I85PREP
⌘ K
PCCP study materials
PCCP study resources

PCCP - Professional Cloud Certified Professional: Designing and planning a cloud solution architecture

PCCP · Work through your exam, one topic at a time.

All guide sections

Designing and planning a cloud solution architecture

Designing and planning a cloud solution architecture

This domain tests your ability to translate business and technical requirements into a GCP architecture: resource hierarchy, network topology, compute and storage choices, and explicit trade-offs among cost, latency, availability, and compliance. Expect scenario questions that give constraints (SLO, RTO/RPO, budget, data residency) and ask for the best design.

Core rules and concepts

  • Start with requirements: define availability target (e.g., 99.9% vs 99.99%), RTO/RPO, latency, compliance, and budget before choosing products.
  • Resource hierarchy: Organization > Folders > Projects > resources. IAM and Org Policy inherit downward; use folders for environment/team separation; projects are billing and quota boundaries.
  • Network: a VPC is global and subnets are regional. Use Shared VPC when a central team manages networking across projects; VPC peering connects networks; Cloud Interconnect or VPN for hybrid; Private Google Access for internal IPs to reach Google APIs.
  • Compute: GKE for container orchestration and portability; Cloud Run for serverless containers; Cloud Functions for event-driven; Compute Engine for full control; managed instance groups for autoscaling and healing.
  • Storage and data: Cloud Storage classes (Standard, Nearline, Coldline, Archive); BigQuery for analytics; Cloud SQL for regional OLTP; Spanner for global strong consistency; Firestore for mobile/web.
  • Design for failure: multi-zone is the default for regional resources; multi-region only when the SLO justifies cost.

Worked example

A retailer needs a global web app at 99.95% availability with a relational backend. Reason: use a global external HTTP(S) load balancer with anycast IP and Cloud CDN; run regional GKE clusters in two regions behind it; use Cloud SQL with a cross-region replica (or Spanner if global writes are needed); health checks and autoscaling. This meets the SLO with two regions, avoiding the cost of many regions.

Traps

Over-engineering multi-region when multi-zone meets the SLO; assuming a VPC is regional; using Cloud SQL for global write scale; ignoring Org Policy and quotas; forgetting that Shared VPC requires host/service project design.