The purpose of traffic shaping
Traffic shaping, or traffic management, controls the bandwidth available and sets the priority of traffic processed by the policy to control the volume of traffic for a specific period (bandwidth throttling) or rate the traffic is sent (rate limiting).
Traffic shaping attempts to normalize traffic peaks and bursts to prioritize certain flows over others. But there is a physical limitation to the amount of data which can be buffered and to the length of time. Once these thresholds have been surpassed, frames and packets will be dropped, and sessions will be affected in other ways.
A basic traffic shaping approach is to prioritize certain traffic flows over other traffic whose potential loss is less disadvantageous. This would mean that you accept certain sacrifices in performance and stability on low-priority traffic, to increase or guarantee performance and stability to high-priority traffic.
If, for example, you are applying bandwidth limitations to certain flows, you must accept the fact that these sessions can be limited and therefore negatively impacted.
Note that traffic shaping is effective for normal IP traffic at normal traffic rates. Traffic shaping is not effective during periods when traffic exceeds the capacity of the FortiGate unit. Because packets must be received by the FortiGate unit before they are subject to traffic shaping, if the FortiGate unit cannot process all of the traffic it receives, then dropped packets, delays, and latency are likely to occur.
To ensure that traffic shaping is working at its best, make sure that the interface Ethernet statistics show no errors, collisions or buffer overruns.
Accelerated interfaces (NPx network processors and CE) affect traffic shaping. For more information, see the FortiOS Whats New in Hardware Acceleration for FortiOS 5.4 guide.