To Police or to Shape

[d-fillmore]

Hi - We have an  ISP network and sell increments of a 10Gb WAN interface to customers.

At the  moment we are policing inbound at the edges of our network to rate  limuit L2VPN bandwidth and policing inbound and outbound at the egress  of out network for IP transit. We had a few problems with our initial  policing config in that it was quite brutal and the effect for L2VPN  users was that anything over the 50Mb was just getting dropped and their  TCP applications were throttling back loads and the perceived bandwidth  (by the customer) was much lower than the 50Mb. The policer we used was  this.

policy-map  50Meg
  class class-default
   police 50000000 1500 1500     conform-action transmit     exceed-action drop
!

Having  looked at a few other ISP's configs, I can see shapers being used  extensively instead of policers. I know the technical differences  between the two, but want to know if we should look to implement shaping  instead of policing to provide a better experience for the customer  whilst still restricting bandwidth?

Also, I can understand the variables  used in policers and shapers, but are there some standard fugures that  you would use for say 20, 40 and 50 Mb policers as I can imagine just  guessing them is not going to work

Many Thanks in advance

Dom

[Giuseppe Larosa]

Hello Dom,

shaping has to be preferred if supported because it allows to control traffic volume without dropping packets but holding exceeding packets in buffer.

Unfortunately, shaping is supported only outbound.

And it is not supported in all platforms for example you cannot configure it on ethernet linecards of C6500/C7600.

So it is not always a viable option

Hope to help

Giuseppe

[Ganesh Hariharan]

Hi - We have an  ISP network and sell increments of a 10Gb WAN interface to customers.
At
the  moment we are policing inbound at the edges of our network to
rate  limuit L2VPN bandwidth and policing inbound and outbound at the
egress  of out network for IP transit. We had a few problems with our
initial  policing config in that it was quite brutal and the effect for
L2VPN  users was that anything over the 50Mb was just getting dropped
and their  TCP applications were throttling back loads and the
perceived bandwidth  (by the customer) was much lower than the 50Mb.
The policer we used was  this.
policy-map  50Meg
  class class-default
   police 50000000 1500 1500     conform-action transmit     exceed-action drop 
!
Having 
looked at a few other ISP's configs, I can see shapers being used 
extensively instead of policers. I know the technical differences 
between the two, but want to know if we should look to implement
shaping  instead of policing to provide a better experience for the
customer  whilst still restricting bandwidth?
Also,
I can understand the variables  used in policers and shapers, but are
there some standard fugures that  you would use for say 20, 40 and 50
Mb policers as I can imagine just  guessing them is not going to work
Many Thanks in advance
Dom

Hi Dom,

Check out the below link for brief differenc between policy and shaping with comparism.

http://www.cisco.com/en/US/tech/tk543/tk545/technologies_tech_note09186a00800a3a25.shtml

Hope to Help !!

Ganesh.H

Remember to rate the helpful post

[d-fillmore]

Thanks for your replies guys, I'm after some example figures for things like Bc on 20, 40 and 50 Mb policers and how we come up with thos figures

I'm going to have to use policers instead as we need inbound and outbound control.

Cheers, Dom