Covers 26:24–30:50. Why a second edition, the collaboration with Chris Riccomini, and the big new theme.

Why and how it happened

  • A second edition had clearly been needed for a couple of years: the first edition was getting dated and technology changes weren't reflected. But Martin now has an academic job (research + teaching), so the update was a side project and progress was very slow — the same context-switching problem as the first edition. 26:25 26:56
  • He also realized he'd lost touch with current industry practice (e.g. data lakes) after going deep into theory in academia. 26:56 27:19

The collaboration with Chris Riccomini

  • Chris Riccomini: old LinkedIn colleague from stream processing, author of The Missing README, writer of the Materialized View newsletter on data systems trends, and a startup investor in that space. 27:19 27:46
  • Division of labor: Chris brought up-to-date knowledge of the industry cutting edge; Martin brought strong opinions on how to teach — precise, carefully chosen wording that remains accessible. "My writing style plus Chris's knowledge of latest industry trends." 28:09 28:38

The big new theme: cloud-native architecture

  • Planned from the start: reflecting cloud-native systems architecture — building data systems on top of cloud services as the foundational abstraction. 28:38 29:03
  • First edition's implicit assumption: machines with local disks; a database instance runs on a machine, writes to local disk, and replication happens at the database level between machines. 29:03
  • What changed: people now build databases on top of object stores, where replication happens at the object-store level. An object store is a genuinely new abstraction — unlike a virtual block device (e.g. EBS), which still mimics a single-node OS block device with a filesystem on top. 29:33 29:59
  • This style was just emerging at the time of the first edition and has since taken off; the idea is integrated throughout the whole book, not confined to one section. 30:23