GroupStudy.com GroupStudy.com - A virtual community of network engineers
 Home  BookStore  StudyNotes  Links  Archives  StudyRooms  HelpWanted  Discounts  Login
RE: wrr-queue bandwidth command posted 01/12/2006
[Chronological Index] [Thread Index] [Top] [Date Prev][Date Next] [Thread Prev][Thread Next]


 "meth-challenged folks" ?

Seems like meth (as in crystal) would only make QOS that much harder to
understand :)   Or did you mean "mAth"?


Chuck 


-----Original Message-----
From: nobody@xxxxxxxxxxxxxx [mailto:nobody@xxxxxxxxxxxxxx] On Behalf Of
HIERS, DAVID (AIT)
Sent: Thursday, January 12, 2006 1:02 PM
To: ccielab@xxxxxxxxxxxxxx
Subject: RE: wrr-queue bandwidth command

The math gets pretty simple if you pick weights that add up to 100.  

Try solving for 10 20 30 40 and 25 25 25 25 for 4 queues, or 20 30 50
and 10 30 60 for 3 queues.

Lazy trick for us meth-challenged folks...

David Hiers

CCIE 10734, CISSP
Systems Engineer
AT&T AESE

-###- 

-----Original Message-----
From: nobody@xxxxxxxxxxxxxx [mailto:nobody@xxxxxxxxxxxxxx] On Behalf Of
de Witt, Duane
Sent: Wednesday, January 11, 2006 11:06 AM
To: Chris Lewis
Cc: Nick; ccielab@xxxxxxxxxxxxxx
Subject: RE: wrr-queue bandwidth command


Yup, logic is perfect. I think I'm running low on caffeine :-)



________________________________

From: Chris Lewis [mailto:chrlewiscsco@xxxxxxxxx]
Sent: 11 January 2006 09:01 PM
To: de Witt, Duane
Cc: Nick; ccielab@xxxxxxxxxxxxxx
Subject: Re: wrr-queue bandwidth command



Oopos a little math error there, 4+6+5 is indeed 15 not 14 as I wrote.
The logic is correct though, any weight applied to queue 4 when teh
priority-queue out is also configured is disregarded.



Chris



On 1/11/06, de Witt, Duane <duane.dewitt@xxxxxxxxxxx> wrote:

Hi

I'm a little confused with this one. If the priority queue is excluded
then the weights are 4+6+5 = 15. Q1 should be 25% of 15 which is 3.75
rounded to 4. Q2 should be 30% of 15 which is 4.5 rounded to 5. How does
the solution get to 6?

-----Original Message-----
From: nobody@xxxxxxxxxxxxxx [mailto: nobody@xxxxxxxxxxxxxx] On Behalf Of
Chris Lewis
Sent: 11 January 2006 04:47 PM
To: Nick
Cc: ccielab@xxxxxxxxxxxxxx
Subject: Re: wrr-queue bandwidth command

Nick,

A good post, but I think one point of detail may be lost here. I think
the
point of detail is missed both in the original question and in the
answer
provided. So I think the person that wrote the requirement was missing a

point of how priority-queue out works.

The priority-queue out command configures the 3550 to use strict
priotity
scheduling for transmit queue 4. This means that any weighting assigned
to
queue 4 by the wrr-queue bandwidth is meaningless, as if there is any
traffic in queue 4, it will be transmitted regardless of weighting
configured for that queue. In essence, when priority-queue out is
configured, the bandwidth allocation algorithm ignores weightings
assigned
to queue 4. The documentation for the priority-queue out command states
the
following:

"When you configure the *priority-queue out* command, the weighted round
robin (WRR) weight ratios are affected because there is one fewer queue
participating in WRR. This means that *weight4* in the *wrr-queue
bandwidth*interface configuration command is ignored (not used in the
ratio
calculation). The expedite queue is a priority queue, and it is serviced

until empty before the other queues are serviced. "

This is how it looks on a switch. I configure F0/1 as given.

interface FastEthernet0/1
switchport mode dynamic desirable
wrr-queue bandwidth 4 6 5 5
wrr-queue cos-map 1 6 7
wrr-queue cos-map 2 0 1 2
wrr-queue cos-map 3 4 5
wrr-queue cos-map 4 3
priority-queue out
end

Now I look at the following show command.

Switch#sho mls qos int f0/1 queueing
FastEthernet0/1
Egress expedite queue: ena
wrr bandwidth weights:
qid-weights
1 - 4
2 - 6
3 - 5
4 - 5    when expedite queue is disabled
Cos-queue map:
cos-qid
0 - 2
1 - 2
2 - 2
3 - 4
4 - 3
5 - 3
6 - 1
7 - 1

The important part to note is that queue 4 will have weighting 5 only if
the
expedite queue is disabled.
What the configuration given means is that traffic in queue 4 will take
whatever it wants, then the remaining queues will have bandwidth
allocated
as follows, as any weighting assigned to queue 4 when the priority-queue
out
command is configured, is ignored:

Q1: 4/14 - (the figure 14 coming from the sum of weights fro queue 1,
2, 3)
Q2: 6/14
Q3: 5/15

Now if I take off the priority-queue command, all your math looks
correct to
me, as verified by the output below:

Switch(config)#int f0/1
Switch(config-if)#no priority-queue out
Switch(config-if)#do sho mls qos int f0/1 que
FastEthernet0/1
Egress expedite queue: dis
wrr bandwidth weights:
qid-weights
1 - 4
2 - 6
3 - 5
4 - 5
Cos-queue map:
cos-qid
0 - 2
1 - 2
2 - 2
3 - 4
4 - 3
5 - 3
6 - 1
7 - 1

Chris



On 1/11/06, Nick <seajay76@xxxxxxxx> wrote:
>
> Please disregard the last strange mail.
>
> (1) With the priority-queue out, the 4th queue of the interface
becomes
> the expedite queue.
>
> int fa0/4
> priority-queue out
>
> (2) you need to put the COS-3 packet int eh 4th queue
>
> int fa0/4
> wrr-queue cos-map 4 3 (they are the queue number and the CoS in order)
>
> (3) Then you can use the other remaining 3 queues.
>
> (4) Allocate the designated bandwidth to the queues.
>
> int fa0/4
> wrr-queue bandwidth 4 6 5 5
>
> Here 4 means 4/(4+6+5+5) , 6 means 6/(4+6+5+5) , 5 means 5/(4+6+5+5)
>
> So you might put it this way.
>
> int fa0/4
> wrr-queue bandwidth 20 30 25 25
>
> (5) Since you allocate the bandwidth to the queues,
>    just put the CoSed packets in each queue
>    You did it for the fourth queue already in (2)
>
> int fa0/4
> wrr-queue cos-map 1 6 7
> wrr-queue cos-map 2 0 1 2
> wrr-queue cos-map 3 4 5
>
> HTH~~
>
> Regards,
> Nick
>
> ----- Original Message -----
> From: <Bill.McKenzie@xxxxxxxxx>
> To: <ccielab@xxxxxxxxxxxxxx>
> Sent: Wednesday, January 11, 2006 10:17 PM
> Subject: wrr-queue bandwidth command
>
>
> > Here is the requirement :
> > Cos-3 must be sent of everything else and match the following cos
values
> to
> > according bandwidth.
> >
> > Cos 3 = 25%
> > CoS 6,7 = 20%
> > CoS 0,1,2 = 30%
> > CoS 4,5 = 25%
> >
> > Here is the solution:
> >
> > wrr-queue bandwidth 4 6 5 5
> > wrr-queue cos-map 1 6 7
> > wrr-queue cos-map 2 0 1 2
> > wrr-queue cos-map 3 4 5
> > wrr-queue cos-map 4 3
> > priority-queue out
> >
> >
> > I don't understand the values in the wrr-queue bandwidth command and

> trying
> > to look at the doc cd is just confusing me more. Can anyone point me
to
> a
> > better explanation of how the values were estimated?
> >
> > Thanks,
> > Bill
> >
> >
> >
> > ============================================
> > STATEMENT OF CONFIDENTIALITY
> >
> > The information contained in this electronic message and any
attachments
> to
> > this message are intended for the exclusive use of the addressee(s)
and
> may
> > contain confidential or privileged information. No representation is
> made
> > on its accuracy or completeness of the information contained in this
> > electronic message.  Certain assumptions may have been made in the
> > preparation of this material as of this date, and are subject to
change
> > without notice.  If you are not the intended recipient, you are
hereby
> > notified that any dissemination, distribution or copying of this
e-mail
> and
> > any attachment(s) is strictly prohibited.
> >
> > Please reply to the sender and destroy all copies of this message
and
> any
> > attachments from your system.
> >
> >
_______________________________________________________________________
> > 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

_______________________________________________________________________
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