Boost Real‑Time App Performance with ReadySet: A Hands‑On Guide to SQL Caching

Introduction

ReadySet is an open‑source SQL cache engine that sits between an application and the database, supporting MySQL and PostgreSQL. It aims to accelerate real‑time applications by caching complex queries without requiring code changes or database migration.

Architecture

ReadySet operates as a proxy layer (Java/PHP application → ReadySet → MySQL) and is implemented as a lightweight in‑memory cache. It provides the same interface as Redis or Memcached but works with standard SQL.

How ReadySet Works

Read operations

Send the SQL query directly to ReadySet.

If the result is cached, ReadySet returns it instantly, achieving millisecond‑level latency.

If not cached, ReadySet forwards the request to the backend database, caches the result, and returns it.

Write operations (INSERT, UPDATE, DELETE)

ReadySet forwards the write request to the backend database and updates the cache accordingly.

Advantages

Uses standard SQL; no new query language to learn.

No code changes required; only the database connection string needs to be updated.

Supports high‑speed caching of complex SQL queries.

Automatically keeps the cache consistent with the database.

Allows selective caching of high‑frequency queries to optimize resource usage.

This design makes ReadySet especially suitable for read‑intensive workloads, delivering significant query‑performance gains while preserving simplicity and eventual consistency.

Installation and Usage

1. Pull the ReadySet Docker image shell> docker pull readysettech/readyset 2. Run the ReadySet service

shell> docker run -d -p 3307:3307 -p 6034:6034 \
    --name readyset \
    -e UPSTREAM_DB_URL=mysql://admin:[email protected]:6666/test \
    -e LISTEN_ADDRESS=0.0.0.0:3307 \
    readysettech/readyset:latest

Note: The UPSTREAM_DB_URL must contain the backend MySQL credentials, host, port, and database name with ALL privileges.

ReadySet exposes two ports: 3307 for query traffic and 6034 for metrics (/metrics endpoint).

After starting, ReadySet connects to the database and replicates the specified tables, which may take seconds to hours depending on size and network conditions.

Check replication status with: shell> docker logs readyset When you see INFO replicators::noria_adapter: MySQL connected, the system is ready to cache queries.

You can also query the status via MySQL client:

When the Snapshot Status column shows Completed, ReadySet is ready.

3. Cache a query

Connect to ReadySet on port 3307 with a MySQL client and run the slow query, then execute: SHOW PROXIED QUERIES; Check the readyset supported column; a value of Yes means the query can be cached.

4. Create a SQL cache CREATE CACHE FROM <query_id>; Example: mysql> CREATE CACHE FROM q_4c1cf3b8080fe634; 5. View cached queries

SHOW CACHES;

6. Test performance

Run the cached slow SQL against ReadySet (port 3307) and directly against the backend MySQL. The ReadySet execution should return results in seconds, demonstrating the cache’s effectiveness after the first execution.

Initial tests show that ReadySet’s caching becomes noticeable immediately after the first query, providing substantial latency reduction.