GCP#
Google Cloud Platform is the third hyperscaler. Strongest in Kubernetes (the project came out of Google), data analytics (BigQuery), and ML (Vertex AI). The catalog is smaller than AWS or Azure, the integration tighter, the network among the best in the industry.
Account model#
The GCP hierarchy:
Organization, the top-level resource tied to a Google Workspace / Cloud Identity domain.
Folder, optional grouping (departments, environments).
Project, the unit of billing, IAM, and quota. Equivalent to an AWS Account; recommended one per environment per service.
Resource, the actual asset.
Identity lives in Cloud Identity (or Google Workspace). Service accounts are first-class resources per project; roles bind identities to projects, folders, or resources.
Catalog at a glance#
Layer |
Headline services |
|---|---|
Compute |
Compute Engine (VMs), GKE (Kubernetes), Cloud Run (serverless containers), Cloud Functions (FaaS), Batch. |
Storage |
Cloud Storage (object), Persistent Disk (block), Filestore (file), Coldline / Archive tiers. |
Database |
Cloud SQL, AlloyDB, Spanner, Firestore, Bigtable, Memorystore. |
Analytics |
BigQuery (data warehouse), Dataflow, Dataproc, Pub/Sub. |
Networking |
VPC, Cloud Load Balancing, Cloud CDN, Cloud DNS, Cloud Interconnect, Network Connectivity Center. |
Identity |
IAM, Cloud Identity, Secret Manager, KMS, Workload Identity Federation. |
Observability |
Cloud Monitoring, Cloud Logging, Cloud Trace, Cloud Profiler, Error Reporting. |
AI / ML |
Vertex AI, Gemini API (Google AI / Vertex), Model Garden, AutoML. |
CLI and SDKs#
The gcloud CLI plus language client libraries (google-cloud-*
packages). Shape: gcloud <group> <verb> [args].
$ gcloud auth login
$ gcloud config set project my-project
$ gcloud compute instances list
$ gcloud container clusters get-credentials prod --region us-central1
$ gcloud storage ls gs://my-bucket
$ gcloud logging read 'severity>=ERROR' --limit 20 --format=json
Regions and zones#
Regions span continents (us-central1, europe-west1). Each region has three or more zones (us-central1-a, b, c) with independent power and networking. Some services (Cloud Storage, BigQuery) are regional or multi-regional by default.
Identity in practice#
Service accounts are first-class IAM principals. Workloads authenticate as one.
Workload Identity Federation is the cross-cloud-trust mechanism. GitHub Actions, AWS workloads, and on-prem clusters can assume GCP service accounts without long-lived keys.
Workload Identity (Kubernetes-specific) ties GKE service accounts to GCP service accounts. The right pattern for pods.
Organization Policy Service caps what projects in the org can do (allowed regions, public IPs, service usage).
Pricing model#
Sustained Use Discounts on Compute Engine apply automatically for instances that run a large fraction of the month, no commitment required.
Committed Use Discounts for 1- or 3-year commitments on Compute Engine, GKE, BigQuery slots.
Spot VMs for interruptible capacity at a steep discount.
BigQuery bills per query bytes scanned (on-demand) or per slot-second (flat rate). The most operator-visible knob on a data warehouse.
When to pick GCP#
When the workload is Kubernetes-first; GKE is the most mature managed Kubernetes.
When the analytics shape is BigQuery (petabyte-scale SQL with no cluster to operate).
When the AI/ML stack is Vertex + Gemini; tighter integration than the multi-vendor patchwork on AWS or Azure.
When global low-latency reads matter (Cloud Load Balancing’s anycast).
When to pick something else#
When the catalog depth or niche-service breadth (AWS) is the driver.
When the estate is Microsoft-shaped (Azure).
When pricing transparency is the constraint; GCP’s pricing surface is wider than AWS’s but harder to predict than a fixed VPS bill.
Datacenter locations#
Code |
Location |
|---|---|
africa-south1 |
Johannesburg, South Africa |
asia-east1 |
Changhua County, Taiwan |
asia-east2 |
Hong Kong |
asia-northeast1 |
Tokyo, Japan |
asia-northeast2 |
Osaka, Japan |
asia-northeast3 |
Seoul, South Korea |
asia-south1 |
Mumbai, India |
asia-south2 |
Delhi, India |
asia-southeast1 |
Jurong West, Singapore |
asia-southeast2 |
Jakarta, Indonesia |
australia-southeast1 |
Sydney, Australia |
australia-southeast2 |
Melbourne, Australia |
europe-central2 |
Warsaw, Poland |
europe-north1 |
Hamina, Finland |
europe-north2 |
Stockholm, Sweden |
europe-southwest1 |
Madrid, Spain |
europe-west1 |
St. Ghislain, Belgium |
europe-west2 |
London, United Kingdom |
europe-west3 |
Frankfurt, Germany |
europe-west4 |
Eemshaven, Netherlands |
europe-west6 |
Zurich, Switzerland |
europe-west8 |
Milan, Italy |
europe-west9 |
Paris, France |
europe-west10 |
Berlin, Germany |
europe-west12 |
Turin, Italy |
me-central1 |
Doha, Qatar |
me-central2 |
Dammam, Saudi Arabia |
me-west1 |
Tel Aviv, Israel |
northamerica-northeast1 |
Montreal, Canada |
northamerica-northeast2 |
Toronto, Canada |
northamerica-south1 |
Querétaro, Mexico |
southamerica-east1 |
Osasco (São Paulo), Brazil |
southamerica-west1 |
Santiago, Chile |
us-central1 |
Council Bluffs, Iowa, USA |
us-east1 |
Moncks Corner, South Carolina, USA |
us-east4 |
Ashburn, Virginia, USA |
us-east5 |
Columbus, Ohio, USA |
us-south1 |
Dallas, Texas, USA |
us-west1 |
The Dalles, Oregon, USA |
us-west2 |
Los Angeles, California, USA |
us-west3 |
Salt Lake City, Utah, USA |
us-west4 |
Las Vegas, Nevada, USA |
References#
Cloud for the cross-cloud taxonomy.