Access and uplinks ports QoS Queuing model On an access switch, does the uplinks egress QoS Queuing model (multiple traffic classes and bandwidth reservation), should be different than the access ports egress Queuing model ? Is there a sense of defining different egress policy-map, with different bandwidth reservation, for access ports and uplinks ports ?

The model is roughly the same, but the marking is different (swap source and destination), and you may allocate queues differently due to higher bandwidth.