QoS: who waits, who gets dropped and why
Every mechanism has a step-by-step animation, the configuration explained line by line and the exam notes. Where a parameter changes the outcome, you get a simulator in which you change it yourself.
TL;DR
QoS doesn't add bandwidth: when a link is congested, it decides which packet leaves first and which gets dropped. You mark traffic only once, at the edge (DSCP: EF 46 for voice, AF for classes with a guarantee, CS for control), you trust only what you control, then every hop applies the same behavior according to the marking. On Cisco everything is written in MQC — class-map, policy-map, service-policy: policing cuts or remarks on input, shaping smooths on output on sub-rate links, LLQ serves voice first with a safety policer, CBWFQ guarantees bandwidth to classes, WRED drops early from TCP to avoid global synchronization. In MPLS the marking becomes a three-bit EXP with three tunneling modes; in the data center the goal is no loss, with PFC and ECN. Most incidents come from marking lost along the way, classes that catch nothing, and congestion in a place other than where the policy is.
How to read the notes: red = critical for the exam, green = theory, blue = numbers, yellow = pitfalls, cyan = differences between platforms and standards. In the animations, the configuration lines used in the current step are highlighted automatically; the ← → arrows change the step when the animation is selected.
The analogy, in short
Think of a hospital with triage: each patient gets a colored wristband at the entrance (the marking), put on by the triage nurse, not by the patient (the trust boundary). The red wristband goes straight into resuscitation (LLQ), the wards have guaranteed beds (bandwidth), the guard at the gate has a limited number of tickets per hour (policing), the small clinic schedules arrivals (shaping), and when the waiting room fills up, a few people with green wristbands are sent home before the door closes for everyone (WRED).
The model holds up to §8. From §9 on (walled-in wards, the wireless courtyard, another hospital with another color code, the ward that calls back), it gets stretched; see the full analogy and its limits in §15.