Back to Papershelf

Papershelf

Bigtable: A Distributed Storage System for Structured Data

A storage paper about tablets, locality, compaction, and the shape of a distributed database built on top of a file system.

databasesdistributed systemsstorage

Why I read it

Bigtable sits in the space between filesystems and databases. It is a good paper for understanding how storage layout, serving paths, and compaction policy shape database behavior.

What it teaches

  • Data model choices show up later as operational behavior.
  • Tablet splitting and locality matter for scale and predictable access.
  • Compaction is not cleanup. It is part of the read and write path.

What I am watching

I am reading it for the backend intuition: how storage engines make tradeoffs before the application ever sends a query.