BigQuery
Usage BasedAnalyze petabytes of data using familiar SQL. BigQuery is Google Cloud's fully managed data warehouse.
Scores
About
BigQuery is Google Cloud's serverless, highly scalable data warehouse designed to run analytical SQL queries across petabyte-scale datasets in seconds. Launched in 2011, it pioneered the separation of storage and compute in cloud data warehouses, eliminating the need to provision or manage clusters.
Storage and compute separation means you pay independently for data stored and for queries executed. Data is stored in Google's Capacitor columnar format across Google's distributed file system; compute (Dremel query engine) scales automatically with query complexity and concurrency.
Pricing models: BigQuery offers two compute pricing modes — on-demand ($6.25/TB of data processed per query, with the first 1 TB/month free) and capacity commitments (flat-rate slots bought by the hour, month, or year for predictable costs on high-volume workloads). Storage is charged at $0.02/GB/month for active storage and $0.01/GB/month for long-term storage (data untouched for 90+ days).
BigQuery ML allows data analysts to build and run machine learning models directly in SQL without exporting data to a separate ML platform. Supported models include linear regression, logistic regression, k-means, matrix factorization, and integration with Vertex AI for custom models.
BigQuery Omni extends BigQuery to run queries against data stored in AWS S3 or Azure Blob Storage without moving the data, using Google's Anthos infrastructure.
Streaming inserts allow near real-time data ingestion at high throughput; batch loading via Cloud Storage, Pub/Sub, Dataflow, or direct API is also supported.
BI Engine is an in-memory analysis service that accelerates queries for connected dashboards (Looker Studio, Looker, Tableau) with sub-second response times on repeated queries.
BigQuery supports federated queries against Cloud Spanner, Cloud SQL, Cloud Bigtable, and Google Sheets without data movement. Its IAM integration, row/column-level security, data masking, and VPC Service Controls make it suitable for regulated data environments.
Key Features
- Serverless, fully managed columnar data warehouse — no clusters to provision or manage
- On-demand pricing ($6.25/TB queried) and capacity commitment slots for predictable costs
- BigQuery ML: build and run ML models in SQL without leaving the warehouse
- BigQuery Omni: federated queries across AWS S3 and Azure Blob Storage
- Streaming inserts for near real-time data ingestion at high throughput
- BI Engine for sub-second in-memory acceleration of dashboard queries
- Row/column-level security, data masking, and VPC Service Controls
Pros
- Truly serverless — zero infrastructure management, auto-scales to petabytes
- First 1 TB of queries per month is free, making it accessible for experimentation
- Deep integration with the broader Google Cloud ecosystem (Dataflow, Pub/Sub, Looker, Vertex AI)
- BigQuery ML eliminates the need to move data out of the warehouse for ML workloads
- Column-level security and row-level access policies simplify data governance at scale
Cons
- On-demand pricing can produce large unexpected bills if queries are written without scanning filters
- Not suited for transactional (OLTP) workloads — optimised for analytics, not row-level writes
- Slot-based capacity pricing requires upfront commitment and careful capacity planning
- Vendor lock-in to Google Cloud — migrating petabyte-scale datasets is costly and slow
- DML operations (UPDATE/DELETE/MERGE) are slower and more expensive than in traditional RDBMS
Pricing
Usage Based- · $6.25 per TB of data processed by queries
- · First 1 TB of queries per month free
- · $0.02/GB/month active storage, $0.01/GB/month long-term storage
- · 10 GB of free storage per month
- · Flat-rate compute slots: pay per slot-hour or commit monthly/annually
- · 100-slot minimums; pricing varies by region
- · Suitable for consistent high-volume query workloads
- · Unused slots can be shared across projects
- · Custom slot commitments with 1-year or 3-year discounts
- · Multi-region replication and CMEK
- · Advanced security and compliance features
- · Dedicated support and SLA options
Possible Stacks
GCP ELT Pipeline
ProjectA fully managed, serverless ELT pipeline on Google Cloud: Fivetran handles ingestion with zero-maintenance connectors; BigQuery stores and queries petabytes without cluster management; dbt transforms data into analytics-ready models; Dagster orchestrates the pipeline as typed, lineage-tracked assets; Metabase provides self-service BI on top.
Related Tools
Works well with (10)
Alternative to (6)
Learning Resources
No resources yet — check back soon.
