I picked this up during my first year at Amazon, and it fundamentally changed how I approach system design.
Replication strategies. Partitioning tradeoffs. The spectrum between consistency and availability. Kleppmann doesn't just explain these concepts — he makes you feel why they matter, with real-world examples that hit different when you're operating services at scale.
The chapter on stream processing directly influenced how I redesigned our event-driven architecture at Buy With Prime. The mental model of "logs as the source of truth" clicked something into place that months of reading AWS documentation hadn't.
If you build anything that stores data, processes events, or serves users at scale — this is the foundation. Everything else is commentary.
Other Reads
Thirty-Three Syllables
On Aryabhata, Sanskrit verse, and the fifteen-century-old algorithm hiding inside every calculator you've ever used.
Borrowed Computation
What if quantum computers aren't fast because of clever engineering — but because they're splitting the universe and borrowing processing power from parallel realities?
The Number That Holds the Sky Together
How ancient Indian sages encoded the Sun-Moon distance ratio into a single sacred number — and why total solar eclipses are the proof they were right.