MiniDB

Overview

MiniDB is a database engine built from scratch in Go to demonstrate the moving parts of a real RDBMS — storage, durability, concurrency, indexing, and query execution — without the abstraction layers that hide them in production systems.

Storage: LSM Tree

• MemTable for fast in-memory writes
• SSTables on disk, sorted and immutable
• Background compaction merges SSTables to reclaim space and improve read performance
• Optimised for write-heavy workloads

Durability: Write-Ahead Logging

Every mutation is appended to the WAL before being applied to the MemTable. On restart, the WAL is replayed — no committed write is ever lost, even after a crash.

Transactions: ACID via MVCC

• Multi-Version Concurrency Control — each transaction sees a consistent snapshot
• Snapshot Isolation — readers don't block writers, writers don't block readers
• Full ACID guarantees

Query layer

• Custom SQL parser supporting SELECT, INSERT, and WHERE clauses
• Hash index for O(1) primary-key lookups
• Execution engine with a clean interpreter pattern

Interfaces

• Interactive REPL for ad-hoc exploration:

  minidb> INSERT INTO users VALUES (1, 'Alice')
  minidb> SELECT * FROM users WHERE id = 1
  [1 Alice]

• TCP server mode for client–server use

Why build it

Reading about LSM trees, MVCC, and WAL gives you the concepts. Writing one from scratch is what makes them stick — the failure modes, the trade-offs in compaction strategies, the subtle bugs in snapshot isolation are only visible when you live inside the implementation.