The correct answer is B . In Zero Trust architecture, risky behavior is controlled through continuous evaluation and policy-based response , not through static network constructs such as VLAN quarantine or dependence on standalone appliances. Zscaler’s Zero Trust guidance emphasizes granular, context-based policies that evaluate the user, device, application, and surrounding conditions before and during access. In the ZPA architecture material, Zscaler states that applications should remain inaccessible unless the user is authorized, and policy should be independent of IP address or location.

The strongest architecture match is option B , because Zscaler documentation describes security outcomes such as inline prevention, deception, and threat isolation for compromised or risky users. That means when behavior becomes suspicious, later access attempts can be restricted, misdirected, or blocked based on updated policy context. This is fundamentally different from a legacy response such as placing a device permanently in a VLAN, which remains network-centric and coarse-grained. Logging alone also does not control risk, and simply deploying security appliances does not deliver Zero Trust by itself. Zero Trust controls risky behavior by dynamically adjusting enforcement based on observed context and threat posture, which best aligns with option B.