MEPX
Chapter 7 of 10All chapters
  1. 1Start with numbers
  2. 2Tiers and statelessness
  3. 3Scaling
  4. 4Caching
  5. 5Data at scale
  6. 6Queues and async work
  7. 7Consistency
  8. 8Designing for failure
  9. 9Knowing what is happening
  10. 10Making the trade-off

Chapter 7 of 10

Consistency

CAP, and what it forces.

The trade

When the network splits, a distributed system can stay consistent or stay available, not both. The partition is not optional, so the design decides which one to sacrifice.

  • Banking picks consistency; a social feed usually picks availability.
  • Eventual consistency means correct soon, not correct now.

In practice

Most systems mix both: strong consistency for money and inventory, eventual for counters and feeds. Choosing per feature is more honest than choosing per database.