Trino Cluster
Trino is a distributed SQL query engine that federates queries across heterogeneous data sources, from object storage and Hadoop-style warehouses to relational databases, through a single ANSI SQL interface without first moving the underlying data. It originated as a fork of the Presto project by Presto's original creators and was renamed Trino in 2020.
Baseline topologies
| Topology | Minimum nodes | Roles |
|---|---|---|
| coordinator-plus-workers | 2 | trino-coordinator, trino-worker |
| coordinator-plus-workers-scaled | 4 | trino-coordinator, trino-worker |
What drives the cost
- Worker node count
- Number of Trino worker nodes in the cluster, each running query execution tasks assigned by the coordinator.
- Memory per worker (GB)
- Total JVM heap and off-heap memory allocated to each worker node; this is the primary offer-selection driver for Trino cluster sizing.
- Peak concurrent queries
- Expected peak number of queries executing simultaneously across the cluster, informing coordinator scheduling capacity and worker parallelism requirements.
Where it runs
bare_metal, dedicated, cloud_vm
Size Trino Cluster and get a cost estimate