Skip to content

Share parsed query texts across different connections #345

@vlsi

Description

@vlsi

Cache of prepared statements in #319 uses per-connection cache, so in case of connection pool lots of queries are repeated in memory.

We can use "per-URL" kind of cache to share query texts and avoid parsing on connection renewals.

Current evaluation:
+ shared cache of parsed queries would reduce memory footprint of a connection pool

- additional cross-thread synchronization on con.prepareStatement might reduce throughput
- more complicated implementation might reduce performance and/or complicate support

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions