200-901 Application Deployment and Security Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Host network

  Why it's wrong here

  Host networking places the pod in the node's network namespace, so containers share the node's interfaces rather than an isolated per-pod namespace; localhost then reaches node services, not sibling containers reliably. It suits pods needing raw node-level performance or host port binding, not intra-pod sidecar communication.

• ✗

  Bridge network

  Why it's wrong here

  Bridge networking attaches each container to a shared L2 segment, so containers reach each other by IP, not localhost. It suits multi-host or VM-style topologies needing external connectivity. The scenario requires the shared network namespace that only the pause-container model provides, letting both processes bind the same loopback interface.

• ✗

  Overlay network

  Why it's wrong here

  Overlay networking encapsulates pod-to-pod traffic across nodes via VXLAN or similar tunnels, addressing cross-node connectivity rather than intra-pod loopback. It is the right choice when pods on different hosts must communicate. Within a single pod, containers already share one network namespace, so localhost works without any overlay.

• ✓

  Pod network (containers share the same IP)

  Why this is correct

  Containers within a single pod share one network namespace, hence one IP address and port space, so the sidecar proxy and main application reach each other over localhost. This shared pod network model is exactly what the stem's localhost communication requires.

Go deeper

Related to this question