Trino

Definition

Trino is an open-source, distributed SQL query engine, originally forked from Presto as PrestoSQL, built for fast, interactive analytics across large-scale and federated data sources.

Overview

Trino was created when several of the original engineers behind Presto at Facebook left to found a new company and continued development of the engine under a new name, first as PrestoSQL and later rebranded to Trino to avoid confusion with the original project. Architecturally, Trino follows the same core design as Presto: a coordinator node plans queries, which are then executed in parallel and largely in memory across a cluster of worker nodes, without Trino itself storing any data. Like Presto, Trino connects to external systems — data lakes, relational databases, NoSQL stores, and message queues — through a pluggable connector framework, so a single SQL query can join data living in a data lake with data in a system like PostgreSQL or Elasticsearch. This makes it a common choice for building a unified SQL access layer over a heterogeneous set of storage systems rather than requiring every consumer to speak each system's native query language. Trino is widely paired with open table formats such as Apache Iceberg and Delta Lake to provide the SQL query layer of a modern data lakehouse, and it has become one of the most actively developed engines in that space, used by large technology companies for petabyte-scale interactive analytics.

Key Features

• Distributed, in-memory MPP SQL query execution
• Federated queries across data lakes, databases, and other sources
• Pluggable connector architecture for heterogeneous data
• ANSI SQL support with a familiar query interface
• Horizontal scalability by adding worker nodes
• Active development lineage originating from the Presto project
• Strong integration with open table formats like Iceberg and Delta Lake

Use Cases

Frequently Asked Questions

Related in Database