Part 1 of 7 · 1 min read
Why this page closes the level
In this module you will
- Define system of record and derived data, and give four examples of derived data.
- Show two ways in which dual writes leave systems different for good.
- Explain change data capture (CDC), and compare log reading with triggers.
- Start a new derived system with a snapshot plus a change stream, or with a compacted log.
- Design the update path for a search index and a cache, and say what lag the user may see.
Most real systems keep the same facts in several places: a database, a cache, a search index, a warehouse. Every page of this level led here. Queues and logs carry the changes, delivery guarantees make repeats safe, event sourcing and CQRS describe the idea in application terms, and batch and stream jobs compute the copies. This page puts them together to answer a plain question: when data changes in one place, how do the other places find out, reliably? The books used are "Designing Data-Intensive Applications" (DDIA) and "Understanding Distributed Systems" (UDS). The numbers are made up to show the idea, unless the text names a book.