Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Network Engineer practice bank has an additional 2,437 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your organization operates multiple Google Kubernetes Engine (GKE) clusters in separate producer VPC networks in us-central1. Due to internal RFC 1918 IP exhaustion, every cluster reuses identical Pod and Service CIDR blocks. You must design connectivity and IP allocation to meet these requirements:
us-east4 to privately consume a GKE microservice without exposing Pod IP addresses.Which architecture should you implement?
PRIVATE_SERVICE_CONNECT NAT subnet. Create a PSC endpoint with global access enabled in the consumer VPC, advertise its internal RFC 1918 IP over Cloud Interconnect, and expose the web application using an external Application Load Balancer with an external anycast IP and Cloud Armor.This architecture uses Private Service Connect (PSC) to decouple service-producer GKE VPCs from consumer networks through service-oriented NAT translation, combined with an external Application Load Balancer for Layer 7 public ingress. Each GKE cluster operates inside its own isolated VPC network while exposing only specific internal service endpoints.
PRIVATE_SERVICE_CONNECT subnet, clusters can reuse identical RFC 1918 Pod and Service CIDR blocks without IP overlap conflicts.us-east4: Deploying the PSC endpoint (an internal RFC 1918 forwarding rule IP) in the consumer VPC with PSC global access enabled allows clients connected through Cloud Interconnect in us-east4 to reach the endpoint in us-central1 once the consumer subnet is advertised via Cloud Router.PRIVATE_SERVICE_CONNECT NAT ranges for producer translation, and an external anycast IP strictly at the public load balancer frontend.Unlike VPC Network Peering, Cloud VPN island-mode SNAT, or private services access, PSC natively supports overlapping cluster IP spaces, cross-region hybrid access via global access, and minimal operational overhead while pairing seamlessly with an external Application Load Balancer for public traffic.
This approach implements an island-mode network pattern using Cloud VPN and the GKE IP masquerade agent (nonMasqueradeCIDRs) so nodes perform SNAT from overlapping Pod CIDRs to unique node IP addresses before traffic leaves the cluster VPC. For public access, it uses an external passthrough Network Load Balancer forwarding directly to GKE nodes with external IP addresses.
ip-masq-agent ConfigMaps, Cloud VPN gateways, and custom BGP route advertisements across multiple isolated GKE VPCs adds substantial administrative overhead.You were intentionally avoiding PSC due to legacy constraints requiring node-level IP reachability, had unique non-overlapping node CIDRs available across all clusters, and did not require private-only GKE nodes or Cloud Armor Layer 7 protection.
/22 RFC 1918 CIDR range in the consumer VPC for private services access to reach the GKE producer VPC, and assign the GKE node subnet as the NAT subnet for the internal load balancer's service attachment. Expose the web application using an external Application Load Balancer backed by a Private Service Connect network endpoint group (NEG), and deploy Cloud NAT to handle inbound public traffic.This option mixes private services access (which allocates a /22 CIDR block for VPC Network Peering), a misconfigured PSC service attachment NAT subnet, a PSC network endpoint group (NEG) behind an external Application Load Balancer, and Cloud NAT for public ingress.
PRIVATE_SERVICE_CONNECT, separate from the load balancer and node subnets./22 ranges), exacerbates IP exhaustion, and lacks transitivity across hub-and-spoke topologies.No valid architecture combines these mismatched components; instead, replace private services access with a PSC endpoint with global access, use a dedicated PRIVATE_SERVICE_CONNECT subnet for the service attachment, and let the external Application Load Balancer handle inbound public traffic directly without Cloud NAT.
PRIVATE_SERVICE_CONNECT NAT subnet. Create a regional PSC endpoint without global access in a separate services VPC peered via VPC Network Peering to the consumer VPC hosting the Cloud Interconnect attachment, and expose the web application using an external Application Load Balancer with Cloud Armor.This configuration correctly publishes the GKE microservice through an internal passthrough Network Load Balancer and a Private Service Connect (PSC) service attachment with a dedicated NAT subnet, and it appropriately uses an external Application Load Balancer with Google Cloud Armor for public ingress. However, it places a regional PSC endpoint (without global access) in a separate VPC connected via VPC Network Peering.
us-east4 to us-central1): Fails because PSC global access is disabled on the endpoint. Without global access enabled, clients and Cloud Interconnect VLAN attachments residing in us-east4 cannot route traffic to a regional PSC endpoint in us-central1.allowPscGlobalAccess is explicitly enabled.Instead of peering a separate VPC to a regional PSC endpoint, the PSC endpoint with global access enabled were deployed directly in the consumer VPC that hosts the Cloud Interconnect attachment, or an internal load balancer with a PSC network endpoint group (NEG) backend were deployed in the services VPC to bridge the peered network.
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