Examkin

Question

Difficulty: MediumDesigning Infrastructure for Technical Requirements and High Availability

A healthcare enterprise is designing its cloud architecture on Google Cloud for a patient medical imaging retrieval API and metadata repository. On-premises hospitals must securely stream imaging data to Google Cloud over a dedicated hybrid network connection requiring a sustained throughput of 15 Gbps with 99.99% availability. The imaging retrieval service consists of stateless, containerized HTTP microservices with variable traffic patterns that require low operational overhead. The relational metadata database operates within a single GCP region and requires high availability with automatic zonal failover, but does not require global multi-region write replication. Which TWO architectural decisions should you implement to satisfy these high availability and technical requirements?

  1. Provision redundant Dedicated Interconnect connections with at least two 10 Gbps circuits terminating in separate edge availability domains.Answer
  2. Deploy the stateless HTTP microservices on Cloud Run and utilize Cloud SQL in a High Availability (HA) dual-zone configuration for the metadata database.Answer
  3. C
    Establish an HA Cloud VPN gateway with four active IPsec tunnel pairs to aggregate network bandwidth for the 15 Gbps sustained requirement.
  4. D
    Deploy Cloud Spanner as the relational metadata database engine to ensure maximum regional availability.
  5. E
    Configure Application Load Balancer health checks to execute a SQL SELECT query directly against the backend database engine.

Answer

The optimal architecture requires provisioning redundant Dedicated Interconnect circuits (at least 10 Gbps each across separate edge availability domains) to meet the 15 Gbps throughput and 99.99% SLA requirements, along with deploying stateless containers on Cloud Run backed by a regional Cloud SQL HA instance.
To satisfy 15 Gbps sustained bandwidth and a 99.99% uptime SLA, Dedicated Interconnect with redundant 10 Gbps circuits across distinct edge availability domains is required. For the application tier, Cloud Run provides a serverless platform for stateless HTTP services with minimal operational overhead, and Cloud SQL in HA mode provides automatic cross-zone failover for single-region relational data without the cost and complexity of Cloud Spanner.

Step-by-Step Solution

1
Evaluate hybrid network throughput and SLA requirements
Sustained throughput of 15 Gbps exceeds the standard per-tunnel limit of Cloud VPN (3 Gbps per tunnel). Achieving 99.99% availability requires Dedicated Interconnect configured across two edge availability domains in a single region or metro.
Dedicated Interconnect provides dedicated physical 10 Gbps or 100 Gbps pipes to GCP, satisfying high bandwidth and stringent uptime SLAs.
2
Evaluate compute tier requirements for stateless HTTP microservices
Cloud Run provides fully managed serverless container execution for stateless HTTP workloads, automatically scaling based on incoming web traffic without server or cluster administration.
Cloud Run minimizes operational overhead compared to managing GKE cluster control planes and node pools for simple stateless services.
3
Evaluate database requirements for single-region HA relational storage
Cloud SQL with High Availability (HA) configuration replicates data synchronously to a standby instance in a secondary zone within the same region and supports automatic failover.
Cloud Spanner is over-engineered and costly for workloads that do not require horizontal write scaling across multiple regions.

Key Concept

Designing High-Availability Hybrid Networks and Compute/Storage Tier Selection
Rate this question

Topics

Designing Infrastructure for Technical Requirements and High Availability