Re: [Diffserv-interest] Re: Diffserv-interest digest, Vol 1 #113 - 3 msgs
"John H. Shuler" <johnshuler@mac.com> Mon, 10 November 2003 19:56 UTC
Received: from optimus.ietf.org ([132.151.1.19]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09065 for <diffserv-interest-archive@odin.ietf.org>; Mon, 10 Nov 2003 14:56:19 -0500 (EST)
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org) by optimus.ietf.org with esmtp (Exim 4.20) id 1AJI8p-0005VH-9Q for diffserv-interest-archive@odin.ietf.org; Mon, 10 Nov 2003 14:56:03 -0500
Received: (from exim@localhost) by www1.ietf.org (8.12.8/8.12.8/Submit) id hAAJu2NY021074 for diffserv-interest-archive@odin.ietf.org; Mon, 10 Nov 2003 14:56:02 -0500
Received: from localhost.localdomain ([127.0.0.1] helo=www1.ietf.org) by optimus.ietf.org with esmtp (Exim 4.20) id 1AJI8o-0005Tn-RP; Mon, 10 Nov 2003 14:56:02 -0500
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org) by optimus.ietf.org with esmtp (Exim 4.20) id 1AJI88-0005R4-O3 for diffserv-interest@optimus.ietf.org; Mon, 10 Nov 2003 14:55:20 -0500
Received: from ietf-mx (ietf-mx.ietf.org [132.151.6.1]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id OAA09027 for <diffserv-interest@ietf.org>; Mon, 10 Nov 2003 14:55:06 -0500 (EST)
Received: from ietf-mx ([132.151.6.1]) by ietf-mx with esmtp (Exim 4.12) id 1AJI85-0003Nr-00 for diffserv-interest@ietf.org; Mon, 10 Nov 2003 14:55:17 -0500
Received: from a17-250-248-85.apple.com ([17.250.248.85] helo=smtpout.mac.com) by ietf-mx with esmtp (Exim 4.12) id 1AJI84-0003No-00 for diffserv-interest@ietf.org; Mon, 10 Nov 2003 14:55:16 -0500
Received: from mac.com (smtpin08-en2 [10.13.10.153]) by smtpout.mac.com (Xserve/MantshX 2.0) with ESMTP id hAAJtFIZ015698; Mon, 10 Nov 2003 11:55:15 -0800 (PST)
Received: from [192.168.0.2] (12-235-5-14.client.attbi.com [12.235.5.14]) (authenticated bits=0) by mac.com (Xserve/smtpin08/MantshX 3.0) with ESMTP id hAAJtEZq014350; Mon, 10 Nov 2003 11:55:15 -0800 (PST)
User-Agent: Microsoft-Entourage/10.1.4.030702.0
Date: Mon, 10 Nov 2003 11:55:12 -0800
Subject: Re: [Diffserv-interest] Re: Diffserv-interest digest, Vol 1 #113 - 3 msgs
From: "John H. Shuler" <johnshuler@mac.com>
To: Brian E Carpenter <brc@zurich.ibm.com>
CC: diffserv-interest@ietf.org
Message-ID: <BBD52D20.14B1A%johnshuler@mac.com>
In-Reply-To: <3FAFEBE6.52164C4@zurich.ibm.com>
Mime-version: 1.0
Content-type: text/plain; charset="US-ASCII"
Content-transfer-encoding: 7bit
Content-Transfer-Encoding: 7bit
Sender: diffserv-interest-admin@ietf.org
Errors-To: diffserv-interest-admin@ietf.org
X-BeenThere: diffserv-interest@ietf.org
X-Mailman-Version: 2.0.12
Precedence: bulk
List-Unsubscribe: <https://www1.ietf.org/mailman/listinfo/diffserv-interest>, <mailto:diffserv-interest-request@ietf.org?subject=unsubscribe>
List-Id: Differentiated services general discussion <diffserv-interest.ietf.org>
List-Post: <mailto:diffserv-interest@ietf.org>
List-Help: <mailto:diffserv-interest-request@ietf.org?subject=help>
List-Subscribe: <https://www1.ietf.org/mailman/listinfo/diffserv-interest>, <mailto:diffserv-interest-request@ietf.org?subject=subscribe>
Content-Transfer-Encoding: 7bit
Brian, I agree. My suggestion would be to keep it simple so as to be easily implemented as a common denominator standard that works, then refine it over time after experience demonstrates a need. If it starts out too complicated, it will be too expensive to implement and therefore usurp the purpose. j > From: Brian E Carpenter <brc@zurich.ibm.com> > Organization: IBM > Date: Mon, 10 Nov 2003 20:49:58 +0100 > To: "John H. Shuler" <johnshuler@mac.com> > Cc: diffserv-interest@ietf.org > Subject: Re: [Diffserv-interest] Re: Diffserv-interest digest, Vol 1 #113 - 3 > msgs > > Yes, WFQ is an example solution of course, but it's a relatively easy one to > think > about. > > I do agree we need a best practices document; we're only arguing about its > richness in service classes, I think. > > Brian > > "John H. Shuler" wrote: >> >> Brian, >> >> I don't believe there is a mandate for SPs to use WFQ within the AF class, >> so either WFQ or strict priority queuing is an option. As well, the nature >> and degree of priority -- in fact, the entire policy equation -- seems to be >> left open, except that there is an implication that AF4 is given some higher >> quality/priority than AF1, whether in a weighted or strict sense. >> >> It is my contention, however, that there should be some kind of specific >> guidelines provided for this class, especially, since it is in AF where a >> lot of the subtlety -- and confusion, and cross-carrier screwups -- can >> reside. This is where I believe the proposed draft can have its biggest >> impact. That is, not only in delineating further the treatment of AF within >> class, but also the rules in assigning traffic types to different classes. >> >> The whole idea is that, without some kind of standardization, multi-service >> IP will not take off as it should, because there is no way for a customer to >> know that the video (for example) he hands to his regional SP will make it >> to the other end in a consistent and reliable way. >> >> j >> >>> From: diffserv-interest-request@ietf.org >>> Reply-To: diffserv-interest@ietf.org >>> Date: Mon, 10 Nov 2003 12:00:03 -0500 >>> To: diffserv-interest@ietf.org >>> Subject: Diffserv-interest digest, Vol 1 #113 - 3 msgs >>> >>> Send Diffserv-interest mailing list submissions to >>> diffserv-interest@ietf.org >>> >>> To subscribe or unsubscribe via the World Wide Web, visit >>> https://www1.ietf.org/mailman/listinfo/diffserv-interest >>> or, via email, send a message with subject or body 'help' to >>> diffserv-interest-request@ietf.org >>> >>> You can reach the person managing the list at >>> diffserv-interest-admin@ietf.org >>> >>> When replying, please edit your Subject line so it is more specific >>> than "Re: Contents of Diffserv-interest digest..." >>> >>> >>> Today's Topics: >>> >>> 1. Re: QoS in diffserv network (Brian E Carpenter) >>> 2. Re: QoS in diffserv network (Zoltan.Balint@alcatel.be) >>> 3. Re: QoS in diffserv network (Brian E Carpenter) >>> >>> --__--__-- >>> >>> Message: 1 >>> Date: Sun, 09 Nov 2003 19:26:48 +0100 >>> From: Brian E Carpenter <brc@zurich.ibm.com> >>> Organization: IBM >>> To: Feng Y <feng6@uwindsor.ca> >>> CC: diffserv-interest@ietf.org >>> Subject: Re: [Diffserv-interest] QoS in diffserv network >>> >>> below... >>> >>> Feng Y wrote: >>>> = >>> >>>>>> >>>>>> It is up to the carrier to monitor congestion >>>> periodically to prevent >>>>>> lowest-priority traffic from being too congested. EF >>>> should ALWAYS get >>>>>> through on-time, and AF CIR should always get through, >>>> if somewhat delayed >>>>>> and jittery. The cases when this is NOT true should be >>>> rare and bizarre >>>>>> (earthquakes, major network outages, etc.) unless the >>>> carrier is not doing >>>>>> their job... Which is simply to make sure they engineer >>>> total network >>>>>> bandwidth above the level of peak CIR. >>>>> >>>>> Exactly. Each traffic class (i.e. each queue) needs to be >>>> provisioned for >>>>> the expected load with appropriate over- or >>>> under-provisioning. The new >>>>> thing is that you can over-provision for some traffic and >>>> under-provision >>>>> for other traffic. (New to IP, that is. It's hardly a new >>>> concept.) >>>>> >>>> = >>> >>>> Can we always assume that a higher service class provides >>>> the same or a better service than any lower service class >>>> in AF? = >>> >>> >>> No. There is no "higher" or "lower" built into AF. There >>> is simply the option of having one or more mutually >>> independent AF service classes, with an arbitrary number >>> of four service classes being named AF1, AF2, AF3, AF4 >>> if you choose to use the recommended code points. >>> >>> If you happen to give each of these classes equal weights >>> in a WFQ scheme they are all offering the same level of service. >>> >>> If you happen to give AF2 a higher weight than AF3, you are >>> choosing to offer more service to AF2 traffic. >>> >>> >>>> Can the following scenario happen in Diffserv >>>> network? >>>> The gold service was allocated x percent of a link=92s >>>> bandwidth and silver service was allocated x/2 percent of >>>> the link=92s bandwidth, but the traffic intensity of gold >>>> packets was 10 times higher than that of silver packets. >>> >>> That can certainly happen, if you mean that the offered >>> load to the gold service is ten times greater than the >>> offered load to the silver service. >>> >>>> = >>> >>>> Should the ISP assure that there are not so many gold >>>> packets in service? >>> >>> It depends what the SLA for gold and silver say, doesn't >>> it? There is no rule. It's the ISP's business decision. >>> >>> My assumption and what I would recommend to an ISP is: >>> >>> If the offered load on silver equals the bandwidth assigned >>> to silver, then it should all be transmitted, unless the SLA says >>> otherwise. And the excess traffic offered to the gold service = >>> >>> will be at high risk of being dropped, depending on the utilization >>> of the remaining (100-x-x/2)% of the line. >>> >>> Brian >>> >>> >>> --__--__-- >>> >>> Message: 2 >>> Date: Mon, 10 Nov 2003 09:21:15 +0100 >>> From: Zoltan.Balint@alcatel.be >>> Organization: Alcatel, >>> To: Brian E Carpenter <brc@zurich.ibm.com> >>> Cc: Feng Y <feng6@uwindsor.ca>, diffserv-interest@ietf.org >>> Subject: Re: [Diffserv-interest] QoS in diffserv network >>> >>> <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> >>> <html> >>> <head> >>> <meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1"> >>> <title></title> >>> </head> >>> <body text="#000000" bgcolor="#ffffff"> >>> Brian, Feng,<br> >>> <blockquote type="cite" cite="mid3FAE86E8.3AE456E7@zurich.ibm.com"> >>> <blockquote type="cite"> >>> <pre wrap="">Can we always assume that a higher service class provides >>> the same or a better service than any lower service class >>> in AF? >>> </pre> >>> </blockquote> >>> <pre wrap=""><!----> >>> No. There is no "higher" or "lower" built into AF. There >>> is simply the option of having one or more mutually >>> independent AF service classes, with an arbitrary number >>> of four service classes being named AF1, AF2, AF3, AF4 >>> if you choose to use the recommended code points. >>> >>> If you happen to give each of these classes equal weights >>> in a WFQ scheme they are all offering the same level of service.</pre> >>> </blockquote> >>> That would be the case when the relative load in those classes was >>> equal as well. But, since some<br> >>> classes can be more loaded than others, the service seen in different >>> classes does not <br> >>> relate to the index of the AF class. Unless, of course, the scheduler >>> was strict priority, which does<br> >>> not conform to the requirements in RFC2597 (minimum bandwidth).<br> >>> <br> >>> <br> >>> Regards,<br> >>> Zoltan<br> >>> <br> >>> <snip><br> >>> </body> >>> </html> >>> >>> >>> >>> --__--__-- >>> >>> Message: 3 >>> Date: Mon, 10 Nov 2003 16:55:17 +0100 >>> From: Brian E Carpenter <brc@zurich.ibm.com> >>> Organization: IBM >>> To: Zoltan.Balint@alcatel.be >>> CC: Feng Y <feng6@uwindsor.ca>, diffserv-interest@ietf.org >>> Subject: Re: [Diffserv-interest] QoS in diffserv network >>> >>>> Zoltan.Balint@alcatel.be wrote: >>>> >>>> Brian, Feng, >>>> >>>>>> Can we always assume that a higher service class provides >>>>>> the same or a better service than any lower service class >>>>>> in AF? >>>>>> >>>>>> >>>>> No. There is no "higher" or "lower" built into AF. There >>>>> is simply the option of having one or more mutually >>>>> independent AF service classes, with an arbitrary number >>>>> of four service classes being named AF1, AF2, AF3, AF4 >>>>> if you choose to use the recommended code points. >>>>> >>>>> If you happen to give each of these classes equal weights >>>>> in a WFQ scheme they are all offering the same level of service. >>>>> >>>> That would be the case when the relative load in those classes was equal as >>>> well. But, since some >>>> classes can be more loaded than others, the service seen in different >>>> classes >>>> does not >>>> relate to the index of the AF class. >>> >>> Zoltan, >>> >>> There is no such thing as an index; the digit in AF3 is just part of the >>> name. >>> >>> If two AF classes have equal WFQ weights but different offered loads, they >>> will >>> experience different loss and jitter rates, but the service *offered* is the >>> same. >>> The service *delivered* will vary, but that is a different matter. The two >>> SLAs may >>> be different, by allowing different levels of overbooking, but that is why >>> the >>> diffserv >>> documents distinguish between the SLA and the SLS [service level >>> specification]. >>> >>> I think we are in agreement, but I want to be precise. >>> >>>> Unless, of course, the scheduler was strict priority, which does >>>> not conform to the requirements in RFC2597 (minimum bandwidth). >>> >>> True. >>> >>> Brian >>> >>> >>> >>> --__--__-- >>> >>> _______________________________________________ >>> Diffserv-interest mailing list >>> Diffserv-interest@ietf.org >>> https://www1.ietf.org/mailman/listinfo/diffserv-interest >>> >>> >>> End of Diffserv-interest Digest >> >> _______________________________________________ >> Diffserv-interest mailing list >> Diffserv-interest@ietf.org >> https://www1.ietf.org/mailman/listinfo/diffserv-interest > > -- > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - > Brian E Carpenter > Distinguished Engineer, Internet Standards & Technology, IBM > > NEW ADDRESS <brc@zurich.ibm.com> PLEASE UPDATE ADDRESS BOOK _______________________________________________ Diffserv-interest mailing list Diffserv-interest@ietf.org https://www1.ietf.org/mailman/listinfo/diffserv-interest
- [Diffserv-interest] Re: Diffserv-interest digest,… John H. Shuler
- Re: [Diffserv-interest] Re: Diffserv-interest dig… Brian E Carpenter
- Re: [Diffserv-interest] Re: Diffserv-interest dig… John H. Shuler
- Re: [Diffserv-interest] Re: Diffserv-interest dig… Jing Shen
- Re: [Diffserv-interest] Re: Diffserv-interest dig… Brian E Carpenter
- [Diffserv-interest] Overbooking John H. Shuler