Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Professional Cloud Security Engineer practice bank has 1,343 more like it. Try 10 free: instant score, no account.
An enterprise runs containerized microservices on Google Kubernetes Engine (GKE) in Google Cloud. The security team needs to enforce egress boundary protection that moves beyond basic Layer 4 port and protocol rules.
The security architecture must meet the following requirements:
Which architecture should the security engineer implement to meet these requirements?
This architecture combines Cloud Service Mesh application-layer controls with Google Cloud VPC perimeter security to establish a defense-in-depth egress boundary. Workload identities are established using mutual TLS (mTLS) and X.509 certificates, while outbound traffic is routed through managed egress gateways deployed on a dedicated GKE node pool. The gateway nodes are isolated using Kubernetes taints and tolerations to prevent standard application workloads from co-locating on them.
istio-egress), while application teams deploy workloads in separate application namespaces.This solution ensures workloads cannot bypass mesh proxies via raw sockets or altered iptables rules because the underlying VPC firewall drops all direct external egress from standard workload nodes.
Kubernetes Network Policies provide pod-level network segmentation within a Kubernetes cluster. They specify how groups of pods are allowed to communicate with each other and with other network endpoints. Network policies are scoped at the namespace level and are implemented by the underlying container network interface (CNI) plugin using packet-filtering frameworks like iptables or eBPF.
Kubernetes Network Policies operate exclusively at Layer 3 (IP addresses/CIDR blocks) and Layer 4 (TCP/UDP ports). They do not possess Layer 7 protocol parsing capabilities and cannot inspect HTTP headers, URI paths, user agents, or payload contents. Furthermore, native Kubernetes Network Policies do not support filtering egress traffic based on fully qualified domain names (FQDNs) or application-layer service mesh identities.
The requirement was to enforce Layer 4 transport segmentation between Kubernetes namespaces and restrict pod communication to specific internal IP ranges and TCP ports.
When sidecar proxies (such as Envoy) are injected into workload pods, container startup scripts program local iptables rules within the pod network namespace to intercept incoming and outgoing TCP traffic. The sidecar proxy then handles traffic routing, telemetry, and Layer 7 policy evaluation.
Relying solely on in-pod sidecar proxies is insufficient as a secure perimeter boundary. Container workloads with elevated privileges, modified network capabilities (NET_ADMIN or NET_RAW), or customized networking binaries can bypass local iptables redirection rules and send uninspected packets directly out of the node interface to external hosts. Without complementary node-level VPC firewall rules blocking external egress from workload nodes, workloads can easily evade Layer 7 inspection.
iptables configuration is not a tamper-proof security perimeter against malicious or compromised container workloads.The application environment operated entirely in a trusted, non-adversarial internal development context where bypass protection and perimeter security were not required.
VPC firewall rules operate as distributed stateful firewalls at the network and transport layers (Layer 3 and Layer 4). They allow or deny traffic based on IP address ranges (CIDR blocks), transport protocols (such as TCP, UDP, or ICMP), and destination port numbers. Filtering can be targeted to specific Compute Engine virtual machines or GKE nodes using network tags or service accounts.
Standard VPC firewall rules cannot perform deep packet inspection or evaluate Layer 7 application payloads. They cannot inspect HTTP request headers, HTTP verbs (such as GET or POST), URLs, or user-agent strings. Attempting to use basic VPC firewall rules alone fails to enforce application-layer security policies and cannot authenticate workloads based on cryptographic identities such as X.509 certificates.
The requirement only demanded simple Layer 4 port and CIDR-based packet filtering without Layer 7 inspection or workload identity verification.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated