From: Brian McGahan (bmcgahan@internetworkexpert.com)
Date: Sun Apr 23 2006 - 18:41:04 GMT-3
I'm not sure of the order of operations as the packet moves to
the transmit ring, but how does affect the behavior? Whether it's
fragmented inside the shaping queue or when it leaves the shaping queue
it's still fragmented when it goes onto the transmit ring.
Brian McGahan, CCIE #8593
bmcgahan@internetworkexpert.com
Internetwork Expert, Inc.
http://www.InternetworkExpert.com
Toll Free: 877-224-8987 x 705
Outside US: 775-826-4344 x 705
24/7 Support: http://forum.internetworkexpert.com
Live Chat: http://www.internetworkexpert.com/chat/
> -----Original Message-----
> From: allboutcisco [mailto:frenzeus@streamyx.com]
> Sent: Saturday, April 22, 2006 11:45 PM
> To: Brian McGahan; Scott Smith
> Cc: Cisco certification
> Subject: Re: IE Lab10 - FRTS
>
> Hi Brian,
>
> Thanks for taking the time to clarify on this. Much appreciated. Just
a
> few
> questions that popped up in my mind as i read through...
>
> So if i understand correctly, the clocking rate (AR) definitely
affects
> the
> serialization delay of any packets being transmitted out the interface
> transmit ring. The only way to actually "slow" down the rate of
> transmission
> , so to speak is to actually configure shaping on the interface. Now
> shaping
> queues would usually form on an interface if a certain sized-packet
that
> arrives actually is larger than the configured Bc of the shaping
queues,
> where this packet is actually "shaped" to the rate configured with the
> CIR,
> Bc & Tc.
>
> Now would the packet actually be fragmented before it was placed into
the
> shaping queue or when it is being de-queued to the tx-ring? From your
> explanation it seems that fragmentation occurs after a packet is being
> de-queued from the shaping queue & before it is being placed into the
> tx-ring, thereby the maximum fragment size which is now less than or
equal
> to the Bc & hence will not take more than a Tc to be actually
transmitted
> out the interface.
>
> Hope i've understood correctly, if not really appreciate if i could be
> pointed out to the right direction.
>
> Thanks again brian.
>
> -K
>
> ----- Original Message -----
> From: "Brian McGahan" <bmcgahan@internetworkexpert.com>
> To: "Kenny" <frenzeus@streamyx.com>; "Scott Smith"
<hioctane@gmail.com>
> Cc: "Cisco certification" <ccielab@groupstudy.com>
> Sent: Sunday, April 23, 2006 8:33 AM
> Subject: RE: IE Lab10 - FRTS
>
>
> The transmission time is a function of the serialization delay
> on the transmit ring of the interface, not of the shaping
configuration.
> Take the size of the packet in bits and divide it by the interface
> clocking in bits per second and you get the amount of seconds it takes
> to transmit the packet; this is the serialization delay of that
packet.
>
>
> When shaping is configured all packets sent to the interface go
> into the shaping queue. Those that are above the configured fragment
> size are fragmented before being sent to the interface to be
serialized
> (sent). The rate a which packets or fragments are dequeued from the
> shaping queue to the transmit ring is based on the CIR, Bc, and Tc.
In
> this manner if the maximum packet fragment size is less than the Bc it
> will take no longer than one Tc for any single packet to be dequeued
and
> serialized.
>
>
> HTH,
>
> Brian McGahan, CCIE #8593
> bmcgahan@internetworkexpert.com
>
> Internetwork Expert, Inc.
> http://www.InternetworkExpert.com
> Toll Free: 877-224-8987 x 705
> Outside US: 775-826-4344 x 705
> 24/7 Support: http://forum.internetworkexpert.com
> Live Chat: http://www.internetworkexpert.com/chat/
>
>
> > -----Original Message-----
> > From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf
> Of
> > Kenny
> > Sent: Friday, April 21, 2006 1:50 PM
> > To: Scott Smith
> > Cc: Cisco certification
> > Subject: Re: IE Lab10 - FRTS
> >
> > Hi Scott,
> >
> > I was searching out for this question of Lab10 as well & happen to
> bump
> > into
> > ur mail. I'm actually more concern on specifically when does shaping
&
> > fragmentation actually take place.
> >
> > I read through the Ciscopress QoS 2nd edition book and it seems that
> > shaping
> > actually occurs before fragmentation. Say there is now congestion on
> the
> > software queues, and a packet arrives (which is larger than
320bytes)
> and
> > exceeds the configured Bc, so it gets shaped inside the shaping
queue.
> > According to the book, it seems that only when the shaping scheduler
> > decides
> > to de-queue the packet only then it is checked if the packet
actually
> is
> > larger than 320bytes, and if so it is fragmented. So it seems to me
> that a
> > single packet is actually taking more than 1 interval to be
> transmitted
> > when
> > the packet is in the shaping queue since the actual fragmentation of
> the
> > packet only occurs after it is being shaped.
> >
> > Appreciate if u could clarify if I'm looking at this in the wrong
> > direction.
> > Thanks a lot.
> >
> > Cheers,
> > K
> >
> > ----- Original Message -----
> > From: "Scott Smith" <hioctane@gmail.com>
> > To: "Moin, Imran" <imoin@virtela.net>
> > Cc: "Group Study (E-mail)" <ccielab@groupstudy.com>
> > Sent: Sunday, March 12, 2006 1:47 PM
> > Subject: Re: IE Lab10 - FRTS
> >
> >
> > > Tc = 10ms (smallest possible)
> > > Bc = CIR * Tc
> > > Bc = 256000 * .01 (10/1000 =.01)
> > > Bc = 2560
> > > fragment = 2560 / 8
> > > fragment = 320 (fragment is in bytes so 2560 bits in bytes is 320)
> > >
> > > Changing the fragment to 320 ensures "that a single packet cannot
> take
> > > more than one interval to be transmited".
> > >
> > > -Scott
> > >
> > > On 3/11/06, Moin, Imran <imoin@virtela.net> wrote:
> > >> Hello Folks,
> > >>
> > >> Can someone please help me understand how the values of Bc and
> > >> frame-relay
> > >> fragment were derived in Question number 8.1 (FRTS).
> > >>
> > >> Thanks,
> > >> Imran.
> > >>
> > >>
>
This archive was generated by hypermail 2.1.4 : Mon May 01 2006 - 11:41:59 GMT-3