Why it matters

Athena is AWS ad-hoc analytics. Understanding shapes S3 analytics.

Advertisement

The architecture

Query interface.

Glue Catalog for schema.

Presto engine.

Athena stackSQL queryPresto-basedGlue CatalogschemaS3 dataParquet + partitionedPartitioning + columnar format critical for cost + speed
Athena architecture.
Advertisement

How it works end to end

Pricing: $5 per TB scanned.

Partitioning + Parquet critical for cost.

Federated queries: query beyond S3.

CTAS: create table as select for materialization.