Traffic-Shaping Concepts
Traffic shaping prevents the bit rate of the packets exiting an interface from exceeding a configured
shaping rate. To do so, the shaper monitors the bit rate at which data is being sent. If the configured
rate is exceeded, the shaper delays packets, holding the packets in a shaping queue. The shaper
then releases packets from the queue such that, over time, the overall bit rate does not exceed the
shaping rate.
Traffic shaping solves two general types of problems that can occur in multi-access networks.
First, if a service provider purposefully discards any traffic on a VC when the traffic rate exceeds
the committed information rate (CIR), then it makes sense for the router to not send traffic faster
than the CIR.
Egress blocking is the second type of problem for which shaping provides some relief. Egress
blocking occurs when a router sends data into a Frame Relay or ATM service, and the egress Frame
Relay or ATM switch has to queue the data before it can be sent out to the router on the other end
of the VC. For example, when a T1-connected router sends data, it must be sent at T1 speed. If the
router on the other end of the VC has a link clocked at 256 kbps, the frames/cells will start to back
up in the output queue of the egress switch. Likewise, if that same T1 site has VCs to 20 remote
sites, and each remote site uses a 256-kbps link, then when all 20 remote sites send at about the
same time, frames/cells will be queued, waiting to exit the WAN egress switch to the T1 router. In
this case, shaping can be used to essentially prevent egress queuing, moving the packets back into
a queue in the router, where they can then be manipulated with fancy queuing tools.