From: Carlos G Mendioroz (tron@huapi.ba.ar)
Date: Mon Aug 30 2004 - 12:14:17 GMT-3
I believe I do have another "look" at the requirement:
The task to police P2P says "a lot of complaints about slow network
performance".
The solucion as I understand it does:
All offered traffic
>shape to 5M> (5M traffic)
>police P2P> (less than 5M, clean traffic)
>Tx
Shaping section says provider will drop traffic above 5M.
What I think should be done is:
All offered traffic
>police P2P> (Clean traffic)
>shape to 5M> (5M traffic)
>Tx
Which would have impact in the user complaints of slow performance.
(The first option would not, AFAIK)
You say:
> The p2p traffic will be policed to 8000 bps and all other traffic will
> be shaped as specified above.
The thing may be my bad understanding of what the service-policy command
does. My understanding is that this defines the queue policy applied to
service a queue, usually creating a new set of queues.
May be that's no so. I'd appreciate a document that relates to this...
ccie2be wrote:
> Hi Carlos,
>
> I believe you're not reading/interpreting the task correctly.
>
> The requirement is to have ALL traffic shaped to the specified parameters.
> Since ALL traffic is to be shaped that clearly includes the peer to peer
> traffic.
>
> Then in task 8.4, you need to police the p2p traffic portion of the traffic
> being shaped.
>
> policy-map SHAPE
> class class-default
> shape average 5000000 160000 80000
> service-policy QOS_POLICY
>
> In the above snippet, since "service-policy QOS_POLICY" is within policy-map
> SHAPE, what will happen is this:
>
> The p2p traffic will be policed to 8000 bps and all other traffic will be
> shaped as specified above.
>
> There are a number of show commands you can use to verify this including
> show traffic shape [options], show queue [options].
>
> I suggest you play around with these show commands and make sure you
> understand how to interpret the output. Besides the command reference, I
> also recommend the Wendall Odom book, Cisco DQOS. ISBN 1-58720-058-9.
>
> He has a bunch of examples of this type along with output from various show
> commands which should make this more clear.
>
> HTH, Tim
>
>
>
> ----- Original Message -----
> From: "Carlos G Mendioroz" <tron@huapi.ba.ar>
> To: <ccielab@groupstudy.com>
> Sent: Monday, August 30, 2004 9:51 AM
> Subject: Nested service policies, IEWB Lab 9, Task 8.4
>
>
>
>>Task is to control P2P traffic after doing shaping to 5M
>>on an interface.
>>
>>The solution proposed is to nest two service policies, one to do the
>>shaping and the other to do the policing.
>>
>>But I think the policies are reversed.
>>The sol goes:
>>
>>int atm0/0/0
>> service-policy SHAPE
>>
>>policy-map SHAPE
>> class class-default
>> shape average 5000000 160000 80000
>> service-policy QOS_POLICY
>>
>>policy-map QOS_POLICY
>> class P2P
>> police cir 8000
>>
>>It is my understanding that the interface will shape to 5M ALL the
>>traffic and then the P2P traffic will be policed, so the link will be
>>underutilized (i.e. P2P will not be tx, but nothing else will take the
>>place).
>>
>>Am I wrong ? (Would not be the first time :-)
>>
>>--
>>Carlos G Mendioroz <tron@huapi.ba.ar>
>>
>>_______________________________________________________________________
>>Please help support GroupStudy by purchasing your study materials from:
>>http://shop.groupstudy.com
>>
>>Subscription information may be found at:
>>http://www.groupstudy.com/list/CCIELab.html
>
>
>
-- Carlos G Mendioroz <tron@huapi.ba.ar>
This archive was generated by hypermail 2.1.4 : Fri Sep 03 2004 - 07:02:50 GMT-3