liveBook · Manning

Trino

As explained in Architecting an Apache Iceberg Lakehouse

PrestoSQL was the forked project name that later became Trino.

Trino is an open-source, distributed SQL engine for high-performance analytics across diverse data environments. It targets read-heavy OLAP workloads rather than OLTP, providing one SQL interface for federating queries across lakehouses, relational databases, NoSQL systems, streaming platforms, and analytical stores. section 8.3.5 section 8.3

A coordinator parses and plans each query, then distributes stages, tasks, and splits across workers for parallel execution. Trino can push down filters, minimize data movement, and scale compute independently, making it effective for large or fragmented environments. section 8.3.5 section 8.3.7

Trino’s connector architecture supports:

  • Independent catalogs configured for data lakes and lakehouses such as Iceberg, Delta Lake, Hudi, and Hive. section 8.3.6
  • Connections to cloud storage, relational databases, warehouses, NoSQL stores, Kafka, and analytical systems. section 8.3.6
  • Custom integrations through a service provider interface, allowing developers to add connectors for additional systems. section 8.3.6

Trino focuses on query execution and source connectivity, not native semantic modeling, data curation, or business-facing UI. Organizations needing governed metrics can pair it with tools such as Cube, AtScale, Select Star, or dbt Metrics, or use a commercial Trino platform with added capabilities. section 8.3.9 section 8.3

Deployment ranges from self-managed clusters on premises, virtual machines, or Kubernetes to managed offerings. Self-management preserves control but requires cluster operations and tuning; managed services reduce overhead. Trino Gateway can provide shared access, workload-aware routing, and cluster isolation. section 8.4.2 section 8.3.7

Choosing Trino for fragmented data

A company can query operational data across cloud platforms, relational databases, NoSQL systems, and flat files without first centralizing or replicating it. For occasional Hudi analysis in Amazon S3, AWS Athena provides serverless, pay-per-query access powered by Trino. section 8.5.1 section 8.5.4

Figure 8.9 Trino can connect to databases, data lakes, data warehouses, lakehouse catalogs, and streaming systems, making the results of federated queries accessible via REST API, JDBC, ODBC, and a variety of community-developed MCP servers.

Map Trino’s federation surface across databases, data lakes, warehouses, lakehouse catalogs, streaming systems, and client access interfaces. section 8.3.6

How Architecting an Apache Iceberg Lakehouse covers it

How deep it goes
Explains coordinator-worker execution, connector extensibility, deployment choices, routing, and semantic-layer limitations.
Examples
Uses fragmented enterprise sources, Hudi on Amazon S3 through Athena, and on-premises Cloudera modernization from Impala.
What it assumes
Assumes SQL and distributed data-platform context while distinguishing OLAP from OLTP and explaining the federation architecture.
What you can do after
Operate/configure federated analytics across multiple sources, selecting self-managed or managed deployment according to workload and operational needs.
How it connects
Connects federated querying with Apache Iceberg, Delta Lake, Hudi, Apache Kafka, semantic layers, AWS Athena, and Trino Gateway.
Cover of Architecting an Apache Iceberg Lakehouse
liveBookmanning.com
Read Architecting an Apache Iceberg Lakehouse on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Architecting an Apache Iceberg Lakehouse. The section links point to the parts of the book it draws on.