All projects

systems cli

A key-value store that survives a crash

Write-ahead logging, compaction, and durability you can prove by pulling the plug.

advancedRust~40h

Why it matters

Building a storage engine demystifies databases permanently. Being able to explain a write-ahead log is a genuinely senior-sounding answer.

What you build

  • An append-only log with an in-memory index
  • Crash recovery by replaying the log
  • Compaction that reclaims space
  • A crash test that kills the process mid-write

If you want more

  • Add range queries with a sorted structure
  • Add a Bloom filter

Build it 5 steps

  • Compact away superseded entries and reclaim space

Material for the whole build

Shows you understand durability from the inside, which changes how you use every database afterwards.

Get a plan built around projects like this