A financial firm runs voice trading turrets, interactive video, transactional trading apps and bulk file replication over a congested WAN. Voice needs minimal delay and loss, video needs assured bandwidth, trading needs better-than-best-effort assurance, and replication can use spare capacity. Which PHB mapping strategy correctly reflects EF, AF, CS and default behavior?

DiffServ PHBs encode per-hop behavior where EF provides strict priority with low delay, jitter and loss for voice, AF provides assured forwarding with four classes and three drop precedences for variable-rate assured traffic such as video and critical data, CS provides class selector backward compatibility often for network control, and default DSCP 0 provides best effort for scavenger or bulk traffic. The voice EF plus video AF4 plus trading AF2 plus replication default mapping is correct because it aligns delay sensitivity and loss tolerance to scheduling behavior, giving voice priority queuing, video assured bandwidth with low drop, trading assured delivery, and replication only spare capacity. Marking voice AF4 and video EF fails by concept because it inverts strict-priority and assured behavior, starving voice of priority and over-provisioning bursty video. Marking trading EF and voice default fails by concept because it grants priority to transactional data while leaving delay-sensitive voice in best effort, violating delay budgets. Marking replication EF fails by concept because bulk traffic would hijack the priority queue and cause voice and video starvation. For the exam, associate EF with voice, AF4 with video, AF2 or AF3 with critical data, and default with bulk. In practice, verify trust, policing, queue sizes and that EF is rate-limited to prevent starvation.