Egress traffic shaping vs. normal output interface queue When traffic needs to egress from an interface, it is put to an output queue. When a traffic shaper is shaping egress traffic going out of an interface, it puts the traffic in an output queue also. So can i say 1. The only difference between having a shaper or not is just that the shaper controls how much traffic is allow to egress out of the queue/interface? 2. If an interface outbound queue is full, it doesn't matter if a shaper is in place or not, tail packets will still be drop. - right?

A standard hardware queue is FIFO (First In, First Out), and all the traffic going through the interface passes through this one queue.

Based on your previous question, you probably want to shape based on traffic type, giving priority to things like VoIP. This involves setting up multiple software queues, and giving each queue a certain period of time to send to the hardware queue. A priority queue, like for VoIP, will always get first precedence, up to it defined bandwidth, as long as there is something in the priority queue.

Also, tail drop can cause TCP global synchronization (a bad thing). There are mitigations like RED (Cisco WRED) that randomly drop packets in a queue to prevent the queue from filling and causing tail drop.