constant unresponsive 90Mb/s through a 100Mb/s 
> bottleneck. In parallel some other responsive flows (say 10 non-ECN TCP 
> flows) squeeze themselves into the remaining 10Mb/s. They will cause 
> themselves (say) 0.5% loss probability, while the unresponsive flow will 
> experience 0.5% marking and zero loss.
> * A non-ECN source can just as easily send unresponsively at 90.5Mb/s as 
> 90Mb/s. The other flows will still drive loss to about 0.5%, which the 
> unresponsive flow will now experience as well. Nonetheless, after it 
> retransmits the 0.5% loss it still achieves goodput of about 90Mb/s.

well, sort of.

if the retransmissions are due to dropped packets, it may require retransmitting 
some packets that actually got through (depending on the link latency and rate 
of transmission, those added packets may not be acked before they also get 
retransmitted), this will waste some bandwidth

if the sender slows down, it may drop significantly below the 90Mb/s rate and 
have to ramp up to that rate again.

> {Note 4} Again, feeding back no marks at all would be naive, because it would 
> drive the bottleneck into overload, causing it to turn off ECN (and driving 
> the loss-rate over a cliff). A better strategy is to feedback only a small 
> proportion. Because TCP's rate depends on the square root of the congestion 
> probability, to download N times faster, the receiver should feed back only 
> about 1 in N^2 of the marks or losses. E.g. to go 90 times faster, feed back 
> 1 in 8100 marks (or losses).

that assumes that the raw sending rate is enough higher to drive the link into 
overload itself rather than just being enough higher to demand an unfair share 
of the link

you use the example of a 100Mb link with one app getting 90Mb and trying to get 
more.

Instead think of that same 100Mb link, but with 100 1Mb flows going over it. 
Would ignoring ECN feedback allow one app to get a 50Mb flow, squeezing the 
other apps to .5Mb each?

> {Note 5} There are two classes of apps that use TCP but can get away without 
> reliable delivery:
> i) Some streaming media apps are designed with a loss-tolerant encoding, so 
> they can use TCP but play out the media even if some retransmissions haven't 
> arrived yet (e.g. using a raw socket at the receiver).
> ii) In the specific case of HTTP, a hacked receiver can open another 
> connection to the same server and download the byte-ranges it needs to repair 
> the holes in the other connection.

the TCP stack will demand reliable delivery, even if the app doesn't. There 
isn't any provision for an app to use TCP but allow 'holes' in the data.

David Lang


From nobody Fri Apr 17 06:53:27 2015
Return-Path: <toke@toke.dk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id CE7751B2D05 for <aqm@ietfa.amsl.com>; Fri, 17 Apr 2015 06:53:25 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 2.007
X-Spam-Level: **
X-Spam-Status: No, score=2.007 tagged_above=-999 required=5 tests=[BAYES_50=0.8, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_DK=1.009, MIME_8BIT_HEADER=0.3, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id teSsmg8Ejrwm for <aqm@ietfa.amsl.com>; Fri, 17 Apr 2015 06:53:24 -0700 (PDT)
Received: from mail2.tohojo.dk (mail2.tohojo.dk [77.235.48.147]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 7DE9E1B2D02 for <aqm@ietf.org>; Fri, 17 Apr 2015 06:53:24 -0700 (PDT)
X-Virus-Scanned: amavisd-new at mail2.tohojo.dk
Sender: toke@toke.dk
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=toke.dk; s=201310; t=1429278796; bh=ZzjnTpf7wW0gwfNRMi385wVzVtYAU26XSxtouHJe524=; h=From:To:Cc:Subject:References:Date; b=Qz/hte7xEsGJVIOn/J5kPYiLCdC/+4S12e+sgJWJ3sngXLkEm33c4f0yXWOXBFMbu HgO3hfd2ZFoims/3rFOq4FnHle5yun1uEeYaNNxvN6PX+CvHxbcmS8oE0h18Af1Jbb 7+uJzs/vz5zgBAkpOzV0SfqqcvMWcIrF+Nks8+ms=
Received: by alrua-kau.kau.toke.dk (Postfix, from userid 1000) id A61A7C402B9; Fri, 17 Apr 2015 15:53:15 +0200 (CEST)
From: =?utf-8?Q?Toke_H=C3=B8iland-J=C3=B8rgensen?= <toke@toke.dk>
To: David Lang <david@lang.hm>
References: <1BFAC0A1D7955144A2444E902CB628F85E092996@US70TWXCHMBA11.zam.alcatel-lucent.com> <87vbh010tn.fsf@toke.dk> <1BFAC0A1D7955144A2444E902CB628F85E093A39@US70TWXCHMBA11.zam.alcatel-lucent.com> <87r3rolydh.fsf@toke.dk> <14cb8934af8.27f7.e972a4f4d859b00521b2b659602cb2f9@superduper.net> <87a8yan1g8.fsf@toke.dk> <alpine.DEB.2.02.1504141313400.4306@nftneq.ynat.uz>
Date: Fri, 17 Apr 2015 15:53:15 +0200
X-Clacks-Overhead: GNU Terry Pratchett
Message-ID: <87383y6eus.fsf@toke.dk>
MIME-Version: 1.0
Content-Type: text/plain
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/9oOevUJoLwfSQNAX7lIZOy2sL9Q>
Cc: Simon Barber <simon@superduper.net>, aqm@ietf.org, Michael Welzl <michawe@ifi.uio.no>, "Francini,  Andrea \(Andrea\)" <andrea.francini@alcatel-lucent.com>
Subject: Re: [aqm] References on AQM test results and fq_nocodel
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 17 Apr 2015 13:53:26 -0000

David Lang <david@lang.hm> writes:

> if the fq portion is being gamed, how severe can the imbalance be? Is
> it a matter that if there are N flows without gaming the system, and
> each is getting 1/N bandwith, then if a cheater uses M flows the
> cheater gets M/(N+M) of the bandwidth?

Yes, not counting hash collisions, that would be the case.

> more importantly, is there any way short of a massive number of flows
> (essentially a DDOS), that this gaming of the system can hurt the
> latency of the other flows? or is it just the relative bandwidth of
> the different apps that suffers, and if an app is low bandwidth, but
> latency sensitive (i.e. VoIP), it won't be affected until it's "share"
> of bandwidth drops below the minimum it needs?

Well, since fq_codel is hash based, if you manage to hit the same bucket
as the VoIP flow, that VoIP flow is not going to get the nice implicit
priority it would otherwise from the flow queueing. So yeah, if you,
say, send out 1000 packets at once on different UDP ports, then the VoIP
flow is going to notice. But in this case you're basically DOS'ing the
router; and FIFO queueing and pure AQMs fare pretty badly as well...

> Looping back to the start of the topic, the slides apparently are
> being read that fq without codel is just as good, or better than
> fq_codel, so codel can be dropped, and only fq is needed.
>
> Is this what you are concluding from this? Or is that a misreading of
> something (the data, your work, or me misreading this thread)?

Well, *for the tested scenarios and metrics* that is true. However,
there can be other scenarios where you will benefit from having AQM on
each of the queues. Several such hypothetical scenarios have been
mentioned in this thread. The open question, as far as I'm concerned, is
*how much* a difference it makes. I.e. exactly how much the performance
degrades if it is absent, and in which scenarios. As I said before, I
don't think anyone has ever actually tried it and produced data...

-Toke


From nobody Thu Apr 23 04:19:31 2015
Return-Path: <mirja.kuehlewind@tik.ee.ethz.ch>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id B6F9D1B2D72 for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 04:19:29 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.91
X-Spam-Level: 
X-Spam-Status: No, score=-3.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id xPudNp9rS_rY for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 04:19:27 -0700 (PDT)
Received: from smtp.ee.ethz.ch (smtp.ee.ethz.ch [129.132.2.219]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 8E55B1B2D5E for <aqm@ietf.org>; Thu, 23 Apr 2015 04:19:27 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by smtp.ee.ethz.ch (Postfix) with ESMTP id 1F3ABD930A; Thu, 23 Apr 2015 13:19:26 +0200 (MEST)
X-Virus-Scanned: by amavisd-new on smtp.ee.ethz.ch
Received: from smtp.ee.ethz.ch ([127.0.0.1]) by localhost (.ee.ethz.ch [127.0.0.1]) (amavisd-new, port 10024) with LMTP id vHP-An+D--qm; Thu, 23 Apr 2015 13:19:25 +0200 (MEST)
Received: from [82.130.103.143] (nb-10510.ethz.ch [82.130.103.143]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: mirjak) by smtp.ee.ethz.ch (Postfix) with ESMTPSA id E134BD9303; Thu, 23 Apr 2015 13:19:25 +0200 (MEST)
Message-ID: <5538D53D.1010908@tik.ee.ethz.ch>
Date: Thu, 23 Apr 2015 13:19:25 +0200
From: =?UTF-8?B?TWlyamEgS8O8aGxld2luZA==?= <mirja.kuehlewind@tik.ee.ethz.ch>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: aqm@ietf.org, fred@cisco.com, Gorry Fairhurst <gorry@erg.abdn.ac.uk>,  "Scheffenegger, Richard" <rs@netapp.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/7Q8D6o_3kR1o5mcE0BRZje3GOKA>
Subject: [aqm] AQM and ECN in draft-ietf-aqm-recommendation-11
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 23 Apr 2015 11:19:29 -0000

Hi all,

I know I'm too late for this but want to bring it up anyway:

draft-ietf-aqm-recommendation-11 currently says:
"An AQM algorithm that supports ECN needs to define the threshold and
    algorithm for ECN-marking.  This threshold MAY differ from that used
    for dropping packets that are not marked as ECN-capable, and SHOULD
    be configurable."

But I think it should be:
"An AQM algorithm that supports ECN needs to define the parameters and
    algorithm for ECN-marking.  These parameters MAY differ from that used
    for dropping packets that are not marked as ECN-capable, and SHOULD
    be configurable."

so just s/threshold/parameters/  ...?

I've just caught this now because I'm reading draft-welzl-ecn-benefits-02; 
sorry. Maybe it's still possible to make such a small edit... if other people 
think as well that this should be changed...?

Mirja



From nobody Thu Apr 23 04:57:17 2015
Return-Path: <mirja.kuehlewind@tik.ee.ethz.ch>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 5B3DE1B2FED for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 04:57:16 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.91
X-Spam-Level: 
X-Spam-Status: No, score=-3.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id mfiyGf1bDpY6 for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 04:57:14 -0700 (PDT)
Received: from smtp.ee.ethz.ch (smtp.ee.ethz.ch [129.132.2.219]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 7EF881B2FE9 for <aqm@ietf.org>; Thu, 23 Apr 2015 04:57:02 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by smtp.ee.ethz.ch (Postfix) with ESMTP id B7172D931C; Thu, 23 Apr 2015 13:57:00 +0200 (MEST)
X-Virus-Scanned: by amavisd-new on smtp.ee.ethz.ch
Received: from smtp.ee.ethz.ch ([127.0.0.1]) by localhost (.ee.ethz.ch [127.0.0.1]) (amavisd-new, port 10024) with LMTP id ompphCQchC-6; Thu, 23 Apr 2015 13:57:00 +0200 (MEST)
Received: from [82.130.103.143] (nb-10510.ethz.ch [82.130.103.143]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: mirjak) by smtp.ee.ethz.ch (Postfix) with ESMTPSA id E99DAD937C; Thu, 23 Apr 2015 13:56:59 +0200 (MEST)
Message-ID: <5538DE0B.8070304@tik.ee.ethz.ch>
Date: Thu, 23 Apr 2015 13:56:59 +0200
From: =?windows-1252?Q?Mirja_K=FChlewind?= <mirja.kuehlewind@tik.ee.ethz.ch>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: aqm@ietf.org, fred@cisco.com, Gorry Fairhurst <gorry@erg.abdn.ac.uk>,  "Scheffenegger, Richard" <rs@netapp.com>
References: <5538D53D.1010908@tik.ee.ethz.ch>
In-Reply-To: <5538D53D.1010908@tik.ee.ethz.ch>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/4kUhlyvEEgNDOJ71YcxCMGw7Yb0>
Subject: Re: [aqm] AQM and ECN in draft-ietf-aqm-recommendation-11
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 23 Apr 2015 11:57:16 -0000

Hi again,

I still know I'm late but there are two more points:

draft-welzl-ecn-benefits-02 says:
"Network devices must not drop
    packets solely because these codepoints are used [RFC2309.bis]"
and
"A network device should therefore not
    remark an ECT(0) or ECT(1) mark to zero [RFC2309.bis]."

I could not find this recommendation in draft-ietf-aqm-recommendation-11, at 
least not in the section 4.2.1 (AQM and ECN), but it should be in there (and 
should actually say: a network device MUST NOT remark, ECT(0), ETC(1) or CE to 
zero)...?

Mirja





On 23.04.2015 13:19, Mirja Kühlewind wrote:
> Hi all,
>
> I know I'm too late for this but want to bring it up anyway:
>
> draft-ietf-aqm-recommendation-11 currently says:
> "An AQM algorithm that supports ECN needs to define the threshold and
>      algorithm for ECN-marking.  This threshold MAY differ from that used
>      for dropping packets that are not marked as ECN-capable, and SHOULD
>      be configurable."
>
> But I think it should be:
> "An AQM algorithm that supports ECN needs to define the parameters and
>      algorithm for ECN-marking.  These parameters MAY differ from that used
>      for dropping packets that are not marked as ECN-capable, and SHOULD
>      be configurable."
>
> so just s/threshold/parameters/  ...?
>
> I've just caught this now because I'm reading draft-welzl-ecn-benefits-02;
> sorry. Maybe it's still possible to make such a small edit... if other people
> think as well that this should be changed...?
>
> Mirja
>
>
> _______________________________________________
> aqm mailing list
> aqm@ietf.org
> https://www.ietf.org/mailman/listinfo/aqm
>

-- 
------------------------------------------
Dipl.-Ing. Mirja Kühlewind
Communication Systems Group
Institute TIK, ETH Zürich
Gloriastrasse 35, 8092 Zürich, Switzerland

Room ETZ G93
phone: +41 44 63 26932
email: mirja.kuehlewind@tik.ee.ethz.ch
------------------------------------------


From nobody Thu Apr 23 07:17:09 2015
Return-Path: <mirja.kuehlewind@tik.ee.ethz.ch>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id BDD2D1A8840 for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 07:17:07 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.91
X-Spam-Level: 
X-Spam-Status: No, score=-3.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RpWz6y4P14gb for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 07:17:05 -0700 (PDT)
Received: from smtp.ee.ethz.ch (smtp.ee.ethz.ch [129.132.2.219]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id A28A91A8A7A for <aqm@ietf.org>; Thu, 23 Apr 2015 07:15:49 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by smtp.ee.ethz.ch (Postfix) with ESMTP id 06442D930E; Thu, 23 Apr 2015 16:15:48 +0200 (MEST)
X-Virus-Scanned: by amavisd-new on smtp.ee.ethz.ch
Received: from smtp.ee.ethz.ch ([127.0.0.1]) by localhost (.ee.ethz.ch [127.0.0.1]) (amavisd-new, port 10024) with LMTP id fZp98RSpRDaP; Thu, 23 Apr 2015 16:15:47 +0200 (MEST)
Received: from [82.130.103.143] (nb-10510.ethz.ch [82.130.103.143]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: mirjak) by smtp.ee.ethz.ch (Postfix) with ESMTPSA id B8D3FD9303; Thu, 23 Apr 2015 16:15:47 +0200 (MEST)
Message-ID: <5538FE93.2090501@tik.ee.ethz.ch>
Date: Thu, 23 Apr 2015 16:15:47 +0200
From: =?windows-1252?Q?Mirja_K=FChlewind?= <mirja.kuehlewind@tik.ee.ethz.ch>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: aqm@ietf.org, fred@cisco.com, Gorry Fairhurst <gorry@erg.abdn.ac.uk>,  "Scheffenegger, Richard" <rs@netapp.com>
References: <5538D53D.1010908@tik.ee.ethz.ch> <5538DE0B.8070304@tik.ee.ethz.ch>
In-Reply-To: <5538DE0B.8070304@tik.ee.ethz.ch>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/v6IpJ3wrNQRD6ri2DbNDUMxKk0U>
Subject: Re: [aqm] AQM and ECN in draft-ietf-aqm-recommendation-11
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 23 Apr 2015 14:17:07 -0000

I meant draft-ietf-aqm-ecn-benefits-03 (not draft-welzl-ecn-benefits-02) though...

On 23.04.2015 13:56, Mirja Kühlewind wrote:
> Hi again,
>
> I still know I'm late but there are two more points:
>
> draft-welzl-ecn-benefits-02 says:
> "Network devices must not drop
>      packets solely because these codepoints are used [RFC2309.bis]"
> and
> "A network device should therefore not
>      remark an ECT(0) or ECT(1) mark to zero [RFC2309.bis]."
>
> I could not find this recommendation in draft-ietf-aqm-recommendation-11, at
> least not in the section 4.2.1 (AQM and ECN), but it should be in there (and
> should actually say: a network device MUST NOT remark, ECT(0), ETC(1) or CE to
> zero)...?
>
> Mirja
>
>
>
>
>
> On 23.04.2015 13:19, Mirja Kühlewind wrote:
>> Hi all,
>>
>> I know I'm too late for this but want to bring it up anyway:
>>
>> draft-ietf-aqm-recommendation-11 currently says:
>> "An AQM algorithm that supports ECN needs to define the threshold and
>>       algorithm for ECN-marking.  This threshold MAY differ from that used
>>       for dropping packets that are not marked as ECN-capable, and SHOULD
>>       be configurable."
>>
>> But I think it should be:
>> "An AQM algorithm that supports ECN needs to define the parameters and
>>       algorithm for ECN-marking.  These parameters MAY differ from that used
>>       for dropping packets that are not marked as ECN-capable, and SHOULD
>>       be configurable."
>>
>> so just s/threshold/parameters/  ...?
>>
>> I've just caught this now because I'm reading draft-welzl-ecn-benefits-02;
>> sorry. Maybe it's still possible to make such a small edit... if other people
>> think as well that this should be changed...?
>>
>> Mirja
>>
>>
>> _______________________________________________
>> aqm mailing list
>> aqm@ietf.org
>> https://www.ietf.org/mailman/listinfo/aqm
>>
>

-- 
------------------------------------------
Dipl.-Ing. Mirja Kühlewind
Communication Systems Group
Institute TIK, ETH Zürich
Gloriastrasse 35, 8092 Zürich, Switzerland

Room ETZ G93
phone: +41 44 63 26932
email: mirja.kuehlewind@tik.ee.ethz.ch
------------------------------------------


From nobody Thu Apr 23 08:07:43 2015
Return-Path: <fred@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 4ADF11A92FE for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 08:07:43 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -114.211
X-Spam-Level: 
X-Spam-Status: No, score=-114.211 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5, USER_IN_WHITELIST=-100] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id CQy1sEpP-WAa for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 08:07:41 -0700 (PDT)
Received: from rcdn-iport-5.cisco.com (rcdn-iport-5.cisco.com [173.37.86.76]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id A846B1AC3B7 for <aqm@ietf.org>; Thu, 23 Apr 2015 08:07:14 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=2353; q=dns/txt; s=iport; t=1429801635; x=1431011235; h=from:to:cc:subject:date:message-id:references: in-reply-to:mime-version; bh=2C4DVSDSCU2COZwq6ovabKdPTAP8zPqgHe0L3UilQJs=; b=Nw5RFlB3+EIL0IgTYUbrGHB48UNmhxfDzMDG3WNzU7N802gQXd1xYe+B TK3oiVZHPbohJstworCB1jQuKVOqUmVS78KQeBO/HYysA4z7EZz8fXRXa Djo7OGBaRGWJgMP89H4O9F5+zKpGuPI4cuL+5/QLuBdziSlVGK21v57aP I=;
X-Files: signature.asc : 487
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0DvBgC1CTlV/5JdJa1bDoJ+gS4FgxXJVoESAoE2TAEBAQEBAYELhCABAQEDASNWBQsCAQgYKgICMiUCBAENBQ6IFQi3c5RjAQEBAQEBAQEBAQEBAQEBAQEBAQEBF4s3hQQHgmgvgRYFkUqBcoE3hwyVSyNgglc9bwGBQ4EAAQEB
X-IronPort-AV: E=Sophos;i="5.11,631,1422921600";  d="asc'?scan'208";a="414201554"
Received: from rcdn-core-10.cisco.com ([173.37.93.146]) by rcdn-iport-5.cisco.com with ESMTP; 23 Apr 2015 15:07:14 +0000
Received: from xhc-aln-x07.cisco.com (xhc-aln-x07.cisco.com [173.36.12.81]) by rcdn-core-10.cisco.com (8.14.5/8.14.5) with ESMTP id t3NF7DpC032285 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Thu, 23 Apr 2015 15:07:13 GMT
Received: from xmb-rcd-x09.cisco.com ([169.254.9.172]) by xhc-aln-x07.cisco.com ([173.36.12.81]) with mapi id 14.03.0195.001; Thu, 23 Apr 2015 10:07:13 -0500
From: "Fred Baker (fred)" <fred@cisco.com>
To: =?utf-8?B?TWlyamEgS8O8aGxld2luZA==?= <mirja.kuehlewind@tik.ee.ethz.ch>, Bob Briscoe <bob.briscoe@bt.com>, Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Thread-Topic: AQM and ECN in draft-ietf-aqm-recommendation-11
Thread-Index: AQHQfdctsOd8CA6bH0eQqTzbaO5JzA==
Date: Thu, 23 Apr 2015 15:07:12 +0000
Message-ID: <F16CFD29-365E-417F-8380-378516F3FF07@cisco.com>
References: <5538D53D.1010908@tik.ee.ethz.ch>
In-Reply-To: <5538D53D.1010908@tik.ee.ethz.ch>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: yes
X-MS-TNEF-Correlator: 
x-originating-ip: [10.19.64.121]
Content-Type: multipart/signed; boundary="Apple-Mail=_09169CCD-938A-46FC-8807-D16B626536BB"; protocol="application/pgp-signature"; micalg=pgp-sha1
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/sQ0zbNdgSOu6VH5jR1mPLm1zG8U>
Cc: "Scheffenegger, Richard" <rs@netapp.com>, "aqm@ietf.org list" <aqm@ietf.org>
Subject: Re: [aqm] AQM and ECN in draft-ietf-aqm-recommendation-11
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 23 Apr 2015 15:07:43 -0000

--Apple-Mail=_09169CCD-938A-46FC-8807-D16B626536BB
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

Bob/Gorry:

Didn=E2=80=99t we have essentially the same comment a week or two ago? =
What was our resolution?

I think it was a minor rewording that included a little more than =
s/threshold/parameters/, and which I worried was heading in the =
direction of a much bigger question - "what parameters".

Fred

> On Apr 23, 2015, at 4:19 AM, Mirja K=C3=BChlewind =
<mirja.kuehlewind@tik.ee.ethz.ch> wrote:
>=20
> Hi all,
>=20
> I know I'm too late for this but want to bring it up anyway:
>=20
> draft-ietf-aqm-recommendation-11 currently says:
> "An AQM algorithm that supports ECN needs to define the threshold and
>   algorithm for ECN-marking.  This threshold MAY differ from that used
>   for dropping packets that are not marked as ECN-capable, and SHOULD
>   be configurable."
>=20
> But I think it should be:
> "An AQM algorithm that supports ECN needs to define the parameters and
>   algorithm for ECN-marking.  These parameters MAY differ from that =
used
>   for dropping packets that are not marked as ECN-capable, and SHOULD
>   be configurable."
>=20
> so just s/threshold/parameters/  ...?
>=20
> I've just caught this now because I'm reading =
draft-welzl-ecn-benefits-02; sorry. Maybe it's still possible to make =
such a small edit... if other people think as well that this should be =
changed...?
>=20
> Mirja
>=20
>=20


--Apple-Mail=_09169CCD-938A-46FC-8807-D16B626536BB
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment; filename="signature.asc"
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Message signed with OpenPGP using GPGMail

-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org

iQEVAwUBVTkKoJ9ieig10VPpAQLU+wf/a7B3SPF2wDBJ3rg5+y59w9ozEHIlJq4Z
vp0cpXxgBi4c6RkxsfD7Gcvcs68J8x/c3yFImTW5lVYakPNIrmLBnz96zUXucVks
UbFvIBPYY9p6XjBULNTiEhN45yth4PFThsE/NhFWPt2LtlPC1tNheyBwv6slWnWz
oopCgCqOI3WfJIZeurcCTrNmkoJYsN6s0mE01p0MtSOc/qTLBvN6iQ7u8Ke7b7Ki
6OsvIO11ixG25mpG81yfFdqfLofMhuT7/TQ02JQ9z45wGsGmZWokF5Pu0WZav0XN
o1OqvKGXTUQYGoOQqCUdWLn2bqxoo4weLzEKvRSJe6VzN7xxvmhNdA==
=BlE3
-----END PGP SIGNATURE-----

--Apple-Mail=_09169CCD-938A-46FC-8807-D16B626536BB--


From nobody Thu Apr 23 10:29:36 2015
Return-Path: <mirja.kuehlewind@tik.ee.ethz.ch>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 83DD21ACD98 for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 10:29:34 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.91
X-Spam-Level: 
X-Spam-Status: No, score=-3.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id jkkTHXXV2GJ9 for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 10:29:31 -0700 (PDT)
Received: from smtp.ee.ethz.ch (smtp.ee.ethz.ch [129.132.2.219]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 2A5AA1ACEAC for <aqm@ietf.org>; Thu, 23 Apr 2015 10:28:56 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by smtp.ee.ethz.ch (Postfix) with ESMTP id 2911FD930A; Thu, 23 Apr 2015 19:28:55 +0200 (MEST)
X-Virus-Scanned: by amavisd-new on smtp.ee.ethz.ch
Received: from smtp.ee.ethz.ch ([127.0.0.1]) by localhost (.ee.ethz.ch [127.0.0.1]) (amavisd-new, port 10024) with LMTP id gMyafaHCCGXt; Thu, 23 Apr 2015 19:28:54 +0200 (MEST)
Received: from [82.130.103.143] (nb-10510.ethz.ch [82.130.103.143]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: mirjak) by smtp.ee.ethz.ch (Postfix) with ESMTPSA id CFF64D9303; Thu, 23 Apr 2015 19:28:54 +0200 (MEST)
Message-ID: <55392BD6.501@tik.ee.ethz.ch>
Date: Thu, 23 Apr 2015 19:28:54 +0200
From: =?UTF-8?B?TWlyamEgS8O8aGxld2luZA==?= <mirja.kuehlewind@tik.ee.ethz.ch>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: aqm@ietf.org, Michael Welzl <michawe@ifi.uio.no>,  Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/z9bQJ0K2fHlx00FcqmM4LjXf4pY>
Subject: [aqm] Review of  draft-ietf-aqm-ecn-benefits-03
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 23 Apr 2015 17:29:34 -0000

Hi Gorry, hi Michael,

as promised here is my review of draft-ietf-aqm-ecn-benefits-03.

My overall comment is that even after reading the document (or even slightly 
more than before) I'm not completely sure what the purpose of this document is 
and also what the audience is this documented is directed to. Currently this 
document seems to do two things: 1. it list benefits (which is interesting for 
someone who thinks about enabling ECN) and 2. it kind of outlines needed steps 
for deployment (which would be directed to someone who gets the task from his 
manager to turn on ECN). However, the second point is not clearly spelled out 
and therefore it might be rather confusing for some people to read the second 
part of the document. Also the second part is to some extend still 
work-in-progress, therefor I would recommend to only focus this document on the 
first part.

For the first part (listing benefits) it might also be good to make 
clear/distinguish who has these benefits. I think all benefits that are 
currently listed are only advantageous for the end host/application. Are there 
any benefits for a network operator? Would it be possible to write this document 
such that I could also use it to point network operators to and give them an 
incentive to enable ECN?

Another high level comment is that you say in the introduction that this 
document "also identifies some potential problems that might occur when ECN is 
used" but then you don't really discuss them. I think to show both sides of the 
coin in this document would make the document more useful (and more honest). One 
point that you mention slightly here is that cheating is easier than with loss 
by not providing the feedback. Another point might be fairness between ECN and 
non-ECN traffic as marking will not reduce the queue length and therefore might 
lead to a higher loss rate for the non-ECN traffic instaed. I guess there are 
papers about this; don't have any by hand right now. Are there any other 
problems that should be mentioned?

Find more detailed comment by section below:

Abstract
--------
...says "...potential benefits when applications enable Explicit Congestion 
Notification (ECN)" -> usually an application cannot able ECN because usually 
it's a system setting...?

Section 1
---------
... says "..separate
    configuration of the drop and mark thresholds is known to be
    supported in some network devices and this is recommended
    [RFC2309.bis]."
RFC2309bis does not recommend different settings, it only say that it should be 
possible have different configuration of both. Further, I think this should not 
only concern THE threshold (whatever this is) but usually there are several 
parameters you might want to set independent of each other, e.g. the max 
mark/drop probability in RED.


Section 2
----------
1) I'm not sure I understand the purpose of this section or maybe just the title 
is wrong. I'm currently seeing this section rather as a section that provides 
the needed background knowledge than is talking about deployment. For this 
purpose I'd put all references and potentially a brief summary to other 
RFC/drafts on ECN in this section including RFC2884, RFC4774, RFC5562, RFC6040, 
RFC6679, draft-briscoe-tsvwg-ecn-encap-guidelines and 
draft-ietf-tcpm-accecn-reqs (and rename it).

2) Second paragraph says:
"Network devices must not drop packets solely because these codepoints are used 
[RFC2309.bis]."
Not sure this is the right document to says this (because currently it not seems 
to be directed to network operator/equipment vendors but admins/application 
developers). However, if it says this, it should also say that network devices 
should not bleach these bits.

3) First bullet in list says
"A recent survey reported growing support for ECN on common network paths [TR15]."
This sounds like TR15 shows that ECN is actually used in the Internet. However, 
TR15 only shows that there are only very few cases left where ECN packets are 
dropped or incorrectly altered. Please clarify or remove this sentence here.

4) You could cite draft-bensley-tcpm-dctcp-00 instead of the DCTCP Sigcomm paper 
(or both).

5) I would remove the subsection headings (both 2.1 and 2.2) and just add the 
text there to the main part of the section.

6) "An AQM algorithm that supports ECN needs to define
    the threshold and algorithm for ECN-marking."
This is kind of self-redundant and therefore does not really makes sense to me 
to say; of course an algo that supports ECN needs to say something about ECN...

7) You can use TR15 to provide a reference for the first paragraph in section 2.2:
"Cases have been noted where a sending endpoint marks a packet with a
    non-zero ECN mark, but the packet is received with a zero ECN value
    by the remote endpoint."

8) I'd move the second paragraph of section 2.2. ("The current..") to a 
potentially new problems section, talking about known/previous deployment problems.

9) I would simply remove paragraph 3-4 of section 2.2 because this was basically 
as already mentioned by referring to 2309bis and rfc6040 in section 2.1.

Section 3.2
------------
1) Don't understand why there is a listing here...? Just remove the listing and 
make text out of it...?

2) The sentence "This also
       avoids the inefficiency of dropping data that has already made it
       across at least part of the network path."
does not belong in this section. This sentence should just be moved to section 
3.1 (or in an own section) and must be further explained, saying that dropping 
packet at the of the path has already blocked resources that other traffic could 
have used otherwise.

Section 3.3
-----------
1) I'd say this section misses on part of the discussion. It is true that if by 
chance your last  packet(s) get lost ECN can help. However, this section reads a 
little like, with ECN it is save to send packet bursts. Which is not true 
because even if ECN is used by a network device, the queue might be too small to 
hold the whole burst. I believe this case happen very often which might be a 
reason for the higher tail loss probability that sometimes is experienced with 
IW10. Please add this point to the discussion.

2) I don't really get the point of the second paragraph. First of all it is 
confusion that this paragraph starts which "In addition to avoiding HOL 
blocking,.."; I guess that is left over from a previous version of this text...? 
And then you talk about a connection that is currently idle, so why is the 
performance of this connection that is currently not sending anything reduced?

3) I don't understand what "applications that send intermittent bursts of data, 
and rely upon timer-based recovery of packet loss" are...? Isn't the transport 
responsible to not send bursts and care about recovery...?

4) For the last paragraph in section 3.3 note that stacks often remember RTT 
measurements for a certain IP address and set the initial RTO based on this 
information.

Section 3.4
-----------
You still need FEC or some kind of error concealment even if ECN is used because 
you can never be sure that your packet are not get dropped (by non-ECN-enable 
devices or other reasons). Therefore using ECN will clearly not reduce 
complexity. The only thing you can do is to potentially reduce the amount of 
redundancy you send if you know that a certain path is ECN enables or don't see 
losses at the beginning of a connection. This can save network resources but 
actually might not improve user experience; in fact the user experience might be 
worse in case there are sudden losses.

Further the text says "negative impact of using loss-hiding mechanisms"; I don't 
really think that FEC has a negative impact as long as you've send enough 
redundancy...? Error concealment might but is used less and less. I'd recommend 
to talk about error concealment only in this last paragraph and explain a little 
further.

Section 3.5
----------
"Recording the presence of CE-marked packets can therefore provide
    information about the performance of the network path."
Would change to:
"Recording the presence of CE-marked packets in absence of loss can therefore 
provide
    information about the performance of the network path."
And also say more concretely what is meant with 'performance of the network 
path' -> congestion level or no drops by other middleboxes on this path...

Section 3.6
-----------
1) I like the section but I would phrase it differently; also it's not clear who 
needs to support what in this case. I'd like to propose the following text [not 
sure about the heading...]:

"3.6 Opportunity to provide an improved congestion feedback signal

Loss and ECN marking are both used as an indication for congestion. However, 
while the amount of feedback that is provided by loss should naturally be 
minimized, this is not the case for ECN. With ECN a network node could provide 
richer and more frequent feedback on the congestion state of a link which then 
could be used by the control mechanisms implemented in end host to make a more 
appropriate decision on how to react to congestion and to react faster to 
changes in congestion state.

Further while drop-based AQM mechanisms usually operate on a smoothed queue 
length estimation (instead of the instantaneous queue length) and therefore 
slightly delay the feedback signal to avoid unnecessary losses in case of 
transient congestion, this would be not necessary for ECN. If congestion is only 
transient due to short traffic bursts that are active for less than one RTT, the 
congestion signal would reach the sender at a time where the congestion is 
already cleared up. However, instead delaying the feedback in the network, the 
end host could reduce its sending rate incrementally based on the extend of 
congestion (that was experienced over e.g. the last RTT) similar as DCTCP. In 
case if the congestion is only transient, the end host would only reduce its 
rate slightly and be able to catch up quickly again. However, in case the 
congestion is persistent, this would help to remove additional delays from the 
network and resolve congestion faster which after all reduces the average 
queuing delay.

However, current ECN is defined as a 'drop equivalent' in RFC3168. To change the 
semantics of ECN both the AQM in the network nodes and the control mechanism in 
the end hosts would still need to cope with nodes or end hosts that rely on the 
old semantics. Therefore changing the semantics can be done more easily in 
confined environment such as a data center. DCTCP is an example that changes 
both the configuration of the used AQM as well as the congestion response in the 
end host and relies on that fact that all nodes in data center are configured 
the same way. [Deployment strategies to change the semantics of ECN in the 
Internet are currently under discussion in the IETF.]"

2) I'd move the 1. and 2. paragraph of section 3.6.1 to the 
background/deployment section or to the intro depending what you going to do 
with section 2.

Sections 4 & 5
---------
First sentence talks about "operational
    difficulties when the network only partially supports the use of ECN,
    or to respond to the challenges due to misbehaving network devices
    and/or endpoints".
I think these are to very different things. Misbehaving network devices is a 
point for a  problems section (where the lesson learned is that we didn't think 
carefully enough about incremental deployment in the first place but do now). 
However, partial deployment is not a problem but is a thing we simply have to 
cope with. The text sound as if the goal would be that every router in the whole 
Internet would at some point of time be ECN-enabled. I don't think this will 
ever happen and is also not the goal for me. Routers that are very unlikely to 
ever get congested should no be required to look at the ECN bits or monitor the 
queue length to calculate a mark/drop probability.

However as I said at the beginning I don't really thing that sections 4 and 5 
belong in this document. If you decided to keep them (you have to change the 
abstract) and I'd recommend to rename them e.g 4. 'Incremental Deployment 
Strategy' or 'Requirements to enable Incremental Deployment' and 5. 
'Recommendations for enabling ECN in network nodes and end hosts'.


I hope that's helpful! Let me know if you have any questions!

Mirja



From nobody Thu Apr 23 21:38:23 2015
Return-Path: <wes@mti-systems.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id BDF2D1ACF16 for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 21:38:21 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: 0.8
X-Spam-Level: 
X-Spam-Status: No, score=0.8 tagged_above=-999 required=5 tests=[BAYES_50=0.8] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id SjK-JpClsyxe for <aqm@ietfa.amsl.com>; Thu, 23 Apr 2015 21:38:19 -0700 (PDT)
Received: from atl4mhob10.myregisteredsite.com (atl4mhob10.myregisteredsite.com [209.17.115.48]) by ietfa.amsl.com (Postfix) with ESMTP id EFB0F1B2C2B for <aqm@ietf.org>; Thu, 23 Apr 2015 21:37:55 -0700 (PDT)
Received: from mailpod.hostingplatform.com ([10.30.71.203]) by atl4mhob10.myregisteredsite.com (8.14.4/8.14.4) with ESMTP id t3O4bsmO032119 for <aqm@ietf.org>; Fri, 24 Apr 2015 00:37:54 -0400
Received: (qmail 18993 invoked by uid 0); 24 Apr 2015 04:37:54 -0000
X-TCPREMOTEIP: 69.81.157.169
X-Authenticated-UID: wes@mti-systems.com
Received: from unknown (HELO ?192.168.1.112?) (wes@mti-systems.com@69.81.157.169) by 0 with ESMTPA; 24 Apr 2015 04:37:53 -0000
Message-ID: <5539C899.4080704@mti-systems.com>
Date: Fri, 24 Apr 2015 00:37:45 -0400
From: Wesley Eddy <wes@mti-systems.com>
Organization: MTI Systems
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: "aqm@ietf.org" <aqm@ietf.org>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/rpcVmU7QA3MMhIN_TF14bPNAHTg>
Subject: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2015 04:38:21 -0000

To keep moving forward with the set of documents, we'd like to start
a Working Group Last Call on:
http://datatracker.ietf.org/doc/draft-ietf-aqm-ecn-benefits/

Please send any comments you have on this document to the list
by May 8th.  Other than those raised in the thread Mirja has just
opened with her review, I'm not aware of any other open issues.

Thank you to everyone that's been following up on the reviews
pledged at the last meeting!


-- 
Wes Eddy
MTI Systems


From nobody Fri Apr 24 02:25:28 2015
Return-Path: <swmike@swm.pp.se>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 388041B36A7 for <aqm@ietfa.amsl.com>; Fri, 24 Apr 2015 02:25:27 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.961
X-Spam-Level: 
X-Spam-Status: No, score=-3.961 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_SE=0.35, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id m-YzDfZo_Vlf for <aqm@ietfa.amsl.com>; Fri, 24 Apr 2015 02:25:24 -0700 (PDT)
Received: from uplift.swm.pp.se (swm.pp.se [212.247.200.143]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 5C3201B3689 for <aqm@ietf.org>; Fri, 24 Apr 2015 02:25:24 -0700 (PDT)
Received: by uplift.swm.pp.se (Postfix, from userid 501) id 3F2DBA1; Fri, 24 Apr 2015 11:25:22 +0200 (CEST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=swm.pp.se; s=mail; t=1429867522; bh=riTZEC52gd6iPxs+S2YtmXiTrP7DWnIUtbz3se8WuX4=; h=Date:From:To:Subject:In-Reply-To:References:From; b=3a2YIsxF+ZsKlUJR7CgtBv3l/CbzRSYb8vWcF38wh84QJdqjsjt/pczer+EN0puK2 MzmzOs4xXrt98JOQXZG2teaGvwvNXi4uYZiPM0ChQq/hN5IAfIZma4H3lon2SuhKBs l5/fFPcx6Ja1xUrw7aNaEo9KqFFUfEqL7ILFV93M=
Received: from localhost (localhost [127.0.0.1]) by uplift.swm.pp.se (Postfix) with ESMTP id 3BA929F for <aqm@ietf.org>; Fri, 24 Apr 2015 11:25:22 +0200 (CEST)
Date: Fri, 24 Apr 2015 11:25:22 +0200 (CEST)
From: Mikael Abrahamsson <swmike@swm.pp.se>
To: "aqm@ietf.org" <aqm@ietf.org>
In-Reply-To: <5539C899.4080704@mti-systems.com>
Message-ID: <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se>
References: <5539C899.4080704@mti-systems.com>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
Organization: People's Front Against WWW
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/IXBNBgyacNWaDxg4aMLV0EosgWE>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Fri, 24 Apr 2015 09:25:27 -0000

On Fri, 24 Apr 2015, Wesley Eddy wrote:

> To keep moving forward with the set of documents, we'd like to start
> a Working Group Last Call on:
> http://datatracker.ietf.org/doc/draft-ietf-aqm-ecn-benefits/
>
> Please send any comments you have on this document to the list
> by May 8th.  Other than those raised in the thread Mirja has just
> opened with her review, I'm not aware of any other open issues.

I have re-read this document as if I never read it before. Writing my 
thoughts as I read it.

The title says "benefits" but the abstract states that the document will 
talk about both benefits and potential problems. I therefore think the 
title does not reflect the abstract.

1. Introduction: I don't understand the "(potential)", doesn't transport 
always treat loss and CE marking as indications of congestion? Actually, 
the way the introduction reads today one might interpret that TCP and SCTP 
are "Internet transports" but UDP is not, because it's merely "transport".

Can we have a reference to AQM in the second paragraph? What is the 
difference between a router and a middle box?

I don't like the first sentence in the 3rd paragraph, shouldn't it say 
that the device "might mark instead of drop", and not say "as the queue 
builds"? Because if we have a cisco-style policer, there is no queue but 
instead a token-bucket rate limiter. I don't like "as the network queue 
builds" anyhow, if it should stay, it should be rephrased to "interface 
queue size increases" or something like that. What if there are several 
queues? I'd like the last part to not be included at all. Also, is the 
document really saying that ECN should be done before incuring loss for 
non-ECN traffic? The current wording can be interpreted as such.

Fourth parahraph. I have no idea what "incipient" means. I looked it up. 
Can we use the word "developing" or some other synonym?

2.2. Here we have incipient again, and isn't drop or CE marking an 
indicator of actual congestion, not just developing/incipient congestion?

How do we stop older bleaching-boxes that do not understand ECN bits and 
still bleach them, from just forwarding the packet even if CE=1? I 
understand the must requirement and it makes sense, but I don't see how 
this can be achieved with older equipment?

3.1. I would like to see text after 0.02 that says something like (1 
packet in 50) or something like that, just for clarification, or that the 
packet drop rate is 2%. It's not obvious that "the packet drop ratio is 
0.02" means these things.


My summary: I would like to see some text on the downsides of having ECN 
transport, ECN enabled routers and having middle box bleaching ECN bits, 
thus removing the congestion signal. This downside, and the operational 
problem that some old equipment just simply drop packets with ECN bits set 
should be summarized the same way as the benfits. These are the only 
drawbacks of enabling ECN that I can see, and potentially text describing 
how to detect these problems could be added/extended (it's somewhat hinted 
at currently).

Generally, I support the document, it's useful, provides good insight, and 
is valuable to the community. Thanks for writing it!

-- 
Mikael Abrahamsson    email: swmike@swm.pp.se


From nobody Mon Apr 27 04:10:32 2015
Return-Path: <gorry@erg.abdn.ac.uk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 03A5D1B3098 for <aqm@ietfa.amsl.com>; Mon, 27 Apr 2015 04:10:32 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.911
X-Spam-Level: 
X-Spam-Status: No, score=-3.911 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2C5JC7IKLdZl for <aqm@ietfa.amsl.com>; Mon, 27 Apr 2015 04:10:30 -0700 (PDT)
Received: from pegasus.erg.abdn.ac.uk (pegasus.erg.abdn.ac.uk [139.133.204.173]) by ietfa.amsl.com (Postfix) with ESMTP id 2BB341B30A5 for <aqm@ietf.org>; Mon, 27 Apr 2015 04:10:30 -0700 (PDT)
Received: from gorry-mac.erg.abdn.ac.uk (unknown [IPv6:2001:630:241:207:21f:5bff:fe38:7354]) by pegasus.erg.abdn.ac.uk (Postfix) with ESMTPSA id 83D5C1B00283; Mon, 27 Apr 2015 12:10:39 +0100 (BST)
Message-ID: <553E191D.1020009@erg.abdn.ac.uk>
Date: Mon, 27 Apr 2015 12:10:21 +0100
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Organization: The University of Aberdeen is a charity registered in Scotland,  No SC013683. 
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: =?UTF-8?B?TWlyamEgS8O8aGxld2luZA==?= <mirja.kuehlewind@tik.ee.ethz.ch>, aqm@ietf.org, fred@cisco.com,  "Scheffenegger, Richard" <rs@netapp.com>
References: <5538D53D.1010908@tik.ee.ethz.ch>
In-Reply-To: <5538D53D.1010908@tik.ee.ethz.ch>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/p9bnTCJzPAdLBZLsBrEFntpZUVc>
Subject: Re: [aqm] AQM and ECN in draft-ietf-aqm-recommendation-11
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: gorry@erg.abdn.ac.uk
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2015 11:10:32 -0000

On 23/04/2015 12:19, Mirja K�ĵhlewind wrote:
> Hi all,
>
> I know I'm too late for this but want to bring it up anyway:
>
> draft-ietf-aqm-recommendation-11 currently says:
> "An AQM algorithm that supports ECN needs to define the threshold and
>     algorithm for ECN-marking.  This threshold MAY differ from that used
>     for dropping packets that are not marked as ECN-capable, and SHOULD
>     be configurable."
>
> But I think it should be:
> "An AQM algorithm that supports ECN needs to define the parameters and
>     algorithm for ECN-marking.  These parameters MAY differ from that used
>     for dropping packets that are not marked as ECN-capable, and SHOULD
>     be configurable."
>
> so just s/threshold/parameters/  ...?
>
> I've just caught this now because I'm reading
> draft-welzl-ecn-benefits-02; sorry. Maybe it's still possible to make
> such a small edit... if other people think as well that this should be
> changed...?
>
> Mirja
>
>

We are indeed considering one very minor change to the wording here to 
prevent future confusion. Once we have text, we'll share with the WG to 
be sure this is acceptable.

Gorry & Fred


From nobody Mon Apr 27 04:11:08 2015
Return-Path: <gorry@erg.abdn.ac.uk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id C31BB1B30AA for <aqm@ietfa.amsl.com>; Mon, 27 Apr 2015 04:11:06 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.911
X-Spam-Level: 
X-Spam-Status: No, score=-3.911 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id iJb8D0ai9hmf for <aqm@ietfa.amsl.com>; Mon, 27 Apr 2015 04:11:05 -0700 (PDT)
Received: from pegasus.erg.abdn.ac.uk (pegasus.erg.abdn.ac.uk [139.133.204.173]) by ietfa.amsl.com (Postfix) with ESMTP id 77A841B30A8 for <aqm@ietf.org>; Mon, 27 Apr 2015 04:11:05 -0700 (PDT)
Received: from gorry-mac.erg.abdn.ac.uk (unknown [IPv6:2001:630:241:207:21f:5bff:fe38:7354]) by pegasus.erg.abdn.ac.uk (Postfix) with ESMTPSA id 055AE1B0028D; Mon, 27 Apr 2015 12:11:19 +0100 (BST)
Message-ID: <553E1948.8020206@erg.abdn.ac.uk>
Date: Mon, 27 Apr 2015 12:11:04 +0100
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Organization: The University of Aberdeen is a charity registered in Scotland,  No SC013683. 
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: =?windows-1252?Q?Mirja_K=FChlewind?= <mirja.kuehlewind@tik.ee.ethz.ch>, aqm@ietf.org, fred@cisco.com,  "Scheffenegger, Richard" <rs@netapp.com>
References: <5538D53D.1010908@tik.ee.ethz.ch> <5538DE0B.8070304@tik.ee.ethz.ch> <5538FE93.2090501@tik.ee.ethz.ch>
In-Reply-To: <5538FE93.2090501@tik.ee.ethz.ch>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/R5XUkz3ccF1YiG53z4Kgg-KqhUw>
Subject: Re: [aqm] AQM and ECN in draft-ietf-aqm-recommendation-11
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: gorry@erg.abdn.ac.uk
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Mon, 27 Apr 2015 11:11:06 -0000

The present text is intentionally descriptive only, and ties-in with the 
bleaching text. I personally think this deployment advice is consistent 
with the intention of the AQM recomemndations, but I agree it is not 
specifically called-out in RFC2309.bis, and this reference therefore 
really needs to be removed in the next rev.

This is not the place to make a BCP recommendation.  If "we" get a 
chance, I also think the IETF should assert this using RFC2119 language 
in an appropriate future document.

Gorry

On 23/04/2015 15:15, Mirja Kühlewind wrote:
> I meant draft-ietf-aqm-ecn-benefits-03 (not draft-welzl-ecn-benefits-02)
> though...
>
> On 23.04.2015 13:56, Mirja Kühlewind wrote:
>> Hi again,
>>
>> I still know I'm late but there are two more points:
>>
>> draft-welzl-ecn-benefits-02 says:
>> "Network devices must not drop
>>      packets solely because these codepoints are used [RFC2309.bis]"
>> and
>> "A network device should therefore not
>>      remark an ECT(0) or ECT(1) mark to zero [RFC2309.bis]."
>>
>> I could not find this recommendation in
>> draft-ietf-aqm-recommendation-11, at
>> least not in the section 4.2.1 (AQM and ECN), but it should be in
>> there (and
>> should actually say: a network device MUST NOT remark, ECT(0), ETC(1)
>> or CE to
>> zero)...?
>>
>> Mirja
>>
>>
<snip>


From nobody Mon Apr 27 17:58:27 2015
Return-Path: <internet-drafts@ietf.org>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 79A171ACDD3; Mon, 27 Apr 2015 17:58:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.9
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 tagged_above=-999 required=5 tests=[BAYES_00=-1.9] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id R72dSz_6uO3N; Mon, 27 Apr 2015 17:58:25 -0700 (PDT)
Received: from ietfa.amsl.com (localhost [IPv6:::1]) by ietfa.amsl.com (Postfix) with ESMTP id 16F4C1ACDD7; Mon, 27 Apr 2015 17:58:25 -0700 (PDT)
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
From: internet-drafts@ietf.org
To: <i-d-announce@ietf.org>
X-Test-IDTracker: no
X-IETF-IDTracker: 6.0.2
Auto-Submitted: auto-generated
Precedence: bulk
Message-ID: <20150428005825.13632.20787.idtracker@ietfa.amsl.com>
Date: Mon, 27 Apr 2015 17:58:25 -0700
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/vDYcSSdDOZsU0b8iOqL-QKeSfTc>
Cc: aqm@ietf.org
Subject: [aqm] I-D Action: draft-ietf-aqm-codel-01.txt
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 00:58:26 -0000

A New Internet-Draft is available from the on-line Internet-Drafts directories.
 This draft is a work item of the Active Queue Management and Packet Scheduling Working Group of the IETF.

        Title           : Controlled Delay Active Queue Management
        Authors         : Kathleen Nichols
                          Van Jacobson
                          Andrew McGregor
                          Jana Iyengar
	Filename        : draft-ietf-aqm-codel-01.txt
	Pages           : 28
	Date            : 2015-04-27

Abstract:
   The "persistently full buffer" problem has been discussed in the IETF
   community since the early 80's [RFC896].  The IRTF's End-to-End
   Working Group called for the deployment of active queue management
   (AQM) to solve the problem in 1998 [RFC2309].  Despite the awareness,
   the problem has only gotten worse as Moore's Law growth in memory
   density fueled an exponential increase in buffer pool size.  Efforts
   to deploy AQM have been frustrated by difficult configuration and
   negative impact on network utilization.  This problem, recently
   christened "bufferbloat", [TSVBB2011] [BB2011] has become
   increasingly important throughout the Internet but particularly at
   the consumer edge.

   This document describes a general framework called CoDel (Controlled
   Delay) [CODEL2012] that controls bufferbloat-generated excess delay
   in modern networking environments.  CoDel consists of an estimator, a
   setpoint, and a control loop.  It requires no configuration in normal
   Internet deployments.  CoDel comprises some major technical
   innovations and has been made available as open source so that the
   framework can be applied by the community to a range of problems.  It
   has been implemented in Linux (and available in the Linux
   distribution) and deployed in some networks at the consumer edge.  In
   addition, the framework has been successfully applied in other ways.

   Note: Code Components extracted from this document must include the
   license as included with the code in Section 5.


The IETF datatracker status page for this draft is:
https://datatracker.ietf.org/doc/draft-ietf-aqm-codel/

There's also a htmlized version available at:
https://tools.ietf.org/html/draft-ietf-aqm-codel-01

A diff from the previous version is available at:
https:https://www.ietf.org/rfcdiff?url2=draft-ietf-aqm-codel-01


Please note that it may take a couple of minutes from the time of submission
until the htmlized version and diff are available at tools.ietf.org.

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


From nobody Tue Apr 28 05:37:48 2015
Return-Path: <gorry@erg.abdn.ac.uk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 374751A9029 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 05:37:47 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.911
X-Spam-Level: 
X-Spam-Status: No, score=-1.911 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 9-MTJ24JEA-H for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 05:37:44 -0700 (PDT)
Received: from pegasus.erg.abdn.ac.uk (pegasus.erg.abdn.ac.uk [IPv6:2001:630:241:204::f0f0]) by ietfa.amsl.com (Postfix) with ESMTP id 571321A9039 for <aqm@ietf.org>; Tue, 28 Apr 2015 05:37:44 -0700 (PDT)
Received: from gorry-mac.erg.abdn.ac.uk (unknown [IPv6:2001:630:241:207:21f:5bff:fe38:7354]) by pegasus.erg.abdn.ac.uk (Postfix) with ESMTPSA id E1A8D1B00519; Tue, 28 Apr 2015 13:37:55 +0100 (BST)
Message-ID: <553F7F15.9060509@erg.abdn.ac.uk>
Date: Tue, 28 Apr 2015 13:37:41 +0100
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Organization: The University of Aberdeen is a charity registered in Scotland,  No SC013683. 
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: Mikael Abrahamsson <swmike@swm.pp.se>, "aqm@ietf.org" <aqm@ietf.org>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se>
In-Reply-To: <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/2x6NzcHoSj_8_OxhGlHSdpQ8J74>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: gorry@erg.abdn.ac.uk
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 12:37:47 -0000

It is great to see some detailed comments from you looking afresh at 
this. Please see below for initial replies.

 >On Fri, 24 Apr 2015, Wesley Eddy wrote:
 >
 >To keep moving forward with the set of documents, we'd like to start
 >a Working Group Last Call on:
 >http://datatracker.ietf.org/doc/draft-ietf-aqm-ecn-benefits/
 >>
 >Please send any comments you have on this document to the list
 >by May 8th.  Other than those raised in the thread Mirja has just
 >opened with her review, I'm not aware of any other open issues.
 >
 >I have re-read this document as if I never read it before. Writing my
 >thoughts as I read it.
 >
 >The title says "benefits" but the abstract states that the document will
talk about both benefits and potential problems. I therefore think the
title does not reflect the abstract.
 >
GF/MW: We looked at both approaches, and revised to include "drawbacks" as
well, but then the list discussion seemed to favour focusing on the
"benefits", since deployment issues could change with time. I suggest we
remove the sentence about "problems" from the abstract.
 >
 >1. Introduction: I don't understand the "(potential)", doesn't transport
always treat loss and CE marking as indications of congestion?
Actually, the way the introduction reads today one might interpret that
TCP and SCTP are "Internet transports" but UDP is not, because it's
merely "transport".
 >
GF/MW: We need to add REF to AQM recommendations [RFC2309.bis].
 >
GF/MW:  We'd agree, and be happy to remove "(potential)" from section 1
para 1.
 >
 >Can we have a reference to AQM in the second paragraph? What is the
 >difference between a router and a middle box?
 >
GF/MW: This echoes what is said in the AQM Recommendations draft: "The
mechanisms described in this document may be implemented in
network devices on the path between end-points that include routers,
switches, and other network middleboxes." - The difference may be in the
eyes of the beholder - all are network devices that can deploy AQM.
 >
 >I don't like the first sentence in the 3rd paragraph, shouldn't it say
 >that the device "might mark instead of drop", and not say "as the queue
 >builds"? Because if we have a cisco-style policer, there is no queue but
instead a token-bucket rate limiter. I don't like "as the network queue
builds" anyhow, if it should stay, it should be rephrased to "interface
queue size increases" or something like that.
 >
GF/MW: I agree policiers and other methods can track congestion state and
set a CE-mark. I'm OK with some change to "queue builds", but do not yet
understand the difference between "interface queue size increases", except
that would seem to limit only to an interface. It is OK with us to remove
"as the queue builds".
 >
 >What if there are several queues? I'd like the last part to not be
included at all.
 >
 >I suggest we add this sentence after the next para also, to balance this:
 >"An endpoint that receives a congestion mark
 > causes a congestion control reaction at the sender to reduce
 > the maximum rate permitted by the endpoint."
 >
 >Also, is the  document really saying that ECN should be done before
incurring loss for non-ECN traffic? The current wording can be
interpreted as such.
 >
GF/MW: It enables marking before the point where a drop-tail queue would
have dropped. The ID intentionally does not describe how to implement or
to evaluate if this is done "fairly". The suggested sentence seems to be
unnecessary because it echoes what is in RFC3168, to some degree, but also
uses phrasing that can be misleading, by saying: "maximum rate permitted".
 >
 >Fourth pararaph. I have no idea what "incipient" means. I looked it up.
Can we use the word "developing" or some other synonym?
 >
GF/MW: OK this came from the list discussion. This term has already been
used in early discussion of RED marking (e.g.,
ftp://ftp.ee.lbl.gov/papers/early.pdf) It was used in RFC 2481, and is
used several times in RFC 3168. I suggest we add some explanatory words
(e.g., "where congestion is beginning to be experienced")
 >
 >2.2. Here we have incipient again, and isn't drop or CE marking an
 >indicator of actual congestion, not just developing/incipient congestion?
 >
GF/MW: See above, one advantage of talking about "incipient" congestion is
we avoid talking about "mild" or "initial" or some other term that could
perhaps be misconstrued. It is is congestion, but reported when first
discovered.
 >
 >How do we stop older bleaching-boxes that do not understand ECN bits and
still bleach them, from just forwarding the packet even if CE=1? I
 >understand the must requirement and it makes sense, but I don't see how
 >this can be achieved with older equipment?
 >
GF/MW: We don't see more options than:
* Declare them as not complying with current use
* Go and fix those that can be fixed, and live with the others until in
the future they are finally replaced.
- I'm not sure what extra there is to say about this, apart from to
verify the operation across the path (as in section 4.1).
 >
GF/MW: The phrase:
"A network device should therefore not remark an ECT(0) or ECT(1) mark to
zero [RFC2309.bis]."
- This deployment issue is consistent with the intention of RFC 2309.bis,
but is not specifically called-out in RFC2309.bis, and this reference
needs to be removed.
 >
 >3.1. I would like to see text after 0.02 that says something like (1
 >packet in 50) or something like that, just for clarification, or that
the  packet drop rate is 2%. It's not obvious that "the packet drop
ratio is 0.02" means these things.
 >
GF/MW: We think this change should be made.
 >
 >My summary: I would like to see some text on the downsides of having ECN
transport, ECN enabled routers and having middle box bleaching ECN
bits, thus removing the congestion signal. This downside, and the
operational problem that some old equipment just simply drop packets
with ECN bits set should be summarized the same way as the benfits.
These are the only drawbacks of enabling ECN that I can see, and
potentially text describing how to detect these problems could be
added/extended (it's somewhat hinted at currently).
 >
GF/MW: We could certainly add text, how about adding something like:
 >
After:
A network device should therefore not remark an ECT(0) or ECT(1) mark to 
zero
Add:
  This can result in packets that were originally marked as
  ECN-capable being dropped by ECN-capable network devices further
  along the path. This eliminates the advantage of enabling ECN.
 >
And after:
   Such a packet has already received ECN
   treatment in the network, and remarking it would then hide the
   congestion signal from the endpoints.
Add:
   The endpoint will then be unaware of the signalled congestion, and
   will only trigger a congestion response later if the congested
   queue drops a packet. Note that ECN-enabled
   networks devices need to drop excessive traffic [Section
   4.2.1 of RFC2309.bis] even when this is marked as ECN-capable.
   This eliminates the advantage of enabling ECN, and could slow
   down the response to congestion compared to using AQM, but is not
   expected to be worse than drop-tail performance.
 >
 >Generally, I support the document, it's useful, provides good insight,
and is valuable to the community. Thanks for writing it!
 >
 >--
 >Mikael Abrahamsson    email: swmike@swm.pp.se
 >

We plan another revision of this document based on the received comments.

Please let us know if we have understood correctly.

Gorry & Michael


From nobody Tue Apr 28 05:59:11 2015
Return-Path: <gorry@erg.abdn.ac.uk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id E3CCE1A9100 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 05:59:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.911
X-Spam-Level: 
X-Spam-Status: No, score=-3.911 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id BWjTfI5BkEV4 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 05:59:02 -0700 (PDT)
Received: from pegasus.erg.abdn.ac.uk (pegasus.erg.abdn.ac.uk [139.133.204.173]) by ietfa.amsl.com (Postfix) with ESMTP id 170721A90F7 for <aqm@ietf.org>; Tue, 28 Apr 2015 05:59:02 -0700 (PDT)
Received: from gorry-mac.erg.abdn.ac.uk (unknown [IPv6:2001:630:241:207:21f:5bff:fe38:7354]) by pegasus.erg.abdn.ac.uk (Postfix) with ESMTPSA id 1366D1B00519; Tue, 28 Apr 2015 13:59:15 +0100 (BST)
Message-ID: <553F8414.7050608@erg.abdn.ac.uk>
Date: Tue, 28 Apr 2015 13:59:00 +0100
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Organization: The University of Aberdeen is a charity registered in Scotland,  No SC013683. 
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: =?windows-1252?Q?Mirja_K=FChlewind?= <mirja.kuehlewind@tik.ee.ethz.ch>, aqm@ietf.org,  Michael Welzl <michawe@ifi.uio.no>
References: <55392BD6.501@tik.ee.ethz.ch>
In-Reply-To: <55392BD6.501@tik.ee.ethz.ch>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/NoBlrZ-Qosq9GYNGD4clokcNZm8>
Subject: Re: [aqm] Review of  draft-ietf-aqm-ecn-benefits-03
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: gorry@erg.abdn.ac.uk
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 12:59:09 -0000

Dear Mirja,

Thank you very much for your detailed review! Answers below:

 > Begin forwarded message:
 >
 > Date: 23. april 2015 kl. 19.28.54 CEST
 > From: Mirja Kühlewind <mirja.kuehlewind@tik.ee.ethz.ch>
 > To: <aqm@ietf.org>, Michael Welzl <michawe@ifi.uio.no>, Gorry 
Fairhurst <gorry@erg.abdn.ac.uk>
 > Subject: Review of draft-ietf-aqm-ecn-benefits-03
 >
 > Hi Gorry, hi Michael,
 >
 > as promised here is my review of draft-ietf-aqm-ecn-benefits-03.
 >
 > My overall comment is that even after reading the document (or even 
slightly more than before) I'm not completely sure what the purpose of 
this document is and also what the audience is this documented is 
directed to. Currently this document seems to do two things: 1. it list 
benefits (which is interesting for someone who thinks about enabling 
ECN) and 2. it kind of outlines needed steps for deployment (which would 
be directed to someone who gets the task from his manager to turn on 
ECN). However, the second point is not clearly spelled out and therefore 
it might be rather confusing for some people to read the second part of 
the document. Also the second part is to some extend still 
work-in-progress, therefor I would recommend to only focus this document 
on the first part.
 >
 > For the first part (listing benefits) it might also be good to make 
clear/distinguish who has these benefits. I think all benefits that are 
currently listed are only advantageous for the end host/application. Are 
there any benefits for a network operator? Would it be possible to write 
this document such that I could also use it to point network operators 
to and give them an incentive to enable ECN?

MW/GF: we can only think of one benefit of ECN as currently defined 
(i.e. without basing it on ConEx documents) that obviously targets the 
network operator: making incipient congestion visible (such that it 
could be used e.g. for ConEx). This is addressed in section 3.5. Since 
this is one out of six listed benefits in the table, creating categories 
for end host / application vs. network operator seems unnecessary to us.


 > Another high level comment is that you say in the introduction that 
this document "also identifies some potential problems that might occur 
when ECN is used" but then you don't really discuss them. I think to 
show both sides of the coin in this document would make the document 
more useful (and more honest). One point that you mention slightly here 
is that cheating is easier than with loss by not providing the feedback. 
Another point might be fairness between ECN and non-ECN traffic as 
marking will not reduce the queue length and therefore might lead to a 
higher loss rate for the non-ECN traffic instaed. I guess there are 
papers about this; don't have any by hand right now. Are there any other 
problems that should be mentioned?

MW/GF: This was discussed, and we agreed to remove the "drawbacks" 
discussion, to align with the original proposed work. So, we will remove 
this sentence from the introduction (it is in fact a left-over that 
should have been removed before). As for fairness, it seems to us that 
the related thread has concluded without a clear result. In the absence 
of evidence or references we prefer to stay away from hand-waving about 
this matter in the document.

 > Find more detailed comment by section below:
 >
 > Abstract
 > --------
 > ...says "...potential benefits when applications enable Explicit 
Congestion Notification (ECN)" -> usually an application cannot able ECN 
because usually it's a system setting...?

MW/GF: Good catch! We'll rephrase this as "..when enabling".


 > Section 1
 > ---------
 > ... says "..separate
 >   configuration of the drop and mark thresholds is known to be
 >   supported in some network devices and this is recommended
 >   [RFC2309.bis]."
 > RFC2309bis does not recommend different settings, it only say that it 
should be possible have different configuration of both. Further, I 
think this should not only concern THE threshold (whatever this is) but 
usually there are several parameters you might want to set independent 
of each other, e.g. the max mark/drop probability in RED.

MW/GF: Suggested update:
"While it has often been assumed
that network devices should CE-mark packets at the same level of
congestion at which they would otherwise have dropped them, separate
configuration of the drop and mark conditions. Such separate 
configuration is
known to be supported in some network devices and this is recommended
[RFC2309.bis]."

 > Section 2
 > ----------
 > 1) I'm not sure I understand the purpose of this section or maybe 
just the title is wrong. I'm currently seeing this section rather as a 
section that provides the needed background knowledge than is talking 
about deployment. For this purpose I'd put all references and 
potentially a brief summary to other RFC/drafts on ECN in this section 
including RFC2884, RFC4774, RFC5562, RFC6040, RFC6679, 
draft-briscoe-tsvwg-ecn-encap-guidelines and draft-ietf-tcpm-accecn-reqs 
(and rename it).

MW/GF: This section lists requirements for deployment. Suggestion: 
rename to "ECN deployment requirements"

 >
 > 2) Second paragraph says:
 > "Network devices must not drop packets solely because these 
codepoints are used [RFC2309.bis]."
 > Not sure this is the right document to says this (because currently 
it not seems to be directed to network operator/equipment vendors but 
admins/application developers). However, if it says this, it should also 
say that network devices should not bleach these bits.

MW/GF: suggest: "Network devices must not drop packets solely because 
these codepoints are used or erase these codepoints [RFC2309.bis]."


 > 3) First bullet in list says
 > "A recent survey reported growing support for ECN on common network 
paths [TR15]."
 > This sounds like TR15 shows that ECN is actually used in the 
Internet. However, TR15 only shows that there are only very few cases 
left where ECN packets are dropped or incorrectly altered. Please 
clarify or remove this sentence here.

MW/GF: suggest: "A recent survey reported that incorrect altering of ECN 
bits or consistent dropping of packets carrying the ECN codepoint is 
rare on common network paths [TR15]."


 > 4) You could cite draft-bensley-tcpm-dctcp-00 instead of the DCTCP 
Sigcomm paper (or both).

MW/GF: The paper is a stable reference for now.
But if/when the IETF decides on this, we can add a reference.


 > 5) I would remove the subsection headings (both 2.1 and 2.2) and just 
add the text there to the main part of the section.

MW/GF: OK


 > 6) "An AQM algorithm that supports ECN needs to define
 >   the threshold and algorithm for ECN-marking."
 > This is kind of self-redundant and therefore does not really makes 
sense to me to say; of course an algo that supports ECN needs to say 
something about ECN...

MW/GF: We agree, but suggest to keep it nevertheless, it is a hint to 
document authors to not forget that they should specify ECN rather than 
just assuming some default behaviour.


 > 7) You can use TR15 to provide a reference for the first paragraph in 
section 2.2:
 > "Cases have been noted where a sending endpoint marks a packet with a
 >   non-zero ECN mark, but the packet is received with a zero ECN value
 >   by the remote endpoint."

MW/GF: OK, will add the reference there


 > 8) I'd move the second paragraph of section 2.2. ("The current..") to 
a potentially new problems section, talking about known/previous 
deployment problems.

MW/GF: the document does not accentuate problems in this way, as a 
result of prior discussion. We therefore think that this paragraph is ok 
in its current place.


 > 9) I would simply remove paragraph 3-4 of section 2.2 because this 
was basically as already mentioned by referring to 2309bis and rfc6040 
in section 2.1.

MW/GF: we do think these paragraphs add value here: they describe the 
problem in greater detail than the text before, explaining the problem 
here is different - and, we think, better - than just pointing to 
references.


 > Section 3.2
 > ------------
 > 1) Don't understand why there is a listing here...? Just remove the 
listing and make text out of it...?

MW/GF: This is to help identify the entities that need to collaborate.


 > 2) The sentence "This also
 >      avoids the inefficiency of dropping data that has already made it
 >      across at least part of the network path."
 > does not belong in this section. This sentence should just be moved 
to section 3.1 (or in an own section) and must be further explained, 
saying that dropping packet at the of the path has already blocked 
resources that other traffic could have used otherwise.

MW/GF: agreed. We will insert it at the beginning of section 3.1.


 > Section 3.3
 > -----------
 > 1) I'd say this section misses on part of the discussion. It is true 
that if by chance your last  packet(s) get lost ECN can help. However, 
this section reads a little like, with ECN it is save to send packet 
bursts. Which is not true because even if ECN is used by a network 
device, the queue might be too small to hold the whole burst. I believe 
this case happen very often which might be a reason for the higher tail 
loss probability that sometimes is experienced with IW10. Please add 
this point to the discussion.

MW/GF: we agree that we shouldn't say that "with ECN it is ok to send 
packet bursts" - we want to stay away from such general recommendations 
and just state the potential benefit of ECN when it saves the last 
packet of a burst. See our next comment for more:


 > 2) I don't really get the point of the second paragraph. First of all 
it is confusion that this paragraph starts which "In addition to 
avoiding HOL blocking,.."; I guess that is left over from a previous 
version of this text...? And then you talk about a connection that is 
currently idle, so why is the performance of this connection that is 
currently not sending anything reduced?

MW/GF: indeed it seems that this paragraph has been mangled during 
updates. To address your item 1 and 2, we suggest the following replacement:

***
"While using ECN can never guarantee loss prevention, and thus losses
at the end of a burst can occur with or without ECN, using ECN can increase
the chance for that last packet to be ECN-marked instead of dropped. 
This can allow the
transport to avoid the consequent loss of state about the network path it is
    using, which would have arisen had there been a retransmission
    timeout.  Typical impacts of a transport timeout are to reset path
    estimates such as the RTT, the congestion window, and possibly other
    transport state that can reduce the performance of the transport
    until it again adapts to the path."
***


 > 3) I don't understand what "applications that send intermittent 
bursts of data, and rely upon timer-based recovery of packet loss" 
are...? Isn't the transport responsible to not send bursts and care 
about recovery...?

MW/GF: MPEG-DASH traffic for instance, in particular when used over 
non-paced TCP. UDP-based applications too.


 > 4) For the last paragraph in section 3.3 note that stacks often 
remember RTT measurements for a certain IP address and set the initial 
RTO based on this information.

MW/GF: suggestion: replace:
***
because in this
    case TCP cannot base the timeout period on prior RTT measurements
    from the same connection.
***
with:
***
because in this
    case TCP may not be able to base the timeout period on prior RTT 
measurements.
***


 > Section 3.4
 > -----------
 > You still need FEC or some kind of error concealment even if ECN is 
used because you can never be sure that your packet are not get dropped 
(by non-ECN-enable devices or other reasons). Therefore using ECN will 
clearly not reduce complexity. The only thing you can do is to 
potentially reduce the amount of redundancy you send if you know that a 
certain path is ECN enables or don't see losses at the beginning of a 
connection. This can save network resources but actually might not 
improve user experience; in fact the user experience might be worse in 
case there are sudden losses.

MW/GF: suggestion: remove "add complexity and"


 > Further the text says "negative impact of using loss-hiding 
mechanisms"; I don't really think that FEC has a negative impact as long 
as you've send enough redundancy...? Error concealment might but is used 
less and less. I'd recommend to talk about error concealment only in 
this last paragraph and explain a little further.

MW/GF: error concealment is different from FEC, and it is only mentioned 
in this last paragraph. We suggest to replace "Because this
    reduces the negative impact of using loss-hiding mechanisms,"   with 
"Because this can reduce the potential negative impact that some 
loss-hiding mechanisms can have,"


 > Section 3.5
 > ----------
 > "Recording the presence of CE-marked packets can therefore provide
 >   information about the performance of the network path."
 > Would change to:
 > "Recording the presence of CE-marked packets in absence of loss can 
therefore provide
 >   information about the performance of the network path."

MW/GF: ok


 > And also say more concretely what is meant with 'performance of the 
network path' -> congestion level or no drops by other middleboxes on 
this path...

MW/GF: This intentionally was kept this vague, but we'd welcome a 
concrete recommendation by a ConEx expert      (indeed "or .. or ..." is 
the problem, there are several possibilities here)


 > Section 3.6
 > -----------
 > 1) I like the section but I would phrase it differently; also it's 
not clear who needs to support what in this case. I'd like to propose 
the following text [not sure about the heading...]:
 >
 > "3.6 Opportunity to provide an improved congestion feedback signal
 >
 > Loss and ECN marking are both used as an indication for congestion. 
However, while the amount of feedback that is provided by loss should 
naturally be minimized, this is not the case for ECN. With ECN a network 
node could provide richer and more frequent feedback on the congestion 
state of a link which then could be used by the control mechanisms 
implemented in end host to make a more appropriate decision on how to 
react to congestion and to react faster to changes in congestion state.

MW/GF: ok to add this up to here.


 > Further while drop-based AQM mechanisms usually operate on a smoothed 
queue length estimation (instead of the instantaneous queue length) and 
therefore slightly delay the feedback signal to avoid unnecessary losses 
in case of transient congestion, this would be not necessary for ECN. If 
congestion is only transient due to short traffic bursts that are active 
for less than one RTT, the congestion signal would reach the sender at a 
time where the congestion is already cleared up. However, instead 
delaying the feedback in the network, the end host could reduce its 
sending rate incrementally based on the extend of congestion (that was 
experienced over e.g. the last RTT) similar as DCTCP. In case if the 
congestion is only transient, the end host would only reduce its rate 
slightly and be able to catch up quickly again. However, in case the 
congestion is persistent, this would help to remove additional delays 
from the network and resolve congestion faster which after all reduces 
the average queuing delay.
 >
 > However, current ECN is defined as a 'drop equivalent' in RFC3168. To 
change the semantics of ECN both the AQM in the network nodes and the 
control mechanism in the end hosts would still need to cope with nodes 
or end hosts that rely on the old semantics. Therefore changing the 
semantics can be done more easily in confined environment such as a data 
center. DCTCP is an example that changes both the configuration of the 
used AQM as well as the congestion response in the end host and relies 
on that fact that all nodes in data center are configured the same way. 
[Deployment strategies to change the semantics of ECN in the Internet 
are currently under discussion in the IETF.]"

MW/GF: We think that this goes a bit too far in the direction of hinting 
about implementation and research possibilities that we don't have 
citable proof about (besides: we already refer to DCTCP twice in the 
document, and the 'drop equivalent' semantics are not a MUST in RFC3168).


 > 2) I'd move the 1. and 2. paragraph of section 3.6.1 to the 
background/deployment section or to the intro depending what you going 
to do with section 2.

MW/GF: since we intend section 2 to be about deployment requirements 
only, we don't think this fits and would rather leave these paragraphs 
in section 3.6.1.


 > Sections 4 & 5
 > ---------
 > First sentence talks about "operational
 >   difficulties when the network only partially supports the use of ECN,
 >   or to respond to the challenges due to misbehaving network devices
 >   and/or endpoints".
 > I think these are to very different things. Misbehaving network 
devices is a point for a  problems section (where the lesson learned is 
that we didn't think carefully enough about incremental deployment in 
the first place but do now). However, partial deployment is not a 
problem but is a thing we simply have to cope with. The text sound as if 
the goal would be that every router in the whole Internet would at some 
point of time be ECN-enabled. I don't think this will ever happen and is 
also not the goal for me. Routers that are very unlikely to ever get 
congested should no be required to look at the ECN bits or monitor the 
queue length to calculate a mark/drop probability.

MW/GF: we agree, and suggest to replace this sentence with "Early 
deployment of ECN encountered a number of operational
    difficulties due to misbehaving network devices
    and/or endpoints."


 > However as I said at the beginning I don't really thing that sections 
4 and 5 belong in this document. If you decided to keep them (you have 
to change the abstract) and I'd recommend to rename them e.g 4. 
'Incremental Deployment Strategy' or 'Requirements to enable Incremental 
Deployment' and 5. 'Recommendations for enabling ECN in network nodes 
and end hosts'.

MW/GF: we suggest to insert the fact that we discuss deployment in the 
abstract, and rename these sections to 4.: "Incremental Deployment" and 
"Recommendations for enabling ECN"


 > I hope that's helpful! Let me know if you have any questions!
 >
 > Mirja
 >
 >


Thank you very much,

Michael & Gorry



From nobody Tue Apr 28 06:35:07 2015
Return-Path: <john@jlc.net>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 75F681AC3CA for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 06:35:05 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.21
X-Spam-Level: 
X-Spam-Status: No, score=-4.21 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 2NIz4Z78k89c for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 06:35:03 -0700 (PDT)
Received: from mailhost.jlc.net (mailhost.jlc.net [199.201.159.4]) by ietfa.amsl.com (Postfix) with ESMTP id 46BC81AC3C8 for <aqm@ietf.org>; Tue, 28 Apr 2015 06:35:03 -0700 (PDT)
Received: by mailhost.jlc.net (Postfix, from userid 104) id 22C8CC94A9; Tue, 28 Apr 2015 09:34:51 -0400 (EDT)
Date: Tue, 28 Apr 2015 09:34:51 -0400
From: John Leslie <john@jlc.net>
To: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Message-ID: <20150428133451.GQ63465@verdi>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <553F7F15.9060509@erg.abdn.ac.uk>
User-Agent: Mutt/1.4.1i
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/pzYJKFenxyLCmhIF0IiOlB3yfDk>
Cc: "aqm@ietf.org" <aqm@ietf.org>, Mikael Abrahamsson <swmike@swm.pp.se>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 13:35:05 -0000

Gorry Fairhurst <gorry@erg.abdn.ac.uk> wrote:
> To: Mikael Abrahamsson <swmike@swm.pp.se>, "aqm@ietf.org" <aqm@ietf.org>
> 
>> My summary: I would like to see some text on the downsides of having
>> ECN transport, ECN enabled routers and having middle box bleaching
>> ECN bits, thus removing the congestion signal. This downside, and the
>> operational problem that some old equipment just simply drop packets
>> with ECN bits set should be summarized the same way as the benfits.
>> These are the only drawbacks of enabling ECN that I can see, and
>> potentially text describing how to detect these problems could be
>> added/extended (it's somewhat hinted at currently).
> 
> GF/MW: We could certainly add text, how about adding something like:
> 
> After:
> A network device should therefore not remark an ECT(0) or ECT(1) mark to 
> zero
> Add:
>  This can result in packets that were originally marked as
>  ECN-capable being dropped by ECN-capable network devices further
>  along the path. This eliminates the advantage of enabling ECN.

   This text is a bit confusing -- I had to read it three times to
get (I think) your point. Perhaps:
" 
" This would negate the advantages of ECN for congestion which may
" occur later in the path, in that an ECN-capable network device would
" not be able to forward the packet with congestion signaled by "CE".

   Putting the effect first seems easier to read...

   Also, the forwarding node later in the path might have signaled
congestion via "CE" at an earlier threshhold than it would drop
packets.

> And after:
>   Such a packet has already received ECN
>   treatment in the network, and remarking it would then hide the
>   congestion signal from the endpoints.
> Add:
>   The endpoint will then be unaware of the signalled congestion, and
>   will only trigger a congestion response later if the congested
>   queue drops a packet. Note that ECN-enabled
>   networks devices need to drop excessive traffic [Section
>   4.2.1 of RFC2309.bis] even when this is marked as ECN-capable.
>   This eliminates the advantage of enabling ECN, and could slow
>   down the response to congestion compared to using AQM, but is not
>   expected to be worse than drop-tail performance.

   I'm still failing (after five readings) to extract a clear meaning. :^(

   There are several points:

- clearing a "CE" marking destroys the congestion signal. Barring some
  extraordinary programming, this would make an ECN-enabled flow
  unresponsive.

- with an unresponsive flow, the congested node _might_ reach a level
  of congestion requiring tail-drop. But it also might not. :^(

- _if_ a receiver added a way to observe this abuse (possibly by
  having the sender set an occasional "CE" mark before sending the
  packet), the responsiveness would still be degraded. Besides,
  there's really no chance that _all_ flows would ever be programmed
  to do this.

   Myself, I see no reason to add text to cover all of these: I think
text covering the first point is sufficient.

--
John Leslie <john@jlc.net>


From nobody Tue Apr 28 06:49:49 2015
Return-Path: <swmike@swm.pp.se>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id C0CB11AC432 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 06:49:48 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.661
X-Spam-Level: 
X-Spam-Status: No, score=-1.661 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_SE=0.35, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id C9ntxbSj740X for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 06:49:45 -0700 (PDT)
Received: from uplift.swm.pp.se (ipv6.swm.pp.se [IPv6:2a00:801::f]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 3CDAA1AC40D for <aqm@ietf.org>; Tue, 28 Apr 2015 06:49:45 -0700 (PDT)
Received: by uplift.swm.pp.se (Postfix, from userid 501) id F1B57A1; Tue, 28 Apr 2015 15:49:42 +0200 (CEST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=swm.pp.se; s=mail; t=1430228982; bh=2DWxN3Utupg1PiTWzGfQVvCXq2qxWC2aHfCTOMxyRf0=; h=Date:From:To:cc:Subject:In-Reply-To:References:From; b=g4qYbQGeVwTo3KKl9/B3ESrs3wynh/hk2aa0I40sIyGXPCwrZ5IvqRh7OpwJK04Vg NYs2L2Hbsk1HYu07v4fLiXRl+IKbAB8e9etvSXQLq8ycuFRlelgutHzdKW6ZdIXPRm zNL2/2BOvviX/P9C82snQ2KQl8ZNojeRRNZsia80=
Received: from localhost (localhost [127.0.0.1]) by uplift.swm.pp.se (Postfix) with ESMTP id E817B9F; Tue, 28 Apr 2015 15:49:42 +0200 (CEST)
Date: Tue, 28 Apr 2015 15:49:42 +0200 (CEST)
From: Mikael Abrahamsson <swmike@swm.pp.se>
To: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
In-Reply-To: <553F7F15.9060509@erg.abdn.ac.uk>
Message-ID: <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
Organization: People's Front Against WWW
MIME-Version: 1.0
Content-Type: MULTIPART/MIXED; BOUNDARY="-137064504-2051917898-1430228982=:16871"
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/fWnmqJ7rtZeSw9mPkBBc8lfJrJo>
Cc: "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 13:49:48 -0000

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

---137064504-2051917898-1430228982=:16871
Content-Type: TEXT/PLAIN; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8BIT


Hi,

it's quite hard to read the below email, considering how it's quoted.

I think I need to see the new document to be able to tell if the changes 
works or not.

Also, I actually still don't understand the concept of "developing 
congestion". For me "congestion" means there is a sustained, constant, 
non-trivial buffer fill. So "incipient congestion" doesn't compute at all. 
I understand this term has been used before, but that doesn't help me 
understand the reasoning behind the wording. For me sending a congestion 
signal, means saying there is congestion. It's not saying there is 
developing congestion (whatever that might be).

I would greatly appreciate if someone could point me to text explaining 
this so I can better understand the reasoning behind this wording.

For instance from the linked early.pdf:

"The goal of the gateway is to detect incipient congestion that has 
persisted for a long time (several roundtrip times)."

So we have "developing congestion" that has "persisted for a long time". I 
don't understand this wording at all. If it has persisted for a long time, 
then it's not "developing" anymore, or is there some hidden meaning to 
"incipient" that I don't understand?

On Tue, 28 Apr 2015, Gorry Fairhurst wrote:

> It is great to see some detailed comments from you looking afresh at this. 
> Please see below for initial replies.
>
>> On Fri, 24 Apr 2015, Wesley Eddy wrote:
>>
>> To keep moving forward with the set of documents, we'd like to start
>> a Working Group Last Call on:
>> http://datatracker.ietf.org/doc/draft-ietf-aqm-ecn-benefits/
>>>
>> Please send any comments you have on this document to the list
>> by May 8th.  Other than those raised in the thread Mirja has just
>> opened with her review, I'm not aware of any other open issues.
>>
>> I have re-read this document as if I never read it before. Writing my
>> thoughts as I read it.
>>
>> The title says "benefits" but the abstract states that the document will
> talk about both benefits and potential problems. I therefore think the
> title does not reflect the abstract.
>>
> GF/MW: We looked at both approaches, and revised to include "drawbacks" as
> well, but then the list discussion seemed to favour focusing on the
> "benefits", since deployment issues could change with time. I suggest we
> remove the sentence about "problems" from the abstract.
>>
>> 1. Introduction: I don't understand the "(potential)", doesn't transport
> always treat loss and CE marking as indications of congestion?
> Actually, the way the introduction reads today one might interpret that
> TCP and SCTP are "Internet transports" but UDP is not, because it's
> merely "transport".
>>
> GF/MW: We need to add REF to AQM recommendations [RFC2309.bis].
>>
> GF/MW:  We'd agree, and be happy to remove "(potential)" from section 1
> para 1.
>>
>> Can we have a reference to AQM in the second paragraph? What is the
>> difference between a router and a middle box?
>>
> GF/MW: This echoes what is said in the AQM Recommendations draft: "The
> mechanisms described in this document may be implemented in
> network devices on the path between end-points that include routers,
> switches, and other network middleboxes." - The difference may be in the
> eyes of the beholder - all are network devices that can deploy AQM.
>>
>> I don't like the first sentence in the 3rd paragraph, shouldn't it say
>> that the device "might mark instead of drop", and not say "as the queue
>> builds"? Because if we have a cisco-style policer, there is no queue but
> instead a token-bucket rate limiter. I don't like "as the network queue
> builds" anyhow, if it should stay, it should be rephrased to "interface
> queue size increases" or something like that.
>>
> GF/MW: I agree policiers and other methods can track congestion state and
> set a CE-mark. I'm OK with some change to "queue builds", but do not yet
> understand the difference between "interface queue size increases", except
> that would seem to limit only to an interface. It is OK with us to remove
> "as the queue builds".
>>
>> What if there are several queues? I'd like the last part to not be
> included at all.
>>
>> I suggest we add this sentence after the next para also, to balance this:
>> "An endpoint that receives a congestion mark
>> causes a congestion control reaction at the sender to reduce
>> the maximum rate permitted by the endpoint."
>>
>> Also, is the  document really saying that ECN should be done before
> incurring loss for non-ECN traffic? The current wording can be
> interpreted as such.
>>
> GF/MW: It enables marking before the point where a drop-tail queue would
> have dropped. The ID intentionally does not describe how to implement or
> to evaluate if this is done "fairly". The suggested sentence seems to be
> unnecessary because it echoes what is in RFC3168, to some degree, but also
> uses phrasing that can be misleading, by saying: "maximum rate permitted".
>>
>> Fourth pararaph. I have no idea what "incipient" means. I looked it up.
> Can we use the word "developing" or some other synonym?
>>
> GF/MW: OK this came from the list discussion. This term has already been
> used in early discussion of RED marking (e.g.,
> ftp://ftp.ee.lbl.gov/papers/early.pdf) It was used in RFC 2481, and is
> used several times in RFC 3168. I suggest we add some explanatory words
> (e.g., "where congestion is beginning to be experienced")
>>
>> 2.2. Here we have incipient again, and isn't drop or CE marking an
>> indicator of actual congestion, not just developing/incipient congestion?
>>
> GF/MW: See above, one advantage of talking about "incipient" congestion is
> we avoid talking about "mild" or "initial" or some other term that could
> perhaps be misconstrued. It is is congestion, but reported when first
> discovered.
>>
>> How do we stop older bleaching-boxes that do not understand ECN bits and
> still bleach them, from just forwarding the packet even if CE=1? I
>> understand the must requirement and it makes sense, but I don't see how
>> this can be achieved with older equipment?
>>
> GF/MW: We don't see more options than:
> * Declare them as not complying with current use
> * Go and fix those that can be fixed, and live with the others until in
> the future they are finally replaced.
> - I'm not sure what extra there is to say about this, apart from to
> verify the operation across the path (as in section 4.1).
>>
> GF/MW: The phrase:
> "A network device should therefore not remark an ECT(0) or ECT(1) mark to
> zero [RFC2309.bis]."
> - This deployment issue is consistent with the intention of RFC 2309.bis,
> but is not specifically called-out in RFC2309.bis, and this reference
> needs to be removed.
>>
>> 3.1. I would like to see text after 0.02 that says something like (1
>> packet in 50) or something like that, just for clarification, or that
> the  packet drop rate is 2%. It's not obvious that "the packet drop
> ratio is 0.02" means these things.
>>
> GF/MW: We think this change should be made.
>>
>> My summary: I would like to see some text on the downsides of having ECN
> transport, ECN enabled routers and having middle box bleaching ECN
> bits, thus removing the congestion signal. This downside, and the
> operational problem that some old equipment just simply drop packets
> with ECN bits set should be summarized the same way as the benfits.
> These are the only drawbacks of enabling ECN that I can see, and
> potentially text describing how to detect these problems could be
> added/extended (it's somewhat hinted at currently).
>>
> GF/MW: We could certainly add text, how about adding something like:
>>
> After:
> A network device should therefore not remark an ECT(0) or ECT(1) mark to zero
> Add:
> This can result in packets that were originally marked as
> ECN-capable being dropped by ECN-capable network devices further
> along the path. This eliminates the advantage of enabling ECN.
>>
> And after:
>  Such a packet has already received ECN
>  treatment in the network, and remarking it would then hide the
>  congestion signal from the endpoints.
> Add:
>  The endpoint will then be unaware of the signalled congestion, and
>  will only trigger a congestion response later if the congested
>  queue drops a packet. Note that ECN-enabled
>  networks devices need to drop excessive traffic [Section
>  4.2.1 of RFC2309.bis] even when this is marked as ECN-capable.
>  This eliminates the advantage of enabling ECN, and could slow
>  down the response to congestion compared to using AQM, but is not
>  expected to be worse than drop-tail performance.
>>
>> Generally, I support the document, it's useful, provides good insight,
> and is valuable to the community. Thanks for writing it!
>>
>> --
>> Mikael Abrahamsson    email: swmike@swm.pp.se
>>
>
> We plan another revision of this document based on the received comments.
>
> Please let us know if we have understood correctly.
>
> Gorry & Michael
>

-- 
Mikael Abrahamsson    email: swmike@swm.pp.se
---137064504-2051917898-1430228982=:16871--


From nobody Tue Apr 28 07:00:37 2015
Return-Path: <vm2020@columbia.edu>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 976E41ACDCB for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 07:00:35 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.21
X-Spam-Level: 
X-Spam-Status: No, score=-4.21 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RkgEycfQOq2h for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 07:00:29 -0700 (PDT)
Received: from buckwheat.cc.columbia.edu (buckwheat.cc.columbia.edu [128.59.72.251]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 9C6721ACDC8 for <aqm@ietf.org>; Tue, 28 Apr 2015 07:00:27 -0700 (PDT)
Received: from hazelnut (hazelnut.cc.columbia.edu [128.59.213.250]) by buckwheat.cc.columbia.edu (8.13.8/8.13.8) with ESMTP id t3SDxMBD027858 for <aqm@ietf.org>; Tue, 28 Apr 2015 10:00:26 -0400
Received: from hazelnut (localhost.localdomain [127.0.0.1]) by hazelnut (Postfix) with ESMTP id 73FFE6D for <aqm@ietf.org>; Tue, 28 Apr 2015 10:00:26 -0400 (EDT)
Received: from paneer.cc.columbia.edu (paneer.cc.columbia.edu [128.59.29.4]) by hazelnut (Postfix) with ESMTP id 3E4BE81 for <aqm@ietf.org>; Tue, 28 Apr 2015 10:00:26 -0400 (EDT)
Received: from mail-vn0-f51.google.com (mail-vn0-f51.google.com [209.85.216.51]) by paneer.cc.columbia.edu (8.14.4/8.14.3) with ESMTP id t3SE0PiI015028 (version=TLSv1/SSLv3 cipher=RC4-SHA bits=128 verify=NOT) for <aqm@ietf.org>; Tue, 28 Apr 2015 10:00:26 -0400 (EDT)
Received: by vnbf1 with SMTP id f1so16140818vnb.5 for <aqm@ietf.org>; Tue, 28 Apr 2015 07:00:25 -0700 (PDT)
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:content-type:mime-version:subject:from :in-reply-to:date:cc:message-id:references:to; bh=sSBwYGrfH/frohyRk4ODaWSz19cpJlk3gxq5QT3RYp8=; b=g6KlisY8EmvMEZhTqH5abrpBM/BwDJYbdSS4X/+8sQHA92UoZEuie53mt7S11IAe+Q pTdxbjmJK8WjAhYr4ezSqziH0smgZk0VqxroQnMzzRYNa09/DFR6nDQR+IOjVAgZYh7J BCnQLZis1lGuEQikKx8FO+xfhPXeXK3fcGW20qxjJMxhNEFAVbRtwDd8ynsi2OAahTCa FWLDJfJcEsV3FNXN44WhQrMc3rlWSSzES+pLy1hqG56V8OaE52UospAOotAcASh6XvLD 6GjOLcjRrv0hN1Xrmx1ZU98Z05OXFSBbGiIXEIzWaQA+yvoqanE2jj7TxJBuwGMMxF/u IGjA==
X-Gm-Message-State: ALoCoQli9oFrSmKQk3rVkr3EhaKkR1EOtwNWR15tIZgKjk1tYMS/H8lzPoGfoRLQSD0ta2AuwQeNVuxodspJ6MUquZTtE+KfR+WmVnmApCZSrv0PCZbXcBgB+IL4ZR3RWRb+9a2RKsW8
X-Received: by 10.52.163.81 with SMTP id yg17mr35904634vdb.97.1430229625473; Tue, 28 Apr 2015 07:00:25 -0700 (PDT)
X-Received: by 10.52.163.81 with SMTP id yg17mr35904555vdb.97.1430229624992; Tue, 28 Apr 2015 07:00:24 -0700 (PDT)
Received: from ?IPv6:2001:18d8:ffff:16:97d:35f5:cc82:f9f8? ([2001:18d8:ffff:16:97d:35f5:cc82:f9f8]) by mx.google.com with ESMTPSA id ef10sm31639827vdc.4.2015.04.28.07.00.24 (version=TLSv1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Tue, 28 Apr 2015 07:00:24 -0700 (PDT)
Content-Type: multipart/alternative; boundary="Apple-Mail=_DC04AA4E-F320-47A4-AEFC-6D4A14596C07"
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2098\))
From: Vishal Misra <misra@cs.columbia.edu>
In-Reply-To: <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se>
Date: Tue, 28 Apr 2015 10:00:23 -0400
Message-Id: <A0E0ADC3-FB07-4CA2-BC21-4C0276DFB33A@cs.columbia.edu>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk> <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se>
To: Mikael Abrahamsson <swmike@swm.pp.se>
X-Mailer: Apple Mail (2.2098)
X-No-Spam-Score: Local
X-Scanned-By: MIMEDefang 2.68 on 128.59.29.4
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/eHUf4x3gxTIoWY2oBsh_Us8JJ88>
Cc: Gorry Fairhurst <gorry@erg.abdn.ac.uk>, "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 14:00:35 -0000

--Apple-Mail=_DC04AA4E-F320-47A4-AEFC-6D4A14596C07
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=windows-1252

One looks at the derivative of the queue length. If that derivative =
continues in the positive direction, you can proactively send a =
congestion signal and ask the flows to slow down before the queue size =
reaches =93congestion levels=94. It is the theory behind almost all =
control systems, and was the idea behind the design of the PI =
controller.

RED was reactive, it waited for the average queue length to rise above a =
threshold and that introduced too much of a delay. PI is pro-active, it =
predicts onset of congestion and responds quickly.

-Vishal
--
http://www.cs.columbia.edu/~misra/


> On Apr 28, 2015, at 9:49 AM, Mikael Abrahamsson <swmike@swm.pp.se> =
wrote:
>=20
>  So "incipient congestion" doesn't compute at all. I understand this =
term has been used before, but that doesn't help me understand the =
reasoning behind the wording. For me sending a congestion signal, means =
saying there is congestion. It's not saying there is developing =
congestion (whatever that might be).


--Apple-Mail=_DC04AA4E-F320-47A4-AEFC-6D4A14596C07
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=windows-1252

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html =
charset=3Dwindows-1252"></head><body style=3D"word-wrap: break-word; =
-webkit-nbsp-mode: space; -webkit-line-break: after-white-space;" =
class=3D"">One looks at the derivative of the queue length. If that =
derivative continues in the positive direction, you can proactively send =
a congestion signal and ask the flows to slow down before the queue size =
reaches =93congestion levels=94. It is the theory behind almost all =
control systems, and was the idea behind the design of the PI =
controller.<div class=3D""><br class=3D""></div><div class=3D"">RED was =
reactive, it waited for the average queue length to rise above a =
threshold and that introduced too much of a delay. PI is pro-active, it =
predicts onset of congestion and responds quickly.</div><div =
class=3D""><br class=3D""></div><div class=3D"">-Vishal<br class=3D""><div=
 apple-content-edited=3D"true" class=3D"">
<span class=3D"Apple-style-span" style=3D"border-collapse: separate; =
border-spacing: 0px 0px; color: rgb(0, 0, 0); font-family: Helvetica; =
font-size: 12px; font-style: normal; font-variant: normal; font-weight: =
normal; letter-spacing: normal; line-height: normal; text-align: auto; =
-khtml-text-decorations-in-effect: none; text-indent: 0px; =
-apple-text-size-adjust: auto; text-transform: none; orphans: 2; =
white-space: normal; widows: 2; word-spacing: 0px; "><div =
class=3D"">--</div><div class=3D""><a =
href=3D"http://www.cs.columbia.edu/~misra/" =
class=3D"">http://www.cs.columbia.edu/~misra/</a></div><br =
class=3D"Apple-interchange-newline"></span>

</div>
<br class=3D""><div><blockquote type=3D"cite" class=3D""><div =
class=3D"">On Apr 28, 2015, at 9:49 AM, Mikael Abrahamsson &lt;<a =
href=3D"mailto:swmike@swm.pp.se" class=3D"">swmike@swm.pp.se</a>&gt; =
wrote:</div><br class=3D"Apple-interchange-newline"><div class=3D""><span =
style=3D"font-family: Helvetica; font-size: 12px; font-style: normal; =
font-variant: normal; font-weight: normal; letter-spacing: normal; =
line-height: normal; orphans: auto; text-align: start; text-indent: 0px; =
text-transform: none; white-space: normal; widows: auto; word-spacing: =
0px; -webkit-text-stroke-width: 0px; float: none; display: inline =
!important;" class=3D""><span =
class=3D"Apple-converted-space">&nbsp;</span>So "incipient congestion" =
doesn't compute at all. I understand this term has been used before, but =
that doesn't help me understand the reasoning behind the wording. For me =
sending a congestion signal, means saying there is congestion. It's not =
saying there is developing congestion (whatever that might =
be).</span></div></blockquote></div><br class=3D""></div></body></html>=

--Apple-Mail=_DC04AA4E-F320-47A4-AEFC-6D4A14596C07--


From nobody Tue Apr 28 07:34:01 2015
Return-Path: <gorry@erg.abdn.ac.uk>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 44B371A0275 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 07:34:00 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.211
X-Spam-Level: 
X-Spam-Status: No, score=-4.211 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id UFEnPPKOGzUg for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 07:33:53 -0700 (PDT)
Received: from pegasus.erg.abdn.ac.uk (pegasus.erg.abdn.ac.uk [139.133.204.173]) by ietfa.amsl.com (Postfix) with ESMTP id 740F01A0362 for <aqm@ietf.org>; Tue, 28 Apr 2015 07:33:53 -0700 (PDT)
Received: from gorry-mac.erg.abdn.ac.uk (unknown [IPv6:2001:630:241:207:21f:5bff:fe38:7354]) by pegasus.erg.abdn.ac.uk (Postfix) with ESMTPSA id 8DF351B00519; Tue, 28 Apr 2015 15:34:05 +0100 (BST)
Message-ID: <553F9A4F.8060901@erg.abdn.ac.uk>
Date: Tue, 28 Apr 2015 15:33:51 +0100
From: Gorry Fairhurst <gorry@erg.abdn.ac.uk>
Organization: The University of Aberdeen is a charity registered in Scotland,  No SC013683. 
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: John Leslie <john@jlc.net>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk> <20150428133451.GQ63465@verdi>
In-Reply-To: <20150428133451.GQ63465@verdi>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/_zdkE5YkmmQAlkzBN8dAU57T5MM>
Cc: "aqm@ietf.org" <aqm@ietf.org>, Mikael Abrahamsson <swmike@swm.pp.se>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
Reply-To: gorry@erg.abdn.ac.uk
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 14:34:00 -0000

On 28/04/2015 14:34, John Leslie wrote:
> Gorry Fairhurst <gorry@erg.abdn.ac.uk> wrote:
>> To: Mikael Abrahamsson <swmike@swm.pp.se>, "aqm@ietf.org" <aqm@ietf.org>
>>
>>> My summary: I would like to see some text on the downsides of having
>>> ECN transport, ECN enabled routers and having middle box bleaching
>>> ECN bits, thus removing the congestion signal. This downside, and the
>>> operational problem that some old equipment just simply drop packets
>>> with ECN bits set should be summarized the same way as the benfits.
>>> These are the only drawbacks of enabling ECN that I can see, and
>>> potentially text describing how to detect these problems could be
>>> added/extended (it's somewhat hinted at currently).
>>
>> GF/MW: We could certainly add text, how about adding something like:
>>
>> After:
>> A network device should therefore not remark an ECT(0) or ECT(1) mark to
>> zero
>> Add:
>>   This can result in packets that were originally marked as
>>   ECN-capable being dropped by ECN-capable network devices further
>>   along the path. This eliminates the advantage of enabling ECN.
>
>     This text is a bit confusing -- I had to read it three times to
> get (I think) your point. Perhaps:
> "
> " This would negate the advantages of ECN for congestion which may
> " occur later in the path, in that an ECN-capable network device would
> " not be able to forward the packet with congestion signaled by "CE".
>
>     Putting the effect first seems easier to read...
>
>     Also, the forwarding node later in the path might have signaled
> congestion via "CE" at an earlier threshhold than it would drop
> packets.
>
I agree and so we will improve the words.

However, I'm not keen on delving into what happens if (and when) 
thresholds differ (or the AQM algorithm is different).

>> And after:
>>    Such a packet has already received ECN
>>    treatment in the network, and remarking it would then hide the
>>    congestion signal from the endpoints.
>> Add:
>>    The endpoint will then be unaware of the signalled congestion, and
>>    will only trigger a congestion response later if the congested
>>    queue drops a packet. Note that ECN-enabled
>>    networks devices need to drop excessive traffic [Section
>>    4.2.1 of RFC2309.bis] even when this is marked as ECN-capable.
>>    This eliminates the advantage of enabling ECN, and could slow
>>    down the response to congestion compared to using AQM, but is not
>>    expected to be worse than drop-tail performance.
>
>     I'm still failing (after five readings) to extract a clear meaning. :^(
>
We will look at simpler words.

>     There are several points:
>
> - clearing a "CE" marking destroys the congestion signal. Barring some
>    extraordinary programming, this would make an ECN-enabled flow
>    unresponsive.
>
It would be less responsive than an ECN-enabled flow should be and could 
congest a queue - perhaps negating all benefits - but unresponsive? - I 
hope not, if it really contributes to congestion the queue should 
finally drop and the flow adapt before congestion collapse.

> - with an unresponsive flow, the congested node _might_ reach a level
>    of congestion requiring tail-drop. But it also might not. :^(
>
Could be so.

> - _if_ a receiver added a way to observe this abuse (possibly by
>    having the sender set an occasional "CE" mark before sending the
>    packet), the responsiveness would still be degraded. Besides,
>    there's really no chance that _all_ flows would ever be programmed
>    to do this.
>
Yes, there are ways to be smart.

>     Myself, I see no reason to add text to cover all of these: I think
> text covering the first point is sufficient.
>
> --
> John Leslie <john@jlc.net>
>
> _______________________________________________
> aqm mailing list
> aqm@ietf.org
> https://www.ietf.org/mailman/listinfo/aqm
>

Gorry & Michael


From nobody Tue Apr 28 07:44:54 2015
Return-Path: <john@jlc.net>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id D434E1ACE81 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 07:44:52 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.21
X-Spam-Level: 
X-Spam-Status: No, score=-4.21 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id ydX8E096BC5e for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 07:44:51 -0700 (PDT)
Received: from mailhost.jlc.net (mailhost.jlc.net [199.201.159.4]) by ietfa.amsl.com (Postfix) with ESMTP id EB10E1ACE84 for <aqm@ietf.org>; Tue, 28 Apr 2015 07:44:36 -0700 (PDT)
Received: by mailhost.jlc.net (Postfix, from userid 104) id 69320C94BD; Tue, 28 Apr 2015 10:44:23 -0400 (EDT)
Date: Tue, 28 Apr 2015 10:44:23 -0400
From: John Leslie <john@jlc.net>
To: Mikael Abrahamsson <swmike@swm.pp.se>
Message-ID: <20150428144423.GR63465@verdi>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk> <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se>
User-Agent: Mutt/1.4.1i
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/DOvo7WglIzHy_Mb5y242HOXe8bI>
Cc: aqm@ietf.org
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 14:44:53 -0000

Mikael Abrahamsson <swmike@swm.pp.se> wrote:
> 
> Also, I actually still don't understand the concept of "developing 
> congestion". For me "congestion" means there is a sustained, constant, 
> non-trivial buffer fill. So "incipient congestion" doesn't compute at all. 
> I understand this term has been used before, but that doesn't help me 
> understand the reasoning behind the wording. For me sending a congestion 
> signal, means saying there is congestion. It's not saying there is 
> developing congestion (whatever that might be).

   I agree that "developing congestion" doesn't convey the same meaning
to all readers.

   When I suggested we use "incipient congestion", I was thinking of
the queueing-theory definition.

> I would greatly appreciate if someone could point me to text explaining 
> this so I can better understand the reasoning behind this wording.

   I recommend my slides from IETF-77:

https://www.ietf.org/proceedings/77/slides/iccrg-7.pdf

starting with slide 11; and the Bauer_Clark_Lehr_2009 paper it refers
to (Section 3 of that paper).

   Briefly, the Queuing Theory definition is the most useful for
control-loop purposes:
"
" congestion is said to occur if the arrival rate into a system exceeds
" the service rate of the system at a point in time.

   This is what I meant "incipient" congestion to mean.

   (NB it _does_ make sense to talk of "incipient congestion" persisting
for a longer period of time, believe it or not!)

--
John Leslie <john@jlc.net>


From nobody Tue Apr 28 10:32:55 2015
Return-Path: <mirja.kuehlewind@tik.ee.ethz.ch>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 43A061A8939 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 10:32:54 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.91
X-Spam-Level: 
X-Spam-Status: No, score=-3.91 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, MIME_8BIT_HEADER=0.3, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 5CCWjpGAaYjh for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 10:32:49 -0700 (PDT)
Received: from smtp.ee.ethz.ch (smtp.ee.ethz.ch [129.132.2.219]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id C1A231A6FEC for <aqm@ietf.org>; Tue, 28 Apr 2015 10:32:48 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by smtp.ee.ethz.ch (Postfix) with ESMTP id 4DF4BD931C; Tue, 28 Apr 2015 19:32:47 +0200 (MEST)
X-Virus-Scanned: by amavisd-new on smtp.ee.ethz.ch
Received: from smtp.ee.ethz.ch ([127.0.0.1]) by localhost (.ee.ethz.ch [127.0.0.1]) (amavisd-new, port 10024) with LMTP id kRJW60axH2Ym; Tue, 28 Apr 2015 19:32:47 +0200 (MEST)
Received: from [82.130.103.143] (nb-10510.ethz.ch [82.130.103.143]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) (Authenticated sender: mirjak) by smtp.ee.ethz.ch (Postfix) with ESMTPSA id 0E1FAD9319; Tue, 28 Apr 2015 19:32:47 +0200 (MEST)
Message-ID: <553FC43E.1060608@tik.ee.ethz.ch>
Date: Tue, 28 Apr 2015 19:32:46 +0200
From: =?windows-1252?Q?Mirja_K=FChlewind?= <mirja.kuehlewind@tik.ee.ethz.ch>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: gorry@erg.abdn.ac.uk, aqm@ietf.org,  Michael Welzl <michawe@ifi.uio.no>
References: <55392BD6.501@tik.ee.ethz.ch> <553F8414.7050608@erg.abdn.ac.uk>
In-Reply-To: <553F8414.7050608@erg.abdn.ac.uk>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/Nu7LzW1r_hGSuXfDBT2zxYj0-7k>
Subject: Re: [aqm] Review of  draft-ietf-aqm-ecn-benefits-03
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 17:32:54 -0000

Hi Gorry, hi Michael,

btw. you'd think about including a ToC because I think the doc is long enough to 
add one...

otherwise see inline...

>   > For the first part (listing benefits) it might also be good to make
> clear/distinguish who has these benefits. I think all benefits that are
> currently listed are only advantageous for the end host/application. Are
> there any benefits for a network operator? Would it be possible to write
> this document such that I could also use it to point network operators
> to and give them an incentive to enable ECN?
>
> MW/GF: we can only think of one benefit of ECN as currently defined
> (i.e. without basing it on ConEx documents) that obviously targets the
> network operator: making incipient congestion visible (such that it
> could be used e.g. for ConEx). This is addressed in section 3.5. Since
> this is one out of six listed benefits in the table, creating categories
> for end host / application vs. network operator seems unnecessary to us.

Okay. Can you maybe state more clearly in the text (or even organize the 
structure such that is would be clear) who (end/app or network) benefits in each 
section...?

>   > Another high level comment is that you say in the introduction that
> this document "also identifies some potential problems that might occur
> when ECN is used" but then you don't really discuss them. I think to
> show both sides of the coin in this document would make the document
> more useful (and more honest). One point that you mention slightly here
> is that cheating is easier than with loss by not providing the feedback.
> Another point might be fairness between ECN and non-ECN traffic as
> marking will not reduce the queue length and therefore might lead to a
> higher loss rate for the non-ECN traffic instaed. I guess there are
> papers about this; don't have any by hand right now. Are there any other
> problems that should be mentioned?
>
> MW/GF: This was discussed, and we agreed to remove the "drawbacks"
> discussion, to align with the original proposed work. So, we will remove
> this sentence from the introduction (it is in fact a left-over that
> should have been removed before). As for fairness, it seems to us that
> the related thread has concluded without a clear result. In the absence
> of evidence or references we prefer to stay away from hand-waving about
> this matter in the document.

I've looked at the old "Pitfalls when using ECN" section and I agree that this 
section is not really appropriate for this document and therefore was removed 
correctly. But the reason is that this section tries to give recommendations. 
What I would like to see is a section that documents the known problems that 
have been seen in the past. The reason why I would like to see this section is 
that, currently the document tells only positive things about ECN and at the end 
a reader might just wonder why ECN is not deployed if it is so great... I think 
you have to answer this question (by explaining what happen in the past without 
making recommendation for the future).

>
>   > Find more detailed comment by section below:
>   >
>   > Abstract
>   > --------
>   > ...says "...potential benefits when applications enable Explicit
> Congestion Notification (ECN)" -> usually an application cannot able ECN
> because usually it's a system setting...?
>
> MW/GF: Good catch! We'll rephrase this as "..when enabling".

Okay. Or even "... potential benefits for applications (and network operators) 
when enabling ..."...?

>
>
>   > Section 1
>   > ---------
>   > ... says "..separate
>   >   configuration of the drop and mark thresholds is known to be
>   >   supported in some network devices and this is recommended
>   >   [RFC2309.bis]."
>   > RFC2309bis does not recommend different settings, it only say that it
> should be possible have different configuration of both. Further, I
> think this should not only concern THE threshold (whatever this is) but
> usually there are several parameters you might want to set independent
> of each other, e.g. the max mark/drop probability in RED.
>
> MW/GF: Suggested update:
> "While it has often been assumed
> that network devices should CE-mark packets at the same level of
> congestion at which they would otherwise have dropped them, separate
> configuration of the drop and mark conditions. Such separate
> configuration is
> known to be supported in some network devices and this is recommended
> [RFC2309.bis]."

First sentence is missing something but else okay. Second sentence should maybe 
be "Such separate configurability is known to be supported in some network 
devices and this is  recommended [RFC2309.bis]."...?

>
>   > Section 2
>   > ----------
>   > 1) I'm not sure I understand the purpose of this section or maybe
> just the title is wrong. I'm currently seeing this section rather as a
> section that provides the needed background knowledge than is talking
> about deployment. For this purpose I'd put all references and
> potentially a brief summary to other RFC/drafts on ECN in this section
> including RFC2884, RFC4774, RFC5562, RFC6040, RFC6679,
> draft-briscoe-tsvwg-ecn-encap-guidelines and draft-ietf-tcpm-accecn-reqs
> (and rename it).
>
> MW/GF: This section lists requirements for deployment. Suggestion:
> rename to "ECN deployment requirements"

Renaming makes sense then. However I'm also not really sure about requirements. 
It's rather something like 'needed pre-conditions/steps to enable successful 
deployment of ECN'...? Which would then include section 4...? And if this is not 
mentioned to be some kind of background section, would it make more sense to 
move this section after the current section 3 (closer to 4 and 5)?

>
>   >
>   > 2) Second paragraph says:
>   > "Network devices must not drop packets solely because these
> codepoints are used [RFC2309.bis]."
>   > Not sure this is the right document to says this (because currently
> it not seems to be directed to network operator/equipment vendors but
> admins/application developers). However, if it says this, it should also
> say that network devices should not bleach these bits.
>
> MW/GF: suggest: "Network devices must not drop packets solely because
> these codepoints are used or erase these codepoints [RFC2309.bis]."

Yes, but, as stated in a separate mail, the reference probably should be removed 
because [RFC2309.bis] does not give this recommendation.

>
>
>   > 3) First bullet in list says
>   > "A recent survey reported growing support for ECN on common network
> paths [TR15]."
>   > This sounds like TR15 shows that ECN is actually used in the
> Internet. However, TR15 only shows that there are only very few cases
> left where ECN packets are dropped or incorrectly altered. Please
> clarify or remove this sentence here.
>
> MW/GF: suggest: "A recent survey reported that incorrect altering of ECN
> bits or consistent dropping of packets carrying the ECN codepoint is
> rare on common network paths [TR15]."

Okay.

>
>
>   > 4) You could cite draft-bensley-tcpm-dctcp-00 instead of the DCTCP
> Sigcomm paper (or both).
>
> MW/GF: The paper is a stable reference for now.
> But if/when the IETF decides on this, we can add a reference.

Okay.

>
>
>   > 5) I would remove the subsection headings (both 2.1 and 2.2) and just
> add the text there to the main part of the section.
>
> MW/GF: OK
>
>
>   > 6) "An AQM algorithm that supports ECN needs to define
>   >   the threshold and algorithm for ECN-marking."
>   > This is kind of self-redundant and therefore does not really makes
> sense to me to say; of course an algo that supports ECN needs to say
> something about ECN...
>
> MW/GF: We agree, but suggest to keep it nevertheless, it is a hint to
> document authors to not forget that they should specify ECN rather than
> just assuming some default behaviour.

This sentence is already (word-by-word) in [RFC2309.bis] where is makes more 
sense because [RFC2309.bis] talks about all kind of AQM and not only ECN. In 
this document it seems slightly weird because it's anyway only about ECN. I 
would recommend a slight rewording to something like "The used AQM needs to 
support ECN by defining the threshold/parameters and algorithm for ECN marking."

>
>   > 7) You can use TR15 to provide a reference for the first paragraph in
> section 2.2:
>   > "Cases have been noted where a sending endpoint marks a packet with a
>   >   non-zero ECN mark, but the packet is received with a zero ECN value
>   >   by the remote endpoint."
>
> MW/GF: OK, will add the reference there
>
>
>   > 8) I'd move the second paragraph of section 2.2. ("The current..") to
> a potentially new problems section, talking about known/previous
> deployment problems.
>
> MW/GF: the document does not accentuate problems in this way, as a
> result of prior discussion. We therefore think that this paragraph is ok
> in its current place.

Okay. The first sentence of the next paragraph is then redundant with the 
statement of this paragraph. Maybe consider removing this sentence and/or 
merging the two paragraphs.

>
>
>   > 9) I would simply remove paragraph 3-4 of section 2.2 because this
> was basically as already mentioned by referring to 2309bis and rfc6040
> in section 2.1.
>
> MW/GF: we do think these paragraphs add value here: they describe the
> problem in greater detail than the text before, explaining the problem
> here is different - and, we think, better - than just pointing to
> references.

Okay. Please just reread then and check for redundancy with the rest of section 2...

>
>
>   > Section 3.2
>   > ------------
>   > 1) Don't understand why there is a listing here...? Just remove the
> listing and make text out of it...?
>
> MW/GF: This is to help identify the entities that need to collaborate.

The sentence before the listing says: "... using ECN can remove resulting 
delay..." and then there are two bullet points, where one explains how an delay 
is removed but the other doesn't.

The second point says "the transport receiver notes that it has received CE-
       marked packets, and then requests the sender to make an
       appropriate congestion-response to reduce the maximum transmission
       rate for future traffic.".
This is true whenever you use ECN, and is not specific for 'Reducing HOL 
blocking'. So I don't understand why this is listed here...

>
>
>   > 2) The sentence "This also
>   >      avoids the inefficiency of dropping data that has already made it
>   >      across at least part of the network path."
>   > does not belong in this section. This sentence should just be moved
> to section 3.1 (or in an own section) and must be further explained,
> saying that dropping packet at the of the path has already blocked
> resources that other traffic could have used otherwise.
>
> MW/GF: agreed. We will insert it at the beginning of section 3.1.
>
>
>   > Section 3.3
>   > -----------
>   > 1) I'd say this section misses on part of the discussion. It is true
> that if by chance your last  packet(s) get lost ECN can help. However,
> this section reads a little like, with ECN it is save to send packet
> bursts. Which is not true because even if ECN is used by a network
> device, the queue might be too small to hold the whole burst. I believe
> this case happen very often which might be a reason for the higher tail
> loss probability that sometimes is experienced with IW10. Please add
> this point to the discussion.
>
> MW/GF: we agree that we shouldn't say that "with ECN it is ok to send
> packet bursts" - we want to stay away from such general recommendations
> and just state the potential benefit of ECN when it saves the last
> packet of a burst. See our next comment for more:
>
>
>   > 2) I don't really get the point of the second paragraph. First of all
> it is confusion that this paragraph starts which "In addition to
> avoiding HOL blocking,.."; I guess that is left over from a previous
> version of this text...? And then you talk about a connection that is
> currently idle, so why is the performance of this connection that is
> currently not sending anything reduced?
>
> MW/GF: indeed it seems that this paragraph has been mangled during
> updates. To address your item 1 and 2, we suggest the following replacement:
>
> ***
> "While using ECN can never guarantee loss prevention, and thus losses
> at the end of a burst can occur with or without ECN, using ECN can increase
> the chance for that last packet to be ECN-marked instead of dropped.
> This can allow the
> transport to avoid the consequent loss of state about the network path it is
>      using, which would have arisen had there been a retransmission
>      timeout.  Typical impacts of a transport timeout are to reset path
>      estimates such as the RTT, the congestion window, and possibly other
>      transport state that can reduce the performance of the transport
>      until it again adapts to the path."
> ***

Still not fully the point: ECN can't prevent tail loss due to queue overflow, 
however if by chance the last segment(s) was dropped due to a regular AQM-based 
drop decisions, ECN CAN prevent these drops.


>   > 3) I don't understand what "applications that send intermittent
> bursts of data, and rely upon timer-based recovery of packet loss"
> are...? Isn't the transport responsible to not send bursts and care
> about recovery...?
>
> MW/GF: MPEG-DASH traffic for instance, in particular when used over
> non-paced TCP. UDP-based applications too.

I understood which kind of application you were talking about but the sentence 
above seem to be wrong for me, because the application (today) usually choose a 
transport (that might handle pacing and recovery or not) while the application 
might not fully be aware of the consequences of choosing a certain transport e.g 
leading to burst sending. So even though these application largely exist today, 
it might not have been an active choice to send bursts or rely on timers for 
recovery. While this sentence sounds as if the application would require to 
behave this way. Do you know what I mean? Is it maybe possible to simply 
rephrase this sentence? Sorry don't have a good proposal right now...

>
>
>   > 4) For the last paragraph in section 3.3 note that stacks often
> remember RTT measurements for a certain IP address and set the initial
> RTO based on this information.
>
> MW/GF: suggestion: replace:
> ***
> because in this
>      case TCP cannot base the timeout period on prior RTT measurements
>      from the same connection.
> ***
> with:
> ***
> because in this
>      case TCP may not be able to base the timeout period on prior RTT
> measurements.
> ***

Okay.

>
>
>   > Section 3.4
>   > -----------
>   > You still need FEC or some kind of error concealment even if ECN is
> used because you can never be sure that your packet are not get dropped
> (by non-ECN-enable devices or other reasons). Therefore using ECN will
> clearly not reduce complexity. The only thing you can do is to
> potentially reduce the amount of redundancy you send if you know that a
> certain path is ECN enables or don't see losses at the beginning of a
> connection. This can save network resources but actually might not
> improve user experience; in fact the user experience might be worse in
> case there are sudden losses.
>
> MW/GF: suggestion: remove "add complexity and"

Okay.

>
>
>   > Further the text says "negative impact of using loss-hiding
> mechanisms"; I don't really think that FEC has a negative impact as long
> as you've send enough redundancy...? Error concealment might but is used
> less and less. I'd recommend to talk about error concealment only in
> this last paragraph and explain a little further.
>
> MW/GF: error concealment is different from FEC, and it is only mentioned
> in this last paragraph. We suggest to replace "Because this
>      reduces the negative impact of using loss-hiding mechanisms,"   with
> "Because this can reduce the potential negative impact that some
> loss-hiding mechanisms can have,"

Okay. I'd still recommend to use the term 'error concealment' instead of 
'error-hiding' here because you describe error-hiding as FEC or duplication in 
the previous paragraph. Those mechanism only have a negative effect if you send 
to few redundancy while error concealment is used in case you have to few 
redundancy and generates the know artifacts that the user can see.

>
>
>   > Section 3.5
>   > ----------
>   > "Recording the presence of CE-marked packets can therefore provide
>   >   information about the performance of the network path."
>   > Would change to:
>   > "Recording the presence of CE-marked packets in absence of loss can
> therefore provide
>   >   information about the performance of the network path."
>
> MW/GF: ok
>
>
>   > And also say more concretely what is meant with 'performance of the
> network path' -> congestion level or no drops by other middleboxes on
> this path...
>
> MW/GF: This intentionally was kept this vague, but we'd welcome a
> concrete recommendation by a ConEx expert      (indeed "or .. or ..." is
> the problem, there are several possibilities here)

Okay if you intentionally would like to stay vague I'd even use 'state' instead 
of 'performance'. You could say: "... information about the state of the network 
path, such as the current congestion level."...?

>
>
>   > Section 3.6
>   > -----------
>   > 1) I like the section but I would phrase it differently; also it's
> not clear who needs to support what in this case. I'd like to propose
> the following text [not sure about the heading...]:
>   >
>   > "3.6 Opportunity to provide an improved congestion feedback signal
>   >
>   > Loss and ECN marking are both used as an indication for congestion.
> However, while the amount of feedback that is provided by loss should
> naturally be minimized, this is not the case for ECN. With ECN a network
> node could provide richer and more frequent feedback on the congestion
> state of a link which then could be used by the control mechanisms
> implemented in end host to make a more appropriate decision on how to
> react to congestion and to react faster to changes in congestion state.
>
> MW/GF: ok to add this up to here.

Just realized that I didn't state this clearly. These three paragraphs are meant 
to replace 3.6 including 3.6.1 (don't think there is a subheading needed here). 
Therefore the two following paragraph rephrase the third paragraph in section 
3.6.1 (where DCTCP is cited).

Please check again.

>
>
>   > Further while drop-based AQM mechanisms usually operate on a smoothed
> queue length estimation (instead of the instantaneous queue length) and
> therefore slightly delay the feedback signal to avoid unnecessary losses
> in case of transient congestion, this would be not necessary for ECN. If
> congestion is only transient due to short traffic bursts that are active
> for less than one RTT, the congestion signal would reach the sender at a
> time where the congestion is already cleared up. However, instead
> delaying the feedback in the network, the end host could reduce its
> sending rate incrementally based on the extend of congestion (that was
> experienced over e.g. the last RTT) similar as DCTCP. In case if the
> congestion is only transient, the end host would only reduce its rate
> slightly and be able to catch up quickly again. However, in case the
> congestion is persistent, this would help to remove additional delays
> from the network and resolve congestion faster which after all reduces
> the average queuing delay.
>   >
>   > However, current ECN is defined as a 'drop equivalent' in RFC3168. To
> change the semantics of ECN both the AQM in the network nodes and the
> control mechanism in the end hosts would still need to cope with nodes
> or end hosts that rely on the old semantics. Therefore changing the
> semantics can be done more easily in confined environment such as a data
> center. DCTCP is an example that changes both the configuration of the
> used AQM as well as the congestion response in the end host and relies
> on that fact that all nodes in data center are configured the same way.
> [Deployment strategies to change the semantics of ECN in the Internet
> are currently under discussion in the IETF.]"
>
> MW/GF: We think that this goes a bit too far in the direction of hinting
> about implementation and research possibilities that we don't have
> citable proof about (besides: we already refer to DCTCP twice in the
> document, and the 'drop equivalent' semantics are not a MUST in RFC3168).

I'd personally like to see at least one sentence on the point that e.g. RED uses 
a smoothed queue length because it was designed for loss and a congestion 
control that halves its sending rate on congestion. If you design an AQM for ECN 
you could be less conservation. However, its also okay for me to not go into 
more detail here but this is mention already in the current text in the 3. 
paragraph of subsection 3.6.1 that why this is in here.

However, the one point that really should be mentioned is that if you change the 
feedback signal that is provided by ECN, you very likely also have to change the 
congestion control reaction to it and you have to take into account that there 
are non-'updated' senders as well. That was the intention of the last paragraph. 
If you don't want to refer to DCTCP here agian, just take the first sentence (or 
rephrase as you like).

Further, you are right that drop equivalence is not a MUST.
Therefore I would to suggest to rephrase to:
"However, ECN is recommended to be 'drop equivalent' in RFC3168."


>   > 2) I'd move the 1. and 2. paragraph of section 3.6.1 to the
> background/deployment section or to the intro depending what you going
> to do with section 2.
>
> MW/GF: since we intend section 2 to be about deployment requirements
> only, we don't think this fits and would rather leave these paragraphs
> in section 3.6.1.

Sorry, these paragraphs really don't belong here. The subsection heading says 
"Other forms of ECN-Marking/Reactions" while these paragraphs only talk about a 
different way to provide ECN feedback; this does change the marking in the 
router nor the reaction of the sender!

Please consider moving to the intro then (or remove... but I think these things 
should be mentioned somewhere to provide the reference to these docs).

>   > Sections 4 & 5
>   > ---------
>   > First sentence talks about "operational
>   >   difficulties when the network only partially supports the use of ECN,
>   >   or to respond to the challenges due to misbehaving network devices
>   >   and/or endpoints".
>   > I think these are to very different things. Misbehaving network
> devices is a point for a  problems section (where the lesson learned is
> that we didn't think carefully enough about incremental deployment in
> the first place but do now). However, partial deployment is not a
> problem but is a thing we simply have to cope with. The text sound as if
> the goal would be that every router in the whole Internet would at some
> point of time be ECN-enabled. I don't think this will ever happen and is
> also not the goal for me. Routers that are very unlikely to ever get
> congested should no be required to look at the ECN bits or monitor the
> queue length to calculate a mark/drop probability.
>
> MW/GF: we agree, and suggest to replace this sentence with "Early
> deployment of ECN encountered a number of operational
>      difficulties due to misbehaving network devices
>      and/or endpoints."
>
>
>   > However as I said at the beginning I don't really thing that sections
> 4 and 5 belong in this document. If you decided to keep them (you have
> to change the abstract) and I'd recommend to rename them e.g 4.
> 'Incremental Deployment Strategy' or 'Requirements to enable Incremental
> Deployment' and 5. 'Recommendations for enabling ECN in network nodes
> and end hosts'.
>
> MW/GF: we suggest to insert the fact that we discuss deployment in the
> abstract, and rename these sections to 4.: "Incremental Deployment" and
> "Recommendations for enabling ECN"

Okay. Btw. if you add a ToC, I don't think you really need the summary table at 
the end of section 5 anymore.

And please also change the following sentence in section 4:
"This section describes transport mechanisms that allow ECN-enabled
    endpoints to continue to work effectively over a path with partial
    ECN support."
(by replacing 'partial ECN support' with something like 'misbehaving network 
nodes'...?)

Thanks!
Mirja


>
>
>   > I hope that's helpful! Let me know if you have any questions!
>   >
>   > Mirja
>   >
>   >
>
>
> Thank you very much,
>
> Michael & Gorry
>

-- 
------------------------------------------
Dipl.-Ing. Mirja Kühlewind
Communication Systems Group
Institute TIK, ETH Zürich
Gloriastrasse 35, 8092 Zürich, Switzerland

Room ETZ G93
phone: +41 44 63 26932
email: mirja.kuehlewind@tik.ee.ethz.ch
------------------------------------------


From nobody Tue Apr 28 12:50:13 2015
Return-Path: <swmike@swm.pp.se>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 6FDB61A702A for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 12:50:11 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -3.961
X-Spam-Level: 
X-Spam-Status: No, score=-3.961 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HELO_EQ_SE=0.35, RCVD_IN_DNSWL_MED=-2.3, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id zy25-YekcwyR for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 12:50:09 -0700 (PDT)
Received: from uplift.swm.pp.se (swm.pp.se [212.247.200.143]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id A05391A701C for <aqm@ietf.org>; Tue, 28 Apr 2015 12:50:08 -0700 (PDT)
Received: by uplift.swm.pp.se (Postfix, from userid 501) id 0CC22A1; Tue, 28 Apr 2015 21:50:07 +0200 (CEST)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=swm.pp.se; s=mail; t=1430250607; bh=bF/PCD0mggh/Kt1/o8AxKl3lENxAkE6UDGIQUn7wq8k=; h=Date:From:To:cc:Subject:In-Reply-To:References:From; b=U+v4jR23iHVZ3iQe6/s+aCePxKQ5sMbe53f339uGg0SrUHKxCcDaPSk49LYJ5duJh TZB1Fqma41/nsfDHvwfWv2I5MEA9wqqkpBNVltdAX4xwOq0ddiY9uFoHG1WUDiUuvg NuFqCjWvSudViFUB7SKk4DDTY1KOyWloOf4uF3n4=
Received: from localhost (localhost [127.0.0.1]) by uplift.swm.pp.se (Postfix) with ESMTP id 02B7C9F; Tue, 28 Apr 2015 21:50:07 +0200 (CEST)
Date: Tue, 28 Apr 2015 21:50:06 +0200 (CEST)
From: Mikael Abrahamsson <swmike@swm.pp.se>
To: John Leslie <john@jlc.net>
In-Reply-To: <20150428144423.GR63465@verdi>
Message-ID: <alpine.DEB.2.02.1504282125140.16871@uplift.swm.pp.se>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk> <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se> <20150428144423.GR63465@verdi>
User-Agent: Alpine 2.02 (DEB 1266 2009-07-14)
Organization: People's Front Against WWW
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/P__m8lcoR7fZgLibQupBOuLl5mc>
Cc: aqm@ietf.org
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 19:50:11 -0000

On Tue, 28 Apr 2015, John Leslie wrote:

> Mikael Abrahamsson <swmike@swm.pp.se> wrote:
>>
>> Also, I actually still don't understand the concept of "developing
>> congestion". For me "congestion" means there is a sustained, constant,
>> non-trivial buffer fill. So "incipient congestion" doesn't compute at all.
>> I understand this term has been used before, but that doesn't help me
>> understand the reasoning behind the wording. For me sending a congestion
>> signal, means saying there is congestion. It's not saying there is
>> developing congestion (whatever that might be).
>
>   I agree that "developing congestion" doesn't convey the same meaning
> to all readers.
>
>   When I suggested we use "incipient congestion", I was thinking of
> the queueing-theory definition.
>
>> I would greatly appreciate if someone could point me to text explaining
>> this so I can better understand the reasoning behind this wording.
>
>   I recommend my slides from IETF-77:
>
> https://www.ietf.org/proceedings/77/slides/iccrg-7.pdf
>
> starting with slide 11; and the Bauer_Clark_Lehr_2009 paper it refers
> to (Section 3 of that paper).

This link is no longer valid, but I found the paper.

>   Briefly, the Queuing Theory definition is the most useful for
> control-loop purposes:
> "
> " congestion is said to occur if the arrival rate into a system exceeds
> " the service rate of the system at a point in time.
>
>   This is what I meant "incipient" congestion to mean.

Ok, I think I understand the reasoning behind it. Now you and the other 
authors need to decide if you want these documents to be read and 
understood by people like me for instance, who works at service providers 
or other networking professions who are not heavy inty queueing theory but 
instead are just trying to get their gear to do the right thing. I have 
studied english from grade 4 to grade 11, and I was best in my class in 
english. I had never even heard the word "incipient" before 2 days ago, 
and even now I am not sure I understand the full meaning trying to be 
conveyed.

Are people like me the target of these documents? If then, I recommend you 
try to simplify the terminology in these documents, striking "incipient" 
would be a first step, or at least try to explain them in the document, or 
point to a document that does.

If I for instance read RFC2884 is has 3 mentions of "incipient congestion" 
but 95 mentions of "congestion". Am I to understand that these 3 mentions 
of "incipient" means that the congestion that is "incipient" is different 
from the mentions of just "congestion" without "incipient" before it? 
Because that subtle difference is completely lost to me and if I were to 
guess, completely just deleting the 3 mentions of "incipient" from RFC2884 
wouldn't change the document at all. Am I wrong here?

For instance here:

"Jacobson [14] proposed gateways to monitor the average queue size to 
detect incipient congestion, and to randomly drop packets when congestion 
is detected."

I just don't understand the difference between the first mention of 
"incipient congestion" and then the second mention where "incipient" is 
not used. What am I missing here?

>   (NB it _does_ make sense to talk of "incipient congestion" persisting
> for a longer period of time, believe it or not!)

If a queue has a buffer fill of between 40 and 60 ms worth of packets over 
60 seconds, is this link congested or is it experiencing a prolonged 
period of "incipient congestion"? Is the "incipience" (I don't even know 
if that is the correct term) just when the buffer fill time is increasing 
(for instance from 45 to 55 ms) and not when it's decreasing from 55 to 45 
ms, or is it just "incipient congestion" for the entire duration of the 
buffer having a constant non-trivial amount of packets in it?

-- 
Mikael Abrahamsson    email: swmike@swm.pp.se


From nobody Tue Apr 28 13:42:54 2015
Return-Path: <john@jlc.net>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 5F3D81A8849 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 13:42:53 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -4.21
X-Spam-Level: 
X-Spam-Status: No, score=-4.21 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_MED=-2.3, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id kJ2Gw2GXO_RK for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 13:42:51 -0700 (PDT)
Received: from mailhost.jlc.net (mailhost.jlc.net [199.201.159.4]) by ietfa.amsl.com (Postfix) with ESMTP id CDBA71A0068 for <aqm@ietf.org>; Tue, 28 Apr 2015 13:42:50 -0700 (PDT)
Received: by mailhost.jlc.net (Postfix, from userid 104) id E22FDC94A9; Tue, 28 Apr 2015 16:42:37 -0400 (EDT)
Date: Tue, 28 Apr 2015 16:42:37 -0400
From: John Leslie <john@jlc.net>
To: Mikael Abrahamsson <swmike@swm.pp.se>
Message-ID: <20150428204237.GT63465@verdi>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se> <553F7F15.9060509@erg.abdn.ac.uk> <alpine.DEB.2.02.1504281539560.16871@uplift.swm.pp.se> <20150428144423.GR63465@verdi> <alpine.DEB.2.02.1504282125140.16871@uplift.swm.pp.se>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <alpine.DEB.2.02.1504282125140.16871@uplift.swm.pp.se>
User-Agent: Mutt/1.4.1i
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/svjkF1_-vgHNw0xhAZDGL7fZg1I>
Cc: aqm@ietf.org
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Tue, 28 Apr 2015 20:42:53 -0000

Mikael Abrahamsson <swmike@swm.pp.se> wrote:
>>
>>> I would greatly appreciate if someone could point me to text explaining
>>> this so I can better understand the reasoning behind this wording.
>>
>> I recommend my slides from IETF-77:
>>
>> https://www.ietf.org/proceedings/77/slides/iccrg-7.pdf
>>...
>> Briefly, the Queuing Theory definition is the most useful for
>> control-loop purposes:
>>"
>>" congestion is said to occur if the arrival rate into a system exceeds
>>" the service rate of the system at a point in time.
>>
>> This is what I meant "incipient" congestion to mean.
> 
> Ok, I think I understand the reasoning behind it. Now you and the other 
> authors need to decide if you want these documents to be read and 
> understood by people like me for instance, who works at service providers 
> or other networking professions who are not heavy inty queueing theory but 
> instead are just trying to get their gear to do the right thing.

   Speaking for myself, I would like the document to be easy for network
operators to read.

   But I am not an author, and I certainly can't speak for the authors.

> I have studied english from grade 4 to grade 11, and I was best in my
> class in english.

   Alas, that doesn't mean as much as it used to...

> I had never even heard the word "incipient" before 2 days ago, 

   Q.E.D. ;^)

> and even now I am not sure I understand the full meaning trying to be 
> conveyed.

   Alas, you're probably right. :^(

   The temptation becomes overwhelming, at some point in the RFC-generation
process, to use terms that "sound right" without defining them well.

   I have a history of annoying folks by asking for good definitions at
and after this point in the process. But I would like to see a definition
of "incipient congestion" -- perhaps something like:
"
" incipient congestion: the detection of congestion when it is just
"   starting, perhaps by noticing that the arrival rate exceeds the
"   forwarding rate.

   I don't think this is inconsistent with previous usage, but YMMV...

> If I for instance read RFC2884

   "Performance Evaluation of ECN..."

> it has 3 mentions of "incipient congestion" but 95 mentions of
> "congestion". Am I to understand that these 3 mentions of "incipient"
> means that the congestion that is "incipient" is different from the
> mentions of just "congestion" without "incipient" before it? 

   RFC 2884 is water over the hill: we cannot change it.

   My opinion is that it uses "incipient congestion" to indicate that
Random Early Detection signals oncoming congestion before it reaches
the point of tail-drop. "Congestion" used alone seems to be a generic
term covering a quite-wide range of meanings.

   Note, please, that RFC 2884 is informational -- which means that
only the authors are responsible for defining their terms, and that,
unless quite specifically cited, its definitions do not affect the
meaning of the terms it uses when they appear in other documents.

   (That is to say, don't bother looking there for definitions that
we should have included in _this_ document.)

> Because that subtle difference is completely lost to me and if I were to 
> guess, completely just deleting the 3 mentions of "incipient" from RFC2884 
> wouldn't change the document at all. Am I wrong here?

   Yes, IMHO, but YMMV. It meant (IMHO) to distinguish congestion signals
from RED from tail-drop which went before it.

> For instance here:
> 
> "Jacobson [14] proposed gateways to monitor the average queue size to 
> detect incipient congestion, and to randomly drop packets when congestion 
> is detected."

   I don't know the context of what you're quoting here, so I can't
respond.

>> (NB it _does_ make sense to talk of "incipient congestion" persisting
>> for a longer period of time, believe it or not!)
> 
> If a queue has a buffer fill of between 40 and 60 ms worth of packets over 
> 60 seconds, is this link congested or is it experiencing a prolonged 
> period of "incipient congestion"?

   Perhaps yes; perhaps no. Particular AQM algorithms may define this
either way.

> Is the "incipience" (I don't even know if that is the correct term)

   Yes, it's a word, meaning what I think you intend.

> just when the buffer fill time is increasing (for instance from 45 to
> 55 ms) and not when it's decreasing from 55 to 45 ms,

   That would be my preference: when arrival rate exceeds forwarding
rate, it's a bad idea to probe for additional capacity; but when the
forwarding rate exceeds the arrival rate, there probably is unused
capacity which it's reasonable to probe for.

   (Note, of course, that I do _not_ claim this is consistent with
current TCP congestion-control algorithms, which are optimized for
the lack of information which we must live with.)

> or is it just "incipient congestion" for the entire duration of the 
> buffer having a constant non-trivial amount of packets in it?

   Alas, YMMV. The RED algorithm seems to assume that there is some
"incipient congestion" whenever some fraction of the buffer is full.
I doubt we can fix that in this document, but I'd sure like to try!

--
John Leslie <john@jlc.net>


From nobody Tue Apr 28 20:23:51 2015
Return-Path: <fred@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 6774E1A8724 for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 20:23:50 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -114.511
X-Spam-Level: 
X-Spam-Status: No, score=-114.511 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5, USER_IN_WHITELIST=-100] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id uNssjtgqVIgj for <aqm@ietfa.amsl.com>; Tue, 28 Apr 2015 20:23:48 -0700 (PDT)
Received: from rcdn-iport-7.cisco.com (rcdn-iport-7.cisco.com [173.37.86.78]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id C88FC1A87B0 for <aqm@ietf.org>; Tue, 28 Apr 2015 20:23:48 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=2995; q=dns/txt; s=iport; t=1430277828; x=1431487428; h=from:to:cc:subject:date:message-id:references: in-reply-to:mime-version; bh=tT6Bu2wI0EwhrC9yBGTmrF7o/D21Rvj5sDxpY+NoUlQ=; b=ZGuJMNMArcgw1xB5OQ1MB8xQrP7U1NP/nKqyuESCT0cufZJPQwwgMZ9K Q0ADHk77wTP/8ATZmIhI7Fix3lckZ31oNmDQJNj8jmB8dEJCXRP150/xa tS0BYK0wuPq2GlTXtSyEUyu38vElCxXiZfpnRJr2bPo67+wjZGxxyMQe3 w=;
X-Files: signature.asc : 487
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0AcBQAuTkBV/5JdJa1cgwxTXAEEgxXFaIYEAoE2TAEBAQEBAYELhCABAQEDASNWBQsCAQgOCioCAjIlAgQOBQ6IFQgNsnaUFgEBAQEBAQEBAQEBAQEBAQEBAQEBAReLOIUFB4JoL4EWBYZHix6BdIE3WYY7gSM9gwyCcI4QI2CDFG+BRIEBAQEB
X-IronPort-AV: E=Sophos;i="5.11,668,1422921600";  d="asc'?scan'208";a="415559379"
Received: from rcdn-core-10.cisco.com ([173.37.93.146]) by rcdn-iport-7.cisco.com with ESMTP; 29 Apr 2015 03:23:48 +0000
Received: from xhc-aln-x13.cisco.com (xhc-aln-x13.cisco.com [173.36.12.87]) by rcdn-core-10.cisco.com (8.14.5/8.14.5) with ESMTP id t3T3NmRk009565 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Wed, 29 Apr 2015 03:23:48 GMT
Received: from xmb-rcd-x09.cisco.com ([169.254.9.172]) by xhc-aln-x13.cisco.com ([173.36.12.87]) with mapi id 14.03.0195.001; Tue, 28 Apr 2015 22:23:47 -0500
From: "Fred Baker (fred)" <fred@cisco.com>
To: Mikael Abrahamsson <swmike@swm.pp.se>
Thread-Topic: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
Thread-Index: AQHQgivnPopwYxtXo064JSvvoS7ZwQ==
Date: Wed, 29 Apr 2015 03:23:47 +0000
Message-ID: <836F2336-CE02-4FB3-B815-0C2EFC36B4D5@cisco.com>
References: <5539C899.4080704@mti-systems.com> <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se>
In-Reply-To: <alpine.DEB.2.02.1504241052480.16871@uplift.swm.pp.se>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: yes
X-MS-TNEF-Correlator: 
x-originating-ip: [10.19.64.121]
Content-Type: multipart/signed; boundary="Apple-Mail=_3E92C9E8-1708-4FB3-A448-D910731CA4D2"; protocol="application/pgp-signature"; micalg=pgp-sha1
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/8Z0ac02zIsjMZTXJglETNhz7stc>
Cc: "aqm@ietf.org" <aqm@ietf.org>
Subject: Re: [aqm] WGLC for draft-ietf-aqm-ecn-benefits
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Apr 2015 03:23:50 -0000

--Apple-Mail=_3E92C9E8-1708-4FB3-A448-D910731CA4D2
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8

I similarly support the document. One comment, for Mikael.

> On Apr 24, 2015, at 2:25 AM, Mikael Abrahamsson <swmike@swm.pp.se> =
wrote:
>=20
> On Fri, 24 Apr 2015, Wesley Eddy wrote:
>=20
>> To keep moving forward with the set of documents, we'd like to start
>> a Working Group Last Call on:
>> http://datatracker.ietf.org/doc/draft-ietf-aqm-ecn-benefits/
>>=20
>> Please send any comments you have on this document to the list
>> by May 8th.  Other than those raised in the thread Mirja has just
>> opened with her review, I'm not aware of any other open issues.
>=20
> I have re-read this document as if I never read it before. Writing my =
thoughts as I read it.
>=20
> The title says "benefits" but the abstract states that the document =
will talk about both benefits and potential problems. I therefore think =
the title does not reflect the abstract.
>=20
> 1. Introduction: I don't understand the "(potential)", doesn't =
transport always treat loss and CE marking as indications of congestion? =
Actually, the way the introduction reads today one might interpret that =
TCP and SCTP are "Internet transports" but UDP is not, because it's =
merely "transport".

Delay-based transports may or may not. For example, if Caltech FAST is =
in use, the window is changed according to
                       least measured RTT
	cwnd=E2=80=99 =3D cwnd * ------------------ + alpha
                           mean RTT

And the obvious thing to do on receipt of CE might well be to simply =
arrest upward growth on the window, allowing the natural process to =
subsequently reduce it. That would mean, in the above, solving for mean =
RTT when cwnd=E2=80=99 =3D=3D cwnd (what RTT value would have cwnd not =
change in the above equation?), and setting the mean RTT variable to

                                                         cwnd
    mean RTT :=3D min ( Mean RTT, least measured RTT * ------------ )
                                                     cwnd - alpha



--Apple-Mail=_3E92C9E8-1708-4FB3-A448-D910731CA4D2
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment; filename="signature.asc"
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Message signed with OpenPGP using GPGMail

-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org

iQEVAwUBVUBOwZ9ieig10VPpAQLd3ggAlX0v9MWeQPZj8bf1DLCIYSxGJuMpjZMz
/HnuaoJ9fsO76Bp6IN2YMqOAG2T1bU4qrDbI6Uczbcd7gF94upP8/UnOFBq3jNgY
spthuAFgYPkEF26UlDe5dzOAcKSWDtAmk04Ntq/d9DzvacIXJRsIDDmoKOo6OaUz
zrxsyfFVEvnjnZinQrMJrRjxS2r+pD2ERZrJGkmgZ8NFn3oZ4qDNxs0u+Owdne0V
oNui0Ntz3kzOAXZnY3cRn+aBrZvxBkQR6oYjta4hsotFjtn8IQLoESYo1jSxR15Q
wNjTbS4rGjxzjBzO9O/3pd+ElgdAlB6jKL0ldre7XU4jdEcVa22nmw==
=8CIz
-----END PGP SIGNATURE-----

--Apple-Mail=_3E92C9E8-1708-4FB3-A448-D910731CA4D2--


From nobody Wed Apr 29 09:42:28 2015
Return-Path: <bob.briscoe@bt.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id AAE881A8A50 for <aqm@ietfa.amsl.com>; Wed, 29 Apr 2015 09:42:26 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -2.611
X-Spam-Level: 
X-Spam-Status: No, score=-2.611 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, RCVD_IN_DNSWL_LOW=-0.7, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id Wd7_p1YV9C8C for <aqm@ietfa.amsl.com>; Wed, 29 Apr 2015 09:42:24 -0700 (PDT)
Received: from hubrelay-rd.bt.com (hubrelay-rd.bt.com [62.239.224.99]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id 2D09D1A889F for <aqm@ietf.org>; Wed, 29 Apr 2015 09:42:23 -0700 (PDT)
Received: from EVMHR71-UKRD.domain1.systemhost.net (10.36.3.109) by EVMHR68-UKRD.bt.com (10.187.101.23) with Microsoft SMTP Server (TLS) id 8.3.348.2; Wed, 29 Apr 2015 17:42:22 +0100
Received: from EPHR01-UKIP.domain1.systemhost.net (147.149.196.177) by EVMHR71-UKRD.domain1.systemhost.net (10.36.3.109) with Microsoft SMTP Server (TLS) id 8.3.348.2; Wed, 29 Apr 2015 17:42:21 +0100
Received: from bagheera.jungle.bt.co.uk (132.146.168.158) by EPHR01-UKIP.domain1.systemhost.net (147.149.196.177) with Microsoft SMTP Server id 14.3.181.6; Wed, 29 Apr 2015 17:42:18 +0100
Received: from BTP075694.jungle.bt.co.uk ([10.111.115.165])	by bagheera.jungle.bt.co.uk (8.13.5/8.12.8) with ESMTP id t3TGgFqH012984;	Wed, 29 Apr 2015 17:42:15 +0100
Message-ID: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9
Date: Wed, 29 Apr 2015 17:42:15 +0100
To: Richard Scheffenegger <rs@netapp.com>, "Eddy Wesley M. [VZ]" <Wesley.M.Eddy@nasa.gov>
From: Bob Briscoe <bob.briscoe@bt.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed
X-Scanned-By: MIMEDefang 2.56 on 132.146.168.158
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/4kadrxKui3xgfolwLS46S1t8_r4>
Cc: AQM IETF list <aqm@ietf.org>
Subject: [aqm] PIE (and CoDel) drafts: proposed standard vs informational?
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Wed, 29 Apr 2015 16:42:26 -0000

Richard, Wes,

1) The AQM charter says:
"Dec 2014 - Submit first algorithm specification to IESG for 
publication as Proposed Standard"

I volunteered to do a thorough review of the PIE draft, which I'm 
just writing up. One of the problems is that it says 'Proposed 
Standard' at the top, but it's written in an informational style. 
There is no normative language saying what a PIE implementation MUST 
do, what it SHOULD do and what it MAY do.

And actually, I'm not convinced that it would be worthwhile to add 
normative language. On reflection, I believe it would be better left 
as an informative specification of the PIE algorithm and its possible 
variants.

I'm not religious about this. Normative language might be useful. But 
on a pragmatic note, it will take considerable time and argument to 
decide which aspects are the essence of PIE and which are optional, 
because we will then have to consider whether certain combinations of 
options are not workable, or whether taking out too many of the 
optional parts makes it non-viable.

What will be the point of being able to say that a particular 
implementation complies with an RFC that recorded at one snapshot in 
time what we thought defined the line between PIE and not PIE? If a 
future improvement is not described in the RFC, it will be 
non-compliant but better.

BTW, wrt the CoDel draft, it does contain some normative language., 
but there are many aspects of the spec where it is not stated whether 
they are MUST, SHOULD or MAY. So this email really applies to both specs.

2) There /are/ interoperability concerns between AQMs, and between 
delay-based congestion controls (like LEDBAT) and AQMs. Writing a 
Proposed Standard giving the interoperability constraints on all AQMs 
would be a useful exercise. However, giving the special status of 
proposed standard to one design of PIE at one snapshot in time will 
not say anything about interoperability.

3) If someone comes up with an improvement on PIE (or CoDel) next 
week or next month or next year, will the IETF want to standardise 
it? Is the intention that AQM-chair will be a job for life?


Bob


________________________________________________________________
Bob Briscoe,                                                  BT  


From nobody Wed Apr 29 19:46:21 2015
Return-Path: <andrewmcgr@google.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id C05B41A1A9F for <aqm@ietfa.amsl.com>; Wed, 29 Apr 2015 19:46:19 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.388
X-Spam-Level: 
X-Spam-Status: No, score=-1.388 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, FM_FORGED_GMAIL=0.622, HTML_MESSAGE=0.001, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qYXGUePeg103 for <aqm@ietfa.amsl.com>; Wed, 29 Apr 2015 19:46:18 -0700 (PDT)
Received: from mail-qc0-x236.google.com (mail-qc0-x236.google.com [IPv6:2607:f8b0:400d:c01::236]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id DA3FA1A90B9 for <aqm@ietf.org>; Wed, 29 Apr 2015 19:46:17 -0700 (PDT)
Received: by qcpm10 with SMTP id m10so22880942qcp.3 for <aqm@ietf.org>; Wed, 29 Apr 2015 19:46:17 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type; bh=T9wgiY8Vn+b8qEbi6Qb8EiCY/Ue3wmQ8Wp8wvBjIJl4=; b=HpxCVldJQ0SLF2T62+VP7LAnvcqzl/ZaOgXBX2gm4Zr0ICVzRPrhtQg5eulAUs1KoI K+/RJ6yYqH2yGi7TduIbCy9tdVuwB17VkiMtAM6IGIaSAcp7ExEmX+1322EXC+CV2ZB6 WEiudiC7tyoexjObinRBZ0T8TWN/54VPp/Q/9wjAXUeFMwSIYaO3CZ/+II93CbCidy/Q J0Ex3NguVH1+kRaw+dNJTUJtyAdgL/um3DFB9t5tnH1kW/mn8LzKoUU9GIjpu5CdgE2N ZTDcpe/tioA22IWz8nE53bNVlfLSal6mGTZjQITRcd2fRxighiOtJXTbduOT6N7K1kbe gqaQ==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:date :message-id:subject:from:to:cc:content-type; bh=T9wgiY8Vn+b8qEbi6Qb8EiCY/Ue3wmQ8Wp8wvBjIJl4=; b=UV/krxv1wUysAivHXpLGxq8dISb2MdIlW50ex5l7XSxXkIKe8xHmZv7MTNOv9GQ74u DY/nQ643ibq3Nfivdqva7tQorpaZpJ2+bJ9YYXRyriQLzLvriw7+qoTBwnnXC2CDlf/d q/O5KpCQs0/4nECegvuFkc2DmZxSRuRcXRZ7DlelXASMGgcY80bYL/qQJuLjOaAunUbZ u2nfKNPb2cJ0ovVe3xsJ2SVkwN97r0ZBK+GtoYDEWzfHJzNrnrRIzb8Y9jnKJy5fAFkE rcNRvdxw978keKBkLNKzt4gZ3rpwEEvG7F51/b1IclDsGqGILuQqtCKdG9hioXKi8DOJ X8fA==
X-Gm-Message-State: ALoCoQmsbOGhYdF2mH/9ehrWOyGK/ssFd8s4rJY8y2j90kWFHZewC518BsKcFP/NkVH7CqSc2VV/
MIME-Version: 1.0
X-Received: by 10.140.150.19 with SMTP id 19mr2267885qhw.53.1430361977134; Wed, 29 Apr 2015 19:46:17 -0700 (PDT)
Received: by 10.96.139.232 with HTTP; Wed, 29 Apr 2015 19:46:17 -0700 (PDT)
In-Reply-To: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
References: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
Date: Thu, 30 Apr 2015 12:46:17 +1000
Message-ID: <CAPRuP3menOVD-PO5S4o8ajGHLU7dfRfnv-fFarXdvs9PTM47dw@mail.gmail.com>
From: Andrew Mcgregor <andrewmcgr@google.com>
To: Bob Briscoe <bob.briscoe@bt.com>
Content-Type: multipart/alternative; boundary=001a1135a2748c4d5f0514e81980
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/KSjzONYnTQB2Gq7EmXfVuRRYMMU>
Cc: Richard Scheffenegger <rs@netapp.com>, AQM IETF list <aqm@ietf.org>, "Eddy Wesley M. \[VZ\]" <Wesley.M.Eddy@nasa.gov>
Subject: Re: [aqm] PIE (and CoDel) drafts: proposed standard vs informational?
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 30 Apr 2015 02:46:19 -0000

--001a1135a2748c4d5f0514e81980
Content-Type: text/plain; charset=UTF-8

FWIW, the authors and editors have been thinking of CoDel as informational;
I think this comes from a discussion post-chartering which converged to
"there is no one AQM to rule them all, therefore we won't be doing
standards track".

On 30 April 2015 at 02:42, Bob Briscoe <bob.briscoe@bt.com> wrote:

> Richard, Wes,
>
> 1) The AQM charter says:
> "Dec 2014 - Submit first algorithm specification to IESG for publication
> as Proposed Standard"
>
> I volunteered to do a thorough review of the PIE draft, which I'm just
> writing up. One of the problems is that it says 'Proposed Standard' at the
> top, but it's written in an informational style. There is no normative
> language saying what a PIE implementation MUST do, what it SHOULD do and
> what it MAY do.
>
> And actually, I'm not convinced that it would be worthwhile to add
> normative language. On reflection, I believe it would be better left as an
> informative specification of the PIE algorithm and its possible variants.
>
> I'm not religious about this. Normative language might be useful. But on a
> pragmatic note, it will take considerable time and argument to decide which
> aspects are the essence of PIE and which are optional, because we will then
> have to consider whether certain combinations of options are not workable,
> or whether taking out too many of the optional parts makes it non-viable.
>
> What will be the point of being able to say that a particular
> implementation complies with an RFC that recorded at one snapshot in time
> what we thought defined the line between PIE and not PIE? If a future
> improvement is not described in the RFC, it will be non-compliant but
> better.
>
> BTW, wrt the CoDel draft, it does contain some normative language., but
> there are many aspects of the spec where it is not stated whether they are
> MUST, SHOULD or MAY. So this email really applies to both specs.
>
> 2) There /are/ interoperability concerns between AQMs, and between
> delay-based congestion controls (like LEDBAT) and AQMs. Writing a Proposed
> Standard giving the interoperability constraints on all AQMs would be a
> useful exercise. However, giving the special status of proposed standard to
> one design of PIE at one snapshot in time will not say anything about
> interoperability.
>
> 3) If someone comes up with an improvement on PIE (or CoDel) next week or
> next month or next year, will the IETF want to standardise it? Is the
> intention that AQM-chair will be a job for life?
>
>
> Bob
>
>
> ________________________________________________________________
> Bob Briscoe,                                                  BT
> _______________________________________________
> aqm mailing list
> aqm@ietf.org
> https://www.ietf.org/mailman/listinfo/aqm
>



-- 
Andrew McGregor | SRE | andrewmcgr@google.com | +61 4 1071 2221

--001a1135a2748c4d5f0514e81980
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">FWIW, the authors and editors have been thinking of CoDel =
as informational; I think this comes from a discussion post-chartering whic=
h converged to &quot;there is no one AQM to rule them all, therefore we won=
&#39;t be doing standards track&quot;.</div><div class=3D"gmail_extra"><br>=
<div class=3D"gmail_quote">On 30 April 2015 at 02:42, Bob Briscoe <span dir=
=3D"ltr">&lt;<a href=3D"mailto:bob.briscoe@bt.com" target=3D"_blank">bob.br=
iscoe@bt.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" sty=
le=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Richar=
d, Wes,<br>
<br>
1) The AQM charter says:<br>
&quot;Dec 2014 - Submit first algorithm specification to IESG for publicati=
on as Proposed Standard&quot;<br>
<br>
I volunteered to do a thorough review of the PIE draft, which I&#39;m just =
writing up. One of the problems is that it says &#39;Proposed Standard&#39;=
 at the top, but it&#39;s written in an informational style. There is no no=
rmative language saying what a PIE implementation MUST do, what it SHOULD d=
o and what it MAY do.<br>
<br>
And actually, I&#39;m not convinced that it would be worthwhile to add norm=
ative language. On reflection, I believe it would be better left as an info=
rmative specification of the PIE algorithm and its possible variants.<br>
<br>
I&#39;m not religious about this. Normative language might be useful. But o=
n a pragmatic note, it will take considerable time and argument to decide w=
hich aspects are the essence of PIE and which are optional, because we will=
 then have to consider whether certain combinations of options are not work=
able, or whether taking out too many of the optional parts makes it non-via=
ble.<br>
<br>
What will be the point of being able to say that a particular implementatio=
n complies with an RFC that recorded at one snapshot in time what we though=
t defined the line between PIE and not PIE? If a future improvement is not =
described in the RFC, it will be non-compliant but better.<br>
<br>
BTW, wrt the CoDel draft, it does contain some normative language., but the=
re are many aspects of the spec where it is not stated whether they are MUS=
T, SHOULD or MAY. So this email really applies to both specs.<br>
<br>
2) There /are/ interoperability concerns between AQMs, and between delay-ba=
sed congestion controls (like LEDBAT) and AQMs. Writing a Proposed Standard=
 giving the interoperability constraints on all AQMs would be a useful exer=
cise. However, giving the special status of proposed standard to one design=
 of PIE at one snapshot in time will not say anything about interoperabilit=
y.<br>
<br>
3) If someone comes up with an improvement on PIE (or CoDel) next week or n=
ext month or next year, will the IETF want to standardise it? Is the intent=
ion that AQM-chair will be a job for life?<br>
<br>
<br>
Bob<br>
<br>
<br>
________________________________________________________________<br>
Bob Briscoe,=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=
=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 BT=C2=A0 <br>
_______________________________________________<br>
aqm mailing list<br>
<a href=3D"mailto:aqm@ietf.org" target=3D"_blank">aqm@ietf.org</a><br>
<a href=3D"https://www.ietf.org/mailman/listinfo/aqm" target=3D"_blank">htt=
ps://www.ietf.org/mailman/listinfo/aqm</a><br>
</blockquote></div><br><br clear=3D"all"><div><br></div>-- <br><div class=
=3D"gmail_signature"><div dir=3D"ltr"><span style=3D"color:rgb(85,85,85);fo=
nt-family:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px=
 0px;border-style:solid;border-color:rgb(213,15,37);padding-top:2px;margin-=
top:2px">Andrew McGregor=C2=A0|</span><span style=3D"color:rgb(85,85,85);fo=
nt-family:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px=
 0px;border-style:solid;border-color:rgb(51,105,232);padding-top:2px;margin=
-top:2px">=C2=A0SRE=C2=A0|</span><span style=3D"color:rgb(85,85,85);font-fa=
mily:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px 0px;=
border-style:solid;border-color:rgb(0,153,57);padding-top:2px;margin-top:2p=
x">=C2=A0<a href=3D"mailto:andrewmcgr@google.com" target=3D"_blank">andrewm=
cgr@google.com</a>=C2=A0|</span><span style=3D"color:rgb(85,85,85);font-fam=
ily:sans-serif;font-size:small;line-height:1.5em;border-width:2px 0px 0px;b=
order-style:solid;border-color:rgb(238,178,17);padding-top:2px;margin-top:2=
px">=C2=A0+61 4 1071 2221</span><br></div></div>
</div>

--001a1135a2748c4d5f0514e81980--


From nobody Wed Apr 29 21:36:09 2015
Return-Path: <fred@cisco.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 415B21A92FF for <aqm@ietfa.amsl.com>; Wed, 29 Apr 2015 21:36:08 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -114.511
X-Spam-Level: 
X-Spam-Status: No, score=-114.511 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_HI=-5, SPF_PASS=-0.001, T_RP_MATCHES_RCVD=-0.01, USER_IN_DEF_DKIM_WL=-7.5, USER_IN_WHITELIST=-100] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id RDXMeruNKG-1 for <aqm@ietfa.amsl.com>; Wed, 29 Apr 2015 21:36:07 -0700 (PDT)
Received: from rcdn-iport-2.cisco.com (rcdn-iport-2.cisco.com [173.37.86.73]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (No client certificate requested) by ietfa.amsl.com (Postfix) with ESMTPS id E0C871A92FB for <aqm@ietf.org>; Wed, 29 Apr 2015 21:36:06 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=cisco.com; i=@cisco.com; l=1487; q=dns/txt; s=iport; t=1430368568; x=1431578168; h=from:to:cc:subject:date:message-id:references: in-reply-to:mime-version; bh=nXpKTKSvH7qxvKmtYRXrTsNNsyWroKF5v8Xe7gdAe8o=; b=cocDmVE+X8HZLOZJDlaGHytRxQ1nOtbPE5TDGTNUfRAUCvix8j66CZpS o0QdsuWNc6sZR84q73bkfxXYc2qfZT8nwppm12/90sxMRT2qU342hpESa VIIC+UrWI0qsIM6CjgUUAIjsutULWnI89+vtwceuaomBJba2TYAZtKyfd g=;
X-Files: signature.asc : 487
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: A0A4BQCEsEFV/4YNJK1cgwyBLwWDFspIAoFFTAEBAQEBAYELhCABAQEDASNWBQsCAQgYKgICMiUCBA4FDogVCLNOk1ABAQEBAQEBAQEBAQEBAQEBAQEBAQEXiziFBQeCaC+BFgWRaoF3gTiHGZV9I2CDFG+BRIEBAQEB
X-IronPort-AV: E=Sophos;i="5.11,675,1422921600";  d="asc'?scan'208";a="416278241"
Received: from alln-core-12.cisco.com ([173.36.13.134]) by rcdn-iport-2.cisco.com with ESMTP; 30 Apr 2015 04:36:07 +0000
Received: from xhc-rcd-x11.cisco.com (xhc-rcd-x11.cisco.com [173.37.183.85]) by alln-core-12.cisco.com (8.14.5/8.14.5) with ESMTP id t3U4a5Ha026269 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=FAIL); Thu, 30 Apr 2015 04:36:05 GMT
Received: from xmb-rcd-x09.cisco.com ([169.254.9.172]) by xhc-rcd-x11.cisco.com ([173.37.183.85]) with mapi id 14.03.0195.001; Wed, 29 Apr 2015 23:36:05 -0500
From: "Fred Baker (fred)" <fred@cisco.com>
To: Bob Briscoe <bob.briscoe@bt.com>
Thread-Topic: [aqm] PIE (and CoDel) drafts: proposed standard vs informational?
Thread-Index: AQHQgv8rewG8k+FlskCyP58ldo6KdA==
Date: Thu, 30 Apr 2015 04:36:05 +0000
Message-ID: <BE11122C-D4B6-443C-B282-5B59D46011DF@cisco.com>
References: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
In-Reply-To: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
Accept-Language: en-US
Content-Language: en-US
X-MS-Has-Attach: yes
X-MS-TNEF-Correlator: 
x-originating-ip: [10.19.64.121]
Content-Type: multipart/signed; boundary="Apple-Mail=_B7A549C8-CEF8-45A3-83D4-B05CFB07F49D"; protocol="application/pgp-signature"; micalg=pgp-sha1
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/d8yo1lqOJHVT-fZ7cmVgWmzCr74>
Cc: Richard Scheffenegger <rs@netapp.com>, AQM IETF list <aqm@ietf.org>, "Eddy Wesley M. \[VZ\]" <Wesley.M.Eddy@nasa.gov>
Subject: Re: [aqm] PIE (and CoDel) drafts: proposed standard vs informational?
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 30 Apr 2015 04:36:08 -0000

--Apple-Mail=_B7A549C8-CEF8-45A3-83D4-B05CFB07F49D
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=utf-8


> On Apr 29, 2015, at 9:42 AM, Bob Briscoe <bob.briscoe@bt.com> wrote:
>=20
> I volunteered to do a thorough review of the PIE draft, which I'm just =
writing up. One of the problems is that it says 'Proposed Standard' at =
the top, but it's written in an informational style. There is no =
normative language saying what a PIE implementation MUST do, what it =
SHOULD do and what it MAY do.

I=E2=80=99m not sure RFC 2026 says anything about that; that=E2=80=99s =
RFC 2119. PIE describes an algorithm.

--Apple-Mail=_B7A549C8-CEF8-45A3-83D4-B05CFB07F49D
Content-Transfer-Encoding: 7bit
Content-Disposition: attachment; filename="signature.asc"
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Message signed with OpenPGP using GPGMail

-----BEGIN PGP SIGNATURE-----
Comment: GPGTools - http://gpgtools.org

iQEVAwUBVUGxNJ9ieig10VPpAQLwvQf/e6CYSPwwa3kT0OmQGRc2bCXvRLq846Wm
zd+e8BdPZ3cZvab3hgIMYful3ZPDCCNAD/qx0RQWMv1t6yoe1qzpgFgNTCvWLVmP
7L6WQ0+R5fk/9z7t4gNtbsafWs77rYgAyvFxH+ChbWOGPWIiDFkaAEmkq+IQMxay
pPLUFg/Xs7qE4+6grbuN0wVooqQYnMiEtpzxNe7IOusRx13zrq4R9gcett/CfDdv
T2MQtO5DKLlR9P7m0IMHTxUPNGWON+UUhhpAB29B15XTZF2Y4BAV+NQB0rZW/k78
cyr2BSl8pVo6P1z6IVYEeNAGEnsDuCWKpMgA19JMUXcn8W+/R/8G5Q==
=bPtm
-----END PGP SIGNATURE-----

--Apple-Mail=_B7A549C8-CEF8-45A3-83D4-B05CFB07F49D--


From nobody Thu Apr 30 00:22:32 2015
Return-Path: <nicolas.kuhn@telecom-bretagne.eu>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 7383D1ACE9D for <aqm@ietfa.amsl.com>; Thu, 30 Apr 2015 00:22:31 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -1.55
X-Spam-Level: 
X-Spam-Status: No, score=-1.55 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, HELO_EQ_FR=0.35, HTML_MESSAGE=0.001, RCVD_IN_DNSWL_NONE=-0.0001, SPF_PASS=-0.001] autolearn=no
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id lJ3TWjYEKERU for <aqm@ietfa.amsl.com>; Thu, 30 Apr 2015 00:22:29 -0700 (PDT)
Received: from zproxy220.enst-bretagne.fr (zproxy220.enst-bretagne.fr [192.108.117.9]) by ietfa.amsl.com (Postfix) with ESMTP id F2E7C1ACE5B for <aqm@ietf.org>; Thu, 30 Apr 2015 00:22:28 -0700 (PDT)
Received: from localhost (localhost [127.0.0.1]) by zproxy220.enst-bretagne.fr (Postfix) with ESMTP id 18A2F301F3; Thu, 30 Apr 2015 09:22:28 +0200 (CEST)
Received: from zproxy220.enst-bretagne.fr ([127.0.0.1]) by localhost (zproxy220.enst-bretagne.fr [127.0.0.1]) (amavisd-new, port 10032) with ESMTP id Q0RUJhjJ7myO; Thu, 30 Apr 2015 09:22:27 +0200 (CEST)
Received: from localhost (localhost [127.0.0.1]) by zproxy220.enst-bretagne.fr (Postfix) with ESMTP id 140E93022E; Thu, 30 Apr 2015 09:22:27 +0200 (CEST)
X-Virus-Scanned: amavisd-new at zproxy220.enst-bretagne.fr
Received: from zproxy220.enst-bretagne.fr ([127.0.0.1]) by localhost (zproxy220.enst-bretagne.fr [127.0.0.1]) (amavisd-new, port 10026) with ESMTP id LSKEmOORh7I1; Thu, 30 Apr 2015 09:22:26 +0200 (CEST)
Received: from [IPv6:2001:660:7301:3728:2815:186b:71cb:375e] (passerelle-interne.enst-bretagne.fr [192.108.117.210]) by zproxy220.enst-bretagne.fr (Postfix) with ESMTPSA id D0C5D301F3; Thu, 30 Apr 2015 09:22:26 +0200 (CEST)
Content-Type: multipart/alternative; boundary="Apple-Mail=_7FE8E79E-64AD-4D8B-9B1E-DFAC6EDF19B7"
Mime-Version: 1.0 (Mac OS X Mail 8.0 \(1990.1\))
From: Nicolas Kuhn <nicolas.kuhn@telecom-bretagne.eu>
In-Reply-To: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
Date: Thu, 30 Apr 2015 09:22:26 +0200
Message-Id: <41FD727C-9AD2-4173-88D0-0FDC9CF863A2@telecom-bretagne.eu>
References: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
To: Bob Briscoe <bob.briscoe@bt.com>
X-Mailer: Apple Mail (2.1990.1)
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/fI1j-fzftRKDibgf1r5N8VGpbzk>
Cc: Richard Scheffenegger <rs@netapp.com>, AQM IETF list <aqm@ietf.org>, "Eddy Wesley M. \[VZ\]" <Wesley.M.Eddy@nasa.gov>
Subject: Re: [aqm] PIE (and CoDel) drafts: proposed standard vs informational?
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 30 Apr 2015 07:22:31 -0000

--Apple-Mail=_7FE8E79E-64AD-4D8B-9B1E-DFAC6EDF19B7
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii


> On 29 Apr 2015, at 18:42, Bob Briscoe <bob.briscoe@bt.com =
<mailto:bob.briscoe@bt.com>> wrote:
>=20
> Richard, Wes,
>=20
> 1) The AQM charter says:
> "Dec 2014 - Submit first algorithm specification to IESG for =
publication as Proposed Standard"
>=20
> I volunteered to do a thorough review of the PIE draft, which I'm just =
writing up. One of the problems is that it says 'Proposed Standard' at =
the top, but it's written in an informational style. There is no =
normative language saying what a PIE implementation MUST do, what it =
SHOULD do and what it MAY do.
>=20
> And actually, I'm not convinced that it would be worthwhile to add =
normative language. On reflection, I believe it would be better left as =
an informative specification of the PIE algorithm and its possible =
variants.
>=20
> I'm not religious about this. Normative language might be useful. But =
on a pragmatic note, it will take considerable time and argument to =
decide which aspects are the essence of PIE and which are optional, =
because we will then have to consider whether certain combinations of =
options are not workable, or whether taking out too many of the optional =
parts makes it non-viable.
>=20
> What will be the point of being able to say that a particular =
implementation complies with an RFC that recorded at one snapshot in =
time what we thought defined the line between PIE and not PIE? If a =
future improvement is not described in the RFC, it will be non-compliant =
but better.
>=20
> BTW, wrt the CoDel draft, it does contain some normative language., =
but there are many aspects of the spec where it is not stated whether =
they are MUST, SHOULD or MAY. So this email really applies to both =
specs.
>=20
> 2) There /are/ interoperability concerns between AQMs, and between =
delay-based congestion controls (like LEDBAT) and AQMs. Writing a =
Proposed Standard giving the interoperability constraints on all AQMs =
would be a useful exercise. However, giving the special status of =
proposed standard to one design of PIE at one snapshot in time will not =
say anything about interoperability.
>=20

I think that this point 2) should be considered in the evaluation =
guidelines.=20
I propose to add a scenario to assess the performance of delay-based =
congestion controls when there is an AQM.=20
The ToC would become:

4.  Transport Protocols=20
  4.1.  TCP-friendly sender=20
    4.1.1.  TCP-friendly sender with the same initial congestion  window
       4.1.2.  TCP-friendly sender with different initial congestion =
window
  4.2.  Aggressive transport sender
  4.3.  Unresponsive transport sender=20
  4.3.  Delay-based transport sender=20

What do you think ?

Nicolas


> 3) If someone comes up with an improvement on PIE (or CoDel) next week =
or next month or next year, will the IETF want to standardise it? Is the =
intention that AQM-chair will be a job for life?
>=20
>=20
> Bob
>=20
>=20
> ________________________________________________________________
> Bob Briscoe,                                                  BT =20
> _______________________________________________
> aqm mailing list
> aqm@ietf.org <mailto:aqm@ietf.org>
> https://www.ietf.org/mailman/listinfo/aqm =
<https://www.ietf.org/mailman/listinfo/aqm>


--Apple-Mail=_7FE8E79E-64AD-4D8B-9B1E-DFAC6EDF19B7
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html;
	charset=us-ascii

<html><head><meta http-equiv=3D"Content-Type" content=3D"text/html =
charset=3Dus-ascii"><meta http-equiv=3D"Content-Type" content=3D"text/html=
 charset=3Dus-ascii"><meta http-equiv=3D"Content-Type" =
content=3D"text/html charset=3Dus-ascii"></head><body style=3D"word-wrap: =
break-word; -webkit-nbsp-mode: space; -webkit-line-break: =
after-white-space;" class=3D""><br class=3D""><div class=3D""><blockquote =
type=3D"cite" class=3D""><div class=3D"">On 29 Apr 2015, at 18:42, Bob =
Briscoe &lt;<a href=3D"mailto:bob.briscoe@bt.com" =
class=3D"">bob.briscoe@bt.com</a>&gt; wrote:</div><br =
class=3D"Apple-interchange-newline"><div class=3D"">Richard, Wes,<br =
class=3D""><br class=3D"">1) The AQM charter says:<br class=3D"">"Dec =
2014 - Submit first algorithm specification to IESG for publication as =
Proposed Standard"<br class=3D""><br class=3D"">I volunteered to do a =
thorough review of the PIE draft, which I'm just writing up. One of the =
problems is that it says 'Proposed Standard' at the top, but it's =
written in an informational style. There is no normative language saying =
what a PIE implementation MUST do, what it SHOULD do and what it MAY =
do.<br class=3D""><br class=3D"">And actually, I'm not convinced that it =
would be worthwhile to add normative language. On reflection, I believe =
it would be better left as an informative specification of the PIE =
algorithm and its possible variants.<br class=3D""><br class=3D"">I'm =
not religious about this. Normative language might be useful. But on a =
pragmatic note, it will take considerable time and argument to decide =
which aspects are the essence of PIE and which are optional, because we =
will then have to consider whether certain combinations of options are =
not workable, or whether taking out too many of the optional parts makes =
it non-viable.<br class=3D""><br class=3D"">What will be the point of =
being able to say that a particular implementation complies with an RFC =
that recorded at one snapshot in time what we thought defined the line =
between PIE and not PIE? If a future improvement is not described in the =
RFC, it will be non-compliant but better.<br class=3D""><br =
class=3D"">BTW, wrt the CoDel draft, it does contain some normative =
language., but there are many aspects of the spec where it is not stated =
whether they are MUST, SHOULD or MAY. So this email really applies to =
both specs.<br class=3D""><br class=3D"">2) There /are/ interoperability =
concerns between AQMs, and between delay-based congestion controls (like =
LEDBAT) and AQMs. Writing a Proposed Standard giving the =
interoperability constraints on all AQMs would be a useful exercise. =
However, giving the special status of proposed standard to one design of =
PIE at one snapshot in time will not say anything about =
interoperability.<br class=3D""><br class=3D""></div></blockquote><div =
class=3D""><br class=3D""></div><div class=3D"">I think that this point =
2) should be considered in the evaluation guidelines.&nbsp;</div><div =
class=3D"">I propose to add a scenario to assess the performance of =
delay-based congestion controls when there is an AQM.&nbsp;</div><div =
class=3D"">The ToC would become:</div><div class=3D""><br =
class=3D""></div><div class=3D"">4. &nbsp;Transport =
Protocols&nbsp;</div><div class=3D"">&nbsp; 4.1. &nbsp;TCP-friendly =
sender&nbsp;<br class=3D"">&nbsp; &nbsp; 4.1.1. &nbsp;TCP-friendly =
sender with the same initial congestion&nbsp; window<br class=3D"">&nbsp; =
&nbsp; &nbsp; &nbsp;4.1.2. &nbsp;TCP-friendly sender with different =
initial congestion window<br class=3D"">&nbsp; 4.2. &nbsp;Aggressive =
transport sender<br class=3D"">&nbsp; 4.3. &nbsp;Unresponsive transport =
sender&nbsp;<br class=3D""><b class=3D"">&nbsp; 4.3. &nbsp;Delay-based =
transport sender</b>&nbsp;<br class=3D""></div><div class=3D""><br =
class=3D""></div><div class=3D"">What do you think ?</div><div =
class=3D""><br class=3D""></div><div class=3D"">Nicolas</div><div =
class=3D""><br class=3D""></div><br class=3D""><blockquote type=3D"cite" =
class=3D""><div class=3D"">3) If someone comes up with an improvement on =
PIE (or CoDel) next week or next month or next year, will the IETF want =
to standardise it? Is the intention that AQM-chair will be a job for =
life?<br class=3D""><br class=3D""><br class=3D"">Bob<br class=3D""><br =
class=3D""><br =
class=3D"">_______________________________________________________________=
_<br class=3D"">Bob Briscoe, =
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&n=
bsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbs=
p;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;=
&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;BT=
 &nbsp;<br class=3D"">_______________________________________________<br =
class=3D"">aqm mailing list<br class=3D""><a href=3D"mailto:aqm@ietf.org" =
class=3D"">aqm@ietf.org</a><br class=3D""><a =
href=3D"https://www.ietf.org/mailman/listinfo/aqm" =
class=3D"">https://www.ietf.org/mailman/listinfo/aqm</a><br =
class=3D""></div></blockquote></div><br class=3D""></body></html>=

--Apple-Mail=_7FE8E79E-64AD-4D8B-9B1E-DFAC6EDF19B7--


From nobody Thu Apr 30 05:43:36 2015
Return-Path: <wes@mti-systems.com>
X-Original-To: aqm@ietfa.amsl.com
Delivered-To: aqm@ietfa.amsl.com
Received: from localhost (ietfa.amsl.com [127.0.0.1]) by ietfa.amsl.com (Postfix) with ESMTP id 1E8D81B29A7 for <aqm@ietfa.amsl.com>; Thu, 30 Apr 2015 05:43:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at amsl.com
X-Spam-Flag: NO
X-Spam-Score: -0.5
X-Spam-Level: 
X-Spam-Status: No, score=-0.5 tagged_above=-999 required=5 tests=[BAYES_05=-0.5] autolearn=ham
Received: from mail.ietf.org ([4.31.198.44]) by localhost (ietfa.amsl.com [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id qvJA0g4Pp-Zg for <aqm@ietfa.amsl.com>; Thu, 30 Apr 2015 05:43:34 -0700 (PDT)
Received: from atl4mhob09.myregisteredsite.com (atl4mhob09.myregisteredsite.com [209.17.115.47]) by ietfa.amsl.com (Postfix) with ESMTP id 81BF01B29A4 for <aqm@ietf.org>; Thu, 30 Apr 2015 05:43:34 -0700 (PDT)
Received: from mailpod.hostingplatform.com ([10.30.71.204]) by atl4mhob09.myregisteredsite.com (8.14.4/8.14.4) with ESMTP id t3UChWtZ024770 for <aqm@ietf.org>; Thu, 30 Apr 2015 08:43:32 -0400
Received: (qmail 19533 invoked by uid 0); 30 Apr 2015 12:43:32 -0000
X-TCPREMOTEIP: 24.166.126.82
X-Authenticated-UID: wes@mti-systems.com
Received: from unknown (HELO ?192.168.0.3?) (wes@mti-systems.com@24.166.126.82) by 0 with ESMTPA; 30 Apr 2015 12:43:32 -0000
Message-ID: <55422368.5030802@mti-systems.com>
Date: Thu, 30 Apr 2015 08:43:20 -0400
From: Wesley Eddy <wes@mti-systems.com>
Organization: MTI Systems
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:31.0) Gecko/20100101 Thunderbird/31.6.0
MIME-Version: 1.0
To: Bob Briscoe <bob.briscoe@bt.com>, Richard Scheffenegger <rs@netapp.com>, "Eddy Wesley M. [VZ]" <Wesley.M.Eddy@nasa.gov>
References: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
In-Reply-To: <201504291642.t3TGgFqH012984@bagheera.jungle.bt.co.uk>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Archived-At: <http://mailarchive.ietf.org/arch/msg/aqm/XyBu7SbpCclJFSvC9lAOgyQmmIc>
Cc: AQM IETF list <aqm@ietf.org>
Subject: Re: [aqm] PIE (and CoDel) drafts: proposed standard vs informational?
X-BeenThere: aqm@ietf.org
X-Mailman-Version: 2.1.15
Precedence: list
List-Id: "Discussion list for active queue management and flow isolation." <aqm.ietf.org>
List-Unsubscribe: <https://www.ietf.org/mailman/options/aqm>, <mailto:aqm-request@ietf.org?subject=unsubscribe>
List-Archive: <http://www.ietf.org/mail-archive/web/aqm/>
List-Post: <mailto:aqm@ietf.org>
List-Help: <mailto:aqm-request@ietf.org?subject=help>
List-Subscribe: <https://www.ietf.org/mailman/listinfo/aqm>, <mailto:aqm-request@ietf.org?subject=subscribe>
X-List-Received-Date: Thu, 30 Apr 2015 12:43:36 -0000

On 4/29/2015 12:42 PM, Bob Briscoe wrote:
> Richard, Wes,
> 
> 1) The AQM charter says:
> "Dec 2014 - Submit first algorithm specification to IESG for publication
> as Proposed Standard"
> 


Hi Bob, thanks for raising this, since it probably requires some
clarification and discussion.  I thought we'd outlined this a bit
on the list and through discussion a couple meetings ago, but it
might not have been integrated back into the charter.

I think the intention when chartering was to enable us to put
something on Standards Track, if there is consensus, but not
to limit us to only Standards Track.

It seems apparent that there are drafts which the group has
agreed are worth publishing (the ones we've adopted), and part
of submitting them for publication is deciding on the status
they'll be stamped with.  If the group wants to do Informational
or Experimental and feels those are more appropriate, then
obviously that's what we'll do.

I'm pretty sure our ADs support that, but am happy for them to
jump in and correct it, if not!

-- 
Wes Eddy
MTI Systems

