Which statement is true concerning SNATs using SNAT pools and SNATs using automap?

BIG-IP automap SNATs preferentially use nonfloating (local) self-IP addresses for source address translation rather than floating self-IPs. This is by design so that return traffic is directed back to the same device that translated the outbound traffic. Floating self-IPs are used as a fallback only if no nonfloating self-IP exists on the egress VLAN.

SNAT pools do not translate sources randomly to any address in the pool - BIG-IP uses a deterministic selection method (such as least-connections) rather than purely random selection, and the choice text is incomplete.

Automap does not preferentially use floating self-IPs; it actively prefers nonfloating self-IPs to ensure return traffic symmetry on the originating device.

SNAT pool members must be routable IP addresses dedicated for translation, not virtual server addresses, which serve a different purpose as traffic destinations.