Trino
What it is: Fast distributed SQL query engine (formerly PrestoSQL). Query anything—data lakes, databases, NoSQL—with standard SQL. Presto's active fork.
Quick answer: Trino is the actively developed fork of Presto, maintained by Presto's original creators since 2019. It queries data lakes, warehouses, and databases with one SQL dialect through 50+ connectors. It's free under Apache 2.0; choose Trino over Presto for any new deployment, or use Amazon Athena or Starburst Galaxy if you want it managed.
Presto vs Trino: what happened?
Presto's original creators left Meta (then Facebook) in 2018 and forked the project; in 2020 the fork was renamed from PrestoSQL to Trino. Trino gets the large majority of new features today — fault-tolerant execution, mature Apache Iceberg and Delta Lake connectors, faster releases — while Presto continues under the Linux Foundation, used most notably inside Meta itself.
Do you have to run Trino yourself?
No. Amazon Athena runs a Trino-based engine serverless (pay per TB scanned), and Starburst Galaxy is the managed cloud from Trino's creators. Self-managing on Kubernetes gives the most control but requires real ops investment — cluster sizing, memory tuning, and connector upkeep.
What It Does Best
Unified SQL interface. 50+ connectors. Query S3, Delta Lake, PostgreSQL, Kafka, Elasticsearch in one SQL query.
Active development. Fastest-moving fork of Presto. Regular releases, active community, enterprise backing.
Cloud-native. Kubernetes-ready, auto-scaling. Separate compute and storage.
Key Features
50+ connectors: Query any data source with SQL
Cost-based optimizer: Smart query planning across sources
Table formats: Native support for Iceberg, Delta Lake, Hudi
Fault tolerance: Query retries and graceful recovery
ANSI SQL: Standard SQL across all data sources
Pricing
Open Source: Free, Apache 2.0 license
Starburst Enterprise: Contact for pricing (support + features)
AWS EMR: EC2 costs + EMR charge
Self-managed: Cloud infrastructure costs only
When to Use It
✅ Modern data lake analytics (Iceberg, Delta, Hudi)
✅ Cross-database queries and federation
✅ Cloud data warehouse alternative
✅ Real-time analytics on streaming data
✅ New projects (preferred over Presto)
When NOT to Use It
❌ Need data storage (query engine only)
❌ Extremely complex ETL (use Spark)
❌ Small team without ops resources
❌ Very simple queries on small data (overkill)
❌ Need transactional capabilities
Common Use Cases
Data lake SQL: Query Parquet, ORC, Avro files on S3
Table format queries: Apache Iceberg, Delta Lake analytics
Cross-source joins: Combine data from multiple systems
BI tool queries: Power Tableau, Looker on data lakes
Ad-hoc exploration: Interactive queries without data movement
Trino vs Alternatives
vs Presto: Trino more active and modern, choose Trino for new work — see what is Presto SQL for the history
vs Athena: Trino more control and features, Athena serverless and simpler
vs Spark: Trino for interactive queries, Spark for batch processing
vs Starburst: Same engine — Starburst adds management, security, and support on top of Trino
Unique Strengths
Active community: Most active Presto fork with regular releases
Modern table formats: Best support for Iceberg, Delta, Hudi
Query federation: Join across any data source seamlessly
Cloud-native design: Built for modern cloud architectures
Bottom line: The leading SQL engine for modern data stacks. More active than Presto, better features. Excellent for data lakes and federation. Choose Trino over Presto for new projects.