Assume that you have four data centers around the globe, each hosting multiple applications for...

App Connectors should be deployed in redundant pairs for each environment that hosts private applications. A single connector creates an availability risk because ZPA depends on connectors to provide outbound-only reachability to application servers. For four data centers, minimum resilient design means two connectors per data center. That is why Option B (Eight -two per data center) is correct: four data centers multiplied by two connectors per data center equals eight App Connectors.

Why the other options are incorrect:

A. Six - one per data center plus two for cold standby: Cold standby connectors sit idle until needed. ZPA resiliency normally depends on active redundant connectors per hosting location, not a few passive spares.

C. Four - one per data center: One connector per data center leaves that site with a single connector failure point. A maintenance event or connector outage would interrupt private-app access for that data center.

D. Sixteen - to support a full mesh to the other data centers: A full mesh is a network-topology mindset. ZPA App Connectors do not need connector-to-connector mesh links; they need redundant outbound reachability to the Zscaler cloud.