In a Zero Trust architecture, how is the connection to an application provided?

The correct answer is A. Over any network with per-access control. In Zero Trust architecture, access is provided to the specific application , not to the underlying network. This is a foundational design principle in Zscaler’s Universal Zero Trust Network Access (ZTNA) guidance. Users can connect from any location and over any network , while policy is enforced per user, per device, per application, and per session . This differs from legacy approaches that first place the user onto the network and then rely on network segmentation or firewall rules to limit access.

Option B is incorrect because establishing a full network-layer connection is characteristic of legacy VPN-based access, which extends network trust and increases lateral movement risk. Option C is also incorrect because Zero Trust is not defined by building a virtual appliance stack in front of applications. Option D includes TLS, which is used in Zscaler architectures, but the key Zero Trust concept being tested is not merely encrypted transport; it is brokered, granular, per-access connectivity without exposing the application to broad network reachability. Therefore, the most accurate answer is A .