The short answer
Google Cloud Storage (GCS) is GCP's object store: buckets hold objects (files) addressed by name. It is the default landing zone for data lakes on GCP, just as S3 is on AWS.
gs://company-lake/
raw/events/dt=2026-09-05/
curated/dim_customer/GCS storage classes (mental model)
- Standard: hot / frequent
- Nearline: ~once a month access
- Coldline: ~once a quarter
- Archive: rare access / compliance
GCS vs S3
| Topic | GCS | S3 |
|---|---|---|
| URI style | gs://bucket/key | s3://bucket/key |
| Cloud home | GCP | AWS |
| Lake tooling | BigQuery external, Dataflow, Dataproc | Athena, Glue, EMR, Redshift Spectrum |
| Consistency | Strong consistency | Strong consistency (modern S3) |
| Identity | GCP + bucket IAM | AWS IAM + bucket policies |
| Same idea? | Yes: durable for lakes | Yes |
Practical note for freshers
APIs and SDKs differ (google-cloud-storage vs boto3), but the *concept* is the same: cheap durable blobs, partition by prefix, lifecycle cold data, lock down IAM.
Interview tip: "GCS is GCP's S3." Then name one class pair (Standard vs Nearline) and one integration (BigQuery external tables or Dataflow).