MySQL is the backbone of countless applications, but scaling read heavy workloads often requires complex manual caching logic. This talk explores how Readyset takes MySQL to the next level by offloading queries through two distinct caching modes: deep and shallow. We will dive into the underlying architecture of both, focusing on how cache hits are served, how misses are handled, and how the system maintains correctness and efficiency without application changes.

Deep cache maintains materialized views in a streaming dataflow graph. On a hit, Readyset serves results from that precomputed state with sub millisecond latency. We will examine how Readyset executes queries inside its dataflow engine while preserving upstream SQL semantics and incremental maintenance. Unlike a classic database executor that chooses strategies at query time, deep cache fixes the execution path at materialization time (CREATE CACHE). This plan is used both to maintain the view as base data changes and to run the query on a miss, which fundamentally shapes how relational operators are designed and optimized.

Shallow cache provides TTL based result caching for queries that bypass the dataflow graph, supporting a broader set of MySQL features. On a miss, the query executes against the upstream database. Results are keyed by parameterized query templates, utilizing coalescing to prevent thundering herds and configurable refresh behavior to ensure data freshness. We outline the mechanics of parameterization, TTL logic, and the interaction between the cache and the upstream MySQL instance.

The talk concludes with recent improvements in execution internals and the roadmap for optimizing both in engine execution and upstream query paths to further extend MySQL performance.

Audience: Engineers interested in MySQL scaling, query execution, streaming dataflow, and low latency query serving.