Which option is the Cisco recommended mechanism for delivering voice payload?

IntServ (Integrated Services) uses RSVP to reserve bandwidth per flow end-to-end, which is not scalable in large networks and is not Cisco's recommended approach for voice delivery.

DiffServ (Differentiated Services) uses DSCP markings in the IP header to classify and prioritize traffic at each network hop, applying the Expedited Forwarding (EF) per-hop behavior for voice. This approach is scalable across large networks because it does not require per-flow state to be maintained by routers, making it Cisco's recommended QoS model for voice.

Policing drops or re-marks traffic that exceeds a defined rate, which can cause voice packet loss and degrade call quality, making it unsuitable as a primary voice delivery mechanism.

Shaping buffers excess traffic to smooth transmission rates, which introduces delay and jitter into the voice stream, negatively impacting voice quality.