Back to Data & Analytics

Amazon Athena

aws

Serverless, interactive SQL query service querying Amazon S3 data lakes directly with Presto/Trino ($5.00/TB scanned).

DATA & ANALYTICS
Amazon Web Services

Serverless, interactive SQL query service querying Amazon S3 data lakes directly with Presto/Trino ($5.00/TB scanned).

Click any section above to jump directly to it
GCP ➔ AWS Bridge
Fast-Track Mental Model Translation

Coming from Google Cloud?

Compare with Google Cloud BigQuery (On-Demand / BigLake)

Amazon Athena is the direct AWS architectural twin to BigQuery On-Demand / BigLake, charging $5.00 per TB scanned over data lake files.

What's Conceptually Identical
  • 100% serverless SQL engines with $0 idle cost and zero cluster management.
  • Billed strictly on data scanned ($5.00 per TB).
  • Optimized by converting raw data to columnar Parquet/ORC format.
Key Architectural Shifts in AWS
  • Storage: BigQuery stores data in proprietary Capacitor format by default; Athena queries open Parquet/Iceberg files in standard S3 buckets.
  • Metastore: Athena relies on the AWS Glue Data Catalog; BigQuery has an integrated internal metadata engine.
  • SQL Dialect: Athena uses ANSI Presto/Trino SQL; BigQuery uses GoogleSQL.

What is Amazon Athena?

Amazon Athena is an interactive query service that makes it easy to analyze data in Amazon S3 using standard SQL. Athena is serverless, so there is no infrastructure to manage, and you pay only for the queries that you run. It is built on open-source Trino (Presto) and Apache Spark engines, with native integration with the AWS Glue Data Catalog.

Decision Guide: When to Choose This Service

When to Use Amazon Athena

Use Amazon Athena for serverless ad-hoc data lake analytics, querying raw logs in S3, analyzing Apache Iceberg tables, and cost-effective data exploration with $0 idle compute charges.

When NOT to Use (Recommended Alternatives)

Do NOT use Amazon Athena for high-throughput sub-second OLTP transactions, high-concurrency BI dashboards requiring sub-50ms latency (use Redshift with SPICE instead).

Core Architectural Pillars

Serverless $5.00/TB

Pay strictly for bytes scanned from Amazon S3 ($0 idle cost).

Glue Data Catalog

Central Hive-compatible metastore resolving schemas and partitions.

Partition Projection

Calculates S3 folder prefixes dynamically in memory without catalog lookups.

Open Formats

Query Parquet, ORC, JSON, CSV, Apache Iceberg, Delta Lake, and Hudi tables.