I have to find a way to cache results from queries from Trino+Hive, but I don’t find too many documentation about Trino cache performance. I’m thinking about a Redis cache too.
Can you send me a link about cache performance using Trino+Hive?
Core Trino’s only caching opportunity for Hive catalogs is called “Hive connector storage caching” and described at Hive connector storage caching — Trino 427 Documentation. I do not believe any published benchmarks are available for this feature.
It focuses on caching of data from the lake with storage happening on local disks (could be spinning disks, but SSD and/or RAM disks are more appropriate), not “results caching” as you asked about in your question.
Starburst Galaxy is about to be previewing its “Results Cache” and “Repeated Subquery Caching” features, but that is not in core Trino.
Other existing opportunities for cache behaviors are Table Scan Redirections and Materialized Views enabled on Starburst Enterprise via the cache service; Cache service — Starburst Enterprise.
Also, Starburst Galaxy and Starburst Enterprise offer Warp Speed, Warp Speed | Starburst, for autonomous indexing and caching of data lake tables.