Skip to content
Deepak Gupta markBooks

Books

Designing Data-Intensive Applications

Editorial pick

The Big Ideas Behind Reliable, Scalable, and Maintainable Systems

By Martin Kleppmann · O'Reilly Media · 2017

The single textbook every backend engineer should own. Survives every fashionable database wave.

Textbook 350+ pages(616p) Advanced Published 2017
Amazon Publisher Goodreads Author site

Editorial take

Kleppmann's book is a generational textbook. It explains the fundamental ideas behind storage, replication, partitioning, consistency, consensus, and stream processing in a way that's neither tied to a specific database vendor nor allergic to detail. If you've ever wondered what "linearizability" actually means, what trade-offs Kafka is really making, or why your shiny new database vendor is hand-waving the consensus question, this is the book. Dense, but never unnecessarily so. It's the rare technical book that holds up a decade after publication because it's mostly about ideas, not products.

Last hand-checked 2026-05-18.

Read if you …

  • design or operate distributed systems at any non-trivial scale
  • are a senior backend engineer studying for system-design interviews
  • evaluate database / streaming vendors and want to see through marketing

Skip if you …

  • you're a frontend / mobile engineer — most of this won't apply day-to-day
  • you want hands-on tutorials with running code — this is principles-first

If you only read one chapter

Consistency and Consensus

Chapter 9 is the cleanest treatment of distributed-consensus trade-offs you'll find outside academic papers. Worth the price alone.

Key ideas

  • Reliability, scalability, and maintainability are the three north stars — name them explicitly.
  • Replication, partitioning, and transactions are orthogonal — design each one deliberately.
  • Stream and batch processing are increasingly the same thing, viewed at different cadences.
  • Every database vendor's pitch hides a CAP-style trade-off; learn to find it.

About the book

Kleppmann taught at Cambridge before writing this book at O'Reilly. The result is the most ambitious recent attempt to synthesize the distributed-systems literature for working engineers — covering relational and non-relational storage, replication protocols, transactional models, MapReduce-and-after batch processing, and stream processing with Kafka-class systems, with footnotes leading to the relevant academic papers throughout.

The 2nd edition is in progress; the 1st (2017) is still the canonical version. Treat it as both a textbook and a reference — most working engineers read it once linearly, then keep it on the shelf for the next time they're evaluating a database.

Pairs with

If Designing Data-Intensive Applications works for you, these likely will too.