From: George Goglidze (goglidze@gmail.com)
Date: Sat Oct 20 2007 - 13:21:09 ART
Hi,
The traffic shaping works like this:
if it has too big packet, and it exceeds all the tokens it has accumulated
until now.
then the packet is queued, until ofcourse it has enough tokens to send the
whole packet.
so when you have no other traffic, and you make just a ping, the tokens are
easily accumulated,
and the packet is sent after all.
but now let's make the following test.
R6 -> R2 -> R4
R2 config is following:
class-map match-all SHAPE_2
match access-group 198
access-list 198 permit ip any any
!
policy-map SHAPE
class SHAPE_2
shape average 8000 1000 0
!
interface Serial2/0.204 point-to-point !! INTERFACE THAT GOES TO R4
ip address 150.50.24.2 255.255.255.0
frame-relay interface-dlci 204
service-policy output SHAPE
!
now we make a normal ping from R2 to R4 ->
ping 150.50.24.4 source f0/0 repeat 100000
!! little packets, so that we eat up all the tokens. and repeat many times
so that
!! we use tokens all the time.
it succeeds.
now we try and make ping with size 1500 from R6 router,
it will fail, because 1500 bytes will need to wait in a queue until it has
enough tokens all together
to send 1500 bytes +32. this will never happen because we have ping from
R2 to R4 with normal size,
which does not have to wait because it can be sent in less tokens then the
ping of 1500 bytes.
Hope this helps,
Good luck,
On 10/20/07, Brian Dennis <bdennis@internetworkexpert.com> wrote:
>
> You can set the ping timeout to 0 and the router will generate the pings
> as fast as it can without waiting for a reply.
>
> ping 10.1.1.1 timeout 0
>
> Brian Dennis, CCIE4 #2210 (R&S/ISP-Dial/Security/SP)
> bdennis@internetworkexpert.com
>
> Internetwork Expert, Inc.
> http://www.InternetworkExpert.com
> Toll Free: 877-224-8987
> Direct: 775-745-6404 (Outside the US and Canada)
>
>
> >----- Original Message -----
> Subject: Re: Effect of shaping with small Bc, no fragmentation, and a
> large packet?
> Date: Fri, October 19, 2007 10:10
> From: "Gregory Gombas" <ggombas@gmail.com>
>
> > Hi Andrew,
> >
> > I'm not sure ping test is adequate since the router waits for a reply
> > before sending the next packet...I think you need a UDP application to
> > send packets at a fixed rate with specified packet size to really test
> > it....
> >
> >
> > On 10/19/07, Usankin, Andrew <Andrew.Usankin@twtelecom.com> wrote:
> > >
> > > Very good question indeed. I used to thing that packet will go through
> > > but now I'm going to dig dipper to make double sure it is.
> > > For now, here is what I got from labbing:
> > >
> > > R1#sh run int s1/0
> > > Building configuration...
> > >
> > > Current configuration : 129 bytes
> > > !
> > > interface Serial1/0
> > > ip address 172.16.1.1 255.255.255.0
> > > serial restart-delay 0
> > > traffic-shape rate 32000 3200 3200 1000
> > > end
> > >
> > > R1#sh traffic-shape s1/0
> > >
> > > Interface Se1/0
> > > Access Target Byte Sustain Excess Interval Increment
> > > Adapt
> > > VC List Rate Limit bits/int bits/int (ms) (bytes)
> > > Active
> > > - 32000 800 3200 3200 100 400
> -
> > > R1#ping 172.16.1.2 size 1500
> > >
> > > Type escape sequence to abort.
> > > Sending 5, 1500-byte ICMP Echos to 172.16.1.2, timeout is 2 seconds:
> > > !!!!!
> > > Success rate is 100 percent (5/5), round-trip min/avg/max = 12/282/412
> > > ms
> > > R1#
> > >
> > >
> > > As you can see, bytes limit per interval is 800 and I'm sending
> packets
> > > of 1500+24 bites size. They all make it through.
> > > Andrew
> > >
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
> Of
> > > Gregory Gombas
> > > Sent: Thursday, October 18, 2007 7:27 PM
> > > To: Eric Phillips
> > > Cc: ccielab@groupstudy.com
> > > Subject: Re: Effect of shaping with small Bc, no fragmentation, and a
> > > large packet?
> > >
> > > I didn't mean it literally, I just meant that since the packet will
> > > always be exceeding bc it can never be sent...and thus dropped.
> > >
> > > Only way to know for sure is to lab it up!
> > >
> > > On 10/18/07, Eric Phillips <ephillips@squick.cc> wrote:
> > > > Thanks everyone for the replies, I definitely appreciate it!
> > > >
> > > > I did skip a 0 there, sorry... I was aiming for a 10ms Tc to make
> > > > VOIP traffic delayed as little as possible. So yes, the Bc would be
> > > 640 bits.
> > > >
> > > > Herbert, my numbers were referring to Shaping, so I was keeping
> > > > everything in bits. With policing you deal in bytes.
> > > >
> > > > I am very intrigued to see that we have three answers, so perhaps I
> am
> > >
> > > > not the only one that is a bit confused by this. If it is queued
> > > > forever, does it block the pipe while it is queued?
> > > >
> > > > -Eric
> > > >
> > > >
> > > >
> > > > On 10/18/07, Gregory Gombas <ggombas@gmail.com> wrote:
> > > > >
> > > > > Even with shaping configured, if the router queues the packet then
> > > > > it would be queued indefinately considering it will always be
> larger
> > >
> > > > > than bc and never be able to send it...
> > > > >
> > > > > On 10/18/07, Herbert Maosa <asawilunda@googlemail.com> wrote:
> > > > > > Eric,
> > > > > >
> > > > > > I assume you are maintaining bytes as the unit for Bc and ms as
> > > > > > the unit
> > > > > for
> > > > > > TC ? In that case Bc of 64 Bytes and Tc of 10ms is giving me CIR
> =
> > > > > 512Kbps,
> > > > > > using the formula CIR = Bc/Tc.
> > > > > >
> > > > > > In any case, what will happen to a packet that is larger than
> the
> > > > > maximum
> > > > > > rate permitted will depend on whether you are shaping or
> policing.
> > >
> > > > > > If
> > > > > you
> > > > > > are shaping, then the packet will simply be queued in the
> shaping
> > > > > > queue
> > > > > from
> > > > > > which it will then be sent at CIR. If you are policing then it
> > > > > > will
> > > > > depend
> > > > > > on the configured action for non-conforming traffic.
> > > > > >
> > > > > > Herbert.
> > > > > >
> > > > > >
> > > > > > On 10/18/07, Eric Phillips <ephillips@squick.cc> wrote:
> > > > > > >
> > > > > > > Hi all,
> > > > > > >
> > > > > > > I have searched through the GroupStudy archives, read through
> > > > > > > Odem's
> > > > > DQOS
> > > > > > > book, and checked out the DOC CD, but have not found a clear
> > > > > > > answer to
> > > > > a
> > > > > > > question that has been on my mind.
> > > > > > >
> > > > > > > If I enable shaping with a very small Bc, for example, a Tc of
> > > > > > > 10, and
> > > > > a
> > > > > > > Bc
> > > > > > > of 64, that should shape to 64kbit/sec, right? Now if I do
> not
> > > > > > > enable link fragmentation, and a large packet, perhaps 1400
> > > > > > > bytes comes through,
> > > > > what
> > > > > > > will happen? It can not be sent in a single Tc, and it can
> not
> > > > > > > be fragmented because link fragmentation is disabled. Also
> > > > > > > assume that I have no Be, or a Be too small to allow 1400
> bytes
> > > > > > > to accumulate between Bc
> > > > > and
> > > > > > > Be.
> > > > > > >
> > > > > > > This seems like an awfully simple question, but I have not
> found
> > >
> > > > > > > a
> > > > > clear
> > > > > > > answer anywhere.
> > > > > > >
> > > > > > > Thank you very much,
> > > > > > >
> > > > > > > Eric
> > > > > > >
> > > > > > >
> > > > >
> ____________________________________________________________________
> > > > > ___
> > > > > > > Subscription information may be found at:
> > > > > > > http://www.groupstudy.com/list/CCIELab.html
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Kindest regards,
> > > > > > hm
> > > > > >
> > > > > >
> __________________________________________________________________
> > > > > > _____ Subscription information may be found at:
> > > > > > http://www.groupstudy.com/list/CCIELab.html
> > > >
> > > >
> ______________________________________________________________________
> > > > _ Subscription information may be found at:
> > > > http://www.groupstudy.com/list/CCIELab.html
> > >
> > >
> _______________________________________________________________________
> > > Subscription information may be found at:
> > > http://www.groupstudy.com/list/CCIELab.html
> > >
> > > The content contained in this electronic message is not intended to
> > > constitute formation of a contract binding TWTC. TWTC will be
> > > contractually bound only upon execution, by an authorized officer, of
> > > a contract including agreed terms and conditions or by express
> > > application of its tariffs.
> > >
> > > This message is intended only for the use of the individual or entity
> > > to which it is addressed. If the reader of this message is not the
> > > intended recipient, or the employee or agent responsible for
> > > delivering the message to the intended recipient, you are hereby
> > > notified that any dissemination, distribution or copying of this
> > > message is strictly prohibited. If you have received this
> > > communication in error, please notify us immediately by replying to
> > > the sender of this E-Mail or by telephone.
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Fri Nov 16 2007 - 13:11:17 ART