No. 126 of 500 No. 10 in Hands-On AI & Programming

Designing Data-Intensive Applications

The Big Ideas Behind Reliable, Scalable, and Maintainable Systems

By Martin Kleppmann and Chris Riccomini

Print length
670 pages
Publisher
O'Reilly Media
Reading level
Advanced
ISBN-13
978-1098119065
Published
March 24, 2026
Written for
Backend, data and ML platform engineers who design or operate systems that store and move data.

Synopsis

Martin Kleppmann explains the ideas inside the databases, caches, message queues and processing frameworks on which large software systems, including AI platforms, depend. Rather than teaching any single product, the book compares the trade-offs behind relational and document data models, log-structured and B-tree storage engines, leader-based and leaderless replication, partitioning, transactions and isolation levels, and consensus in distributed systems where clocks drift and networks fail. Its final part covers batch and stream processing and makes the case for building systems around immutable event logs and derived data. The first edition appeared in 2017; the second edition, co-written with Chris Riccomini and published by O'Reilly in March 2026, integrates newer technologies, including the design of major cloud data services.

Kleppmann writes with a researcher's precision and a practitioner's sense of what fails in production, relying on diagrams and extensive references to the research literature rather than code listings. Readers gain the ability to reason about consistency, latency and failure, to read vendor claims critically, and to choose tools by their guarantees instead of their marketing. The book does not teach machine learning, but engineers building training-data pipelines, feature stores or retrieval systems face exactly the problems it analyzes.

Scorecard

Editorial score

How it scored

Editorial score: 86.3 / 1005 weighted criteria

  • Influence30%

    27.0/30
  • Authority25%

    22.5/25
  • Practical value20%

    17.0/20
  • Currency15%

    12.8/15
  • Readability10%

    7.0/10
  • Editorial score

    86.3/100

How the rank was computed

Rank No. 126 of 500 comes from a composite of 61.3: observable signals 22.4 of 55 points plus editorial review 38.9 of 45 points.

All 11 inputs
InputPoints
Observable signals22.4/55
Reach (Amazon ratings volume)9.2/16.5
Consensus across best-of lists2.8/13.8
Awards and bestseller evidence3.3/8.3
Reader rating, sample-adjusted1.8/8.3
Amazon sales rank2.5/5.5
Recency2.8/2.8
Editorial review38.9/45
Author authority10.1/11.3
Influence12.2/13.5
Practical value7.7/9.0
Readability3.2/4.5
Currency5.7/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

The standard engineering reference on how databases and distributed data systems work, now updated in a 2026 second edition.

Strengths

  • Explains trade-offs, not vendor products
  • Rigorous yet readable prose with extensive references
  • Second edition updated for cloud-era systems

Keep in mind

It is long and conceptual, with little code and no machine learning instruction, so readers seeking hands-on model building should choose another title.

Key takeaways

  1. Reliability, scalability and maintainability compete with one another, so every architecture choice is a trade-off rather than a free improvement.
  2. Storage engines split into log-structured and B-tree families, whose different read and write costs determine which workloads each handles well.
  3. Replication and partitioning introduce consistency anomalies, so engineers must understand isolation levels and guarantees such as linearizability before trusting a database.
  4. Treating an immutable event log as the source of truth, with derived views maintained by stream processing, simplifies keeping many data systems consistent.

Notable facts

  • The first edition sold 108,079 copies across print, ebook and translations by mid-2020, according to Kleppmann's September 2020 blog post on the economics of book writing. Source (opens in a new tab)
  • Microsoft CTO Kevin Scott wrote that 'This book should be required reading for software engineers.' Source (opens in a new tab)
  • O'Reilly Media published the second edition, co-authored with Chris Riccomini, in March 2026. Source (opens in a new tab)

About the author

Martin Kleppmann and Chris Riccomini

Martin Kleppmann is an associate professor at the University of Cambridge, researching local-first collaboration software and distributed systems security, and one of the people behind the Automerge open-source project. He previously co-founded Rapportive, acquired by LinkedIn in 2012, and worked on large-scale data infrastructure at LinkedIn. The 2026 second edition is co-written with software engineer Chris Riccomini.

Questions

What is Designing Data-Intensive Applications about?
It explains the principles behind modern data systems: data models, storage engines, replication, partitioning, transactions, consistency, consensus, and batch and stream processing. Instead of teaching specific products, Martin Kleppmann compares the trade-offs that shape databases and distributed systems, so engineers can choose tools by their guarantees. A second edition, co-written with Chris Riccomini, was published by O'Reilly in March 2026.
Why does Designing Data-Intensive Applications appear on a list of AI books?
AI products depend on infrastructure that stores, moves and serves data, from training-data pipelines to feature stores and retrieval systems. This book is a widely read explanation of how that infrastructure behaves under load and failure; its first edition sold more than 100,000 copies by 2020. It teaches little about machine learning itself, but ML platform engineers rely on its concepts.
Should I read the first or second edition of Designing Data-Intensive Applications?
Readers buying today should choose the second edition, published in March 2026 and co-written with Chris Riccomini, because it integrates technologies and cloud services that matured after 2017. The first edition's core explanations of replication, partitioning, transactions and stream processing remain sound, so owners of the 2017 edition still have a reliable guide to the fundamentals.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.