Hi,

Traffic shaping differs from traffic policing in that shaping is designed to buffer excess traffic, whereas policing is designed to drop traffic in excess of the target rate, in other words shaping delays (queues) excess traffic to be transmitted at non congested intervals. On the other hand policing simply either drop or mark the excess traffic, and never queues it.

Please do check this document:

http://www.cisco.com/warp/public/105/policevsshape.html

[edit] Guru, sorry for the cross post.

BR,

Mohammed Mahmoud.

Both of dese are used synchronously to achieve wat is termed as proper b/w allocation policy.

Shaping - Used for regulating the traffic flow without dropping any excess b/w demands. It has smoothening effect on the traffic.All excess traffic will be smoothened out by having it stored in buffer queues for later transmission.

Buffer utilisation will increase depending on how much excess traffic is detected.Can be applied only as the output component of an interface as the queues are more used for outward transmission.

Policing - Used for strictly limiting the traffic to a defined limit.No additional buffer allocations are required here as non conforming traffic is basically dropped or sent out with a lowered rate, depending on the configured parameter.NO storage of excess traffic happens.Uses 1-token bucket or 2-token bucket for regulation.

can be applied both inbound & outbound on the interface. Defines excess burst levels as well based on the action configured.

If your need is strict regulation...den go for policing and if its just easing out the traffic flow ..use shaping for better output.

Rate if this helps!!!!!

Shaping behaves much like an interface with fair-queue.

Policing somewhat behaves much like an interface with FIFO, although often like one with a vary small queue. However, with policing, when the "FIFO queue" is exceeded, the packet doesn't have to be dropped, its ToS can be marked differently. Some policers also support three tiers.

Often both the shaper or policer can be configured to allow an initial burst rate, for an interval, larger than sustained rate.

When to use one over the other depends on what you're trying to accomplish, and often what the platform supports. If you need to regulate bandwidth utilization by dropping packets, would you prefer you regulation to use fair-queue or FIFO with a small queue?