Dolt Use Cases: Version-Controlled SQL in Production
Dolt Use Cases: Version-Controlled SQL in Production
Source: https://www.dolthub.com/blog/2024-10-15-dolt-use-cases/
📌 Dolt is a MySQL-compatible database with Git-like branch, merge, and history for table data. Named customer patterns span game config shipping, ML training-data versioning, user-facing branch/merge apps, and compliance-friendly audit workflows.
🗄️ What Dolt Is
MySQL wire protocol and dialect plus Git semantics: branch, merge, clone, push, pull, commit, log, diff, and tag whole databases—not just schema files.
🎮 Game Development
Studios like Scorewarrior (Total Battle) keep game config/content in Dolt, develop on branches, and merge to main for orderly build-time and runtime ship cycles.
🧬 ML Training Data
Flock Safety versions feature/training sets with tags, diffs, and branches for collaboration and explainability; Turbine tracks data for AI cancer-drug / virtual cell simulations.
🔀 Version-Controlled Apps
Threekit exposes branch/merge over 3D product catalogs; Nautobot (Network to Code) routes network-config changes through Dolt-backed review before merge.
📋 Compliance & Audit
Built-in who/when/why history and dual-control review support SOX-like change management without separate CDC or audit-table machinery.
Key facts
| Fact | Value |
|---|---|
| Product | Dolt — Git semantics + MySQL-compatible SQL database |
| Primary post | “Dolt Use Cases” by Zach Musgrave, DoltHub Blog, Oct 15, 2024 |
| Core pattern | Isolated change, review, and history on structured data |
| Game customers | Scorewarrior (Total Battle) and other studios |
| ML customers | Flock Safety (feature store); Turbine (cell simulations) |
| App customers | Threekit (3D catalogs); Nautobot / Network to Code (network config) |
Details
Dolt sits between normal OLTP databases (no first-class branch/merge) and file-level Git (a poor fit for live SQL workloads). Teams get standard SQL clients and queries plus commits, logs, diffs, tags, and remotes for data.
The October 2024 DoltHub survey by Zach Musgrave frames real customer fit, not a pure feature matrix. The recurring need is isolated change, human review, and durable history on structured data—whether shipping game content, pinning training sets, or enforcing dual control on high-stakes configs.
Named examples include Scorewarrior branching game configuration, Flock Safety and Turbine versioning ML/scientific data, Threekit and Nautobot exposing or enforcing branch-and-merge in products, and regulated workflows that lean on built-in history instead of bolted-on audit plumbing.