Ready for mixed practice. Move into case-study questions and explain your trade-offs out loud.

Keep drilling this domain. Revisit the objectives and decision patterns, then retake this quiz.

Google Professional Cloud Architect Domain 2 Guide

What This Domain Tests

Expect implementation choices. The strongest answer usually provisions the right managed service with correct networking, security, data lifecycle, scalability, and operational controls.

Exam Weight

Google lists this domain at ~17.5% of the standard Professional Cloud Architect exam.

How to Think

Read the scenario like an architect: identify constraints, rank trade-offs, and choose the answer that best satisfies the stated business and technical goals.

Study move: For this domain, do not only memorize product names. Practice explaining why the wrong answers are attractive but incomplete.

Ready to drill this domain?

Use the tabs above to move from official objectives to decision patterns, scenario practice, and a quick quiz.

Official Objective Map

Use this as your domain study outline.

1Configure network topologies

Design VPCs, firewalls, routing, load balancing, private access, Shared VPC, peering, and hybrid or multicloud connectivity.
Understand when to extend to on-premises and when to keep boundaries simple.
Secure traffic paths with access control, intrusion protection, and network segmentation.

2Configure storage systems

Choose storage based on access pattern, latency, durability, lifecycle, retention, growth, backup, and recovery requirements.
Provision data processing and compute close enough to data to meet transfer and latency needs.
Apply security and access management to storage from the start.

3Configure compute systems

Map requirements to Compute Engine, GKE, Cloud Run, Cloud Run functions, VMware Engine, specialized workloads, and spot or standard resources.
Understand orchestration, patch management, resource configuration, autoscaling, and container lifecycle expectations.
Use serverless where operational simplicity is a requirement.

4Leverage Vertex AI and prebuilt AI solutions

Understand Vertex AI pipelines, data integration, model training and serving, GPUs, TPUs, AI Hypercomputer, and Model Garden.
Know when prebuilt APIs such as Search, Conversation, Vision, Image, Video, and Audio fit faster than custom ML.
Integrate Gemini Enterprise features, AI agents, and NotebookLM where workflow enhancement is the goal.

Decision Patterns

These are the mental shortcuts that help under exam pressure.

Hybrid connectionChoose dedicated or partner connectivity when bandwidth, reliability, and private access justify it; use VPN for simpler or lower-throughput needs.

Load balancingUse global external load balancing for global apps, internal load balancing for private service access, and regional patterns where locality matters.

Storage lifecycleLet retention, access frequency, recovery needs, and compliance drive lifecycle rules and backup design.

Compute volatilityUse spot resources for fault-tolerant, interruptible work; use standard resources for steady critical workloads.

AI build versus buyUse prebuilt APIs for common perception or conversation tasks; use Vertex AI pipelines for custom lifecycle control.

Mini Scenarios

Open each card, answer in your own words, then compare.

Prompt: A batch analytics job can restart safely and needs lower compute cost.

Strong answer: Use interruptible or spot-style capacity where appropriate, design retry behavior, and avoid placing state only on the ephemeral worker.

Prompt: An enterprise needs private connectivity between on-premises systems and Google Cloud with predictable performance.

Strong answer: Evaluate Cloud Interconnect options, routing, firewall policy, private access, and operational monitoring rather than using public internet paths by default.

Prompt: A team wants to add document summarization to an internal workflow quickly.

Strong answer: Prefer managed Gemini or prebuilt AI capabilities if they satisfy security and data handling requirements; avoid custom model training unless there is a clear need.

Readiness Checklist

Track what you can confidently explain without notes.

0 of 6 complete

Can choose VPC, peering, Shared VPC, load balancing, and private connectivity patterns

Can explain storage lifecycle, retention, backup, recovery, and growth planning

Can map compute choices to Compute Engine, GKE, Cloud Run, functions, and VMware Engine

Can identify when spot resources are safe

Can distinguish prebuilt AI APIs from custom Vertex AI workflows

Can include security and IAM in provisioning decisions

Five-Question Quiz

Use this as a quick readiness pulse, not a score predictor.

Common Traps

These are the answer patterns to catch before exam day.

Too many unrelated VPCs can create routing, security, and operations problems.

Retention, growth, and recovery are exam signals; include them.

GKE is powerful, but Cloud Run or managed services may better fit simpler workloads.

Compute and processing decisions must account for where data lives and how it moves.

Managed AI services can be the best architecture when the requirement is capability, not custom model ownership.

FAQ and Sources

Quick answers plus official references to verify details before exam registration.

Very. Build small labs for VPCs, IAM, Cloud Run, GKE basics, Cloud Storage lifecycle, and load balancing.

Yes. The current guide includes Vertex AI workflows, AI Hypercomputer, Model Garden, Gemini, and prebuilt AI solutions.

Know the selection rules: private versus public, hybrid versus cloud-only, global versus regional, centralized versus distributed.

Choosing infrastructure without explaining security, lifecycle, and operations implications.

Create a table of access pattern, latency, durability, lifecycle, backup, and compliance needs for each major storage option.

Google Cloud Professional Cloud Architect certification page Official Professional Cloud Architect exam guide Professional Cloud Architect learning path