1.  I suppose you could police both directions if you felt like it, but
since it referred to "any port" within the VLAN, I'd say an incoming
measurement would be the most logically accurate.

2.  For outbound-only purposes, shaping could be an option as well.
Typically when the phrase involves words about "not exceeding", a policer is
typically the desired response.  You can't shape inbound though, and so we
wouldn't be targeting the incoming traffic from hosts in that vlan though.

3.  According to the doc CD, the standard policer follows the same format as
the rate-limit command.  normal burst is your target rate converted to bits
then * 1.5.  It's just the thinking of policers and their statistical
measurements.

4.  police cir works "a little" more like a shaper in terms of more specific
measurements.  The rate-limit/regular policer only measures once per second,
and averages over a 5 second period.  Personally, i never liked the use of
"bc" or "be" here because it's NOT as accurate as the way a shaper measures.
but that's just my little rant.  :)

5.  police # is the original/older version.  Police cir is the newer method
which can also involve a pir (dual bucket policing)

6.  I'd have to lab that up, but I don't believe those two commands are
functionally equivalent.

HTH,


Scott Morris, CCIE4 (R&S/ISP-Dial/Security/Service Provider) #4713, JNCIE-M
#153, JNCIS-ER, CISSP, et al.
CCSI/JNCI-M/JNCI-ER
VP - Technical Training - IPexpert, Inc.
IPexpert Sr. Technical Instructor

A Cisco Learning Partner - We Accept Learning Credits!

[EMAIL PROTECTED]

Telephone: +1.810.326.1444
Fax: +1.810.454.0130
http://www.ipexpert.com

 


 



-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Michael Berger
Sent: Saturday, November 10, 2007 10:00 AM
To: [email protected]
Subject: [OSL | CCIE_RS] CCIE R&S: QoS conversion (Workbook Section 17)

Hi There,

   I would have a question regarding exercise 17.3 (well 6 questions
actually :) )

We are asked to  "Configure the 3550 such as any port of vlan 11 does not
exceed 1 Mbps of transmission speed".

There is only fa0/1 in vlan11, so the config will happen there and the
solution applies the following policing

policy-map aPol
   class class-default
      police 1000000 187500 exceed drop

on

int fa0/1
   service-policy input aPol


Q1: why is this applied on the *input* when we are asked to enforce the
*transmission* speed? Why not applying this in "output"?


Q2: we decide to use policing, ok but would shaping have been an option?

Would have this policy made the trick:

policy-map test
  class class-default
   shape average 1000000
!

and then apply it on the "output" of fa0/1?



Q3: Where does this 187500 bytes burst value come from? It is equivalent 
to 1.5Mbps but why is it configured there?


Q4: What is the difference between "burst bytes" and bc in the following 
output:

R2(config-pmap-c)#police cir 1000000 ?
   <1000-512000000>  Burst bytes
   bc                Conform burst
   conform-action    action when rate is less than conform burst
   pir               Peak Information Rate
   <cr>

Q5: What is the difference between policing the CIR or directly a bps 
value in the following output:

R2(config-pmap-c)#police ?
   <8000-2000000000>  Bits per second
   cir                Committed information rate
   rate               Specify police rate


Q6:  Would this command

R2(config-pmap-c)#$police cir 1000000 bc 187500 conform-action tr 
exceed-action drop

be equivalent to:

R2(config-pmap-c)#$police  1000000 187500 conform-action tr 
exceed-action drop

?


Thanks in advance!


Michael

Reply via email to