A multinational genomics research institute processes sensitive sequencing data on Google Cloud across multiple projects. Project-Compute hosts a private Google Kubernetes Engine (GKE) cluster that runs analytical workloads, while Project-Data stores raw genomic files in Cloud Storage buckets. The security policy mandates that: 1) On-premises analyst workstations connected via Dedicated Interconnect must securely administer the GKE cluster master, 2) GKE workloads must read objects from Project-Data without routing over the public internet, and 3) Data exfiltration to external Cloud Storage buckets outside the organization must be strictly blocked, even if a compromised service account has IAM Storage Admin permissions. Which architectural design fulfills all of these security requirements?
- Place Project-Compute and Project-Data within a single VPC Service Controls perimeter, enforce GKE control plane authorized networks restricted to the on-premises subnet CIDR over Dedicated Interconnect, and use Private Service Connect endpoints for Cloud Storage access.Cevap
- BConfigure fine-grained IAM conditional roles on Project-Data buckets restricted to the GKE service account identity, and establish VPC Network Peering between Project-Compute and Project-Data.
- CEstablish VPC Network Peering between the on-premises transit VPC and Project-Compute VPC, assuming traffic will transitively route through Project-Compute to reach Cloud Storage in Project-Data.
- DEnable public endpoint access on the private GKE cluster while disabling authorized networks, and assign the primitive Owner role to the GKE nodes service account to bypass VPC Service Controls perimeter blocks.