Last Minute Notes
  • Github
  • Docs
  • Learn
    @sayanstwt
Lessons
0/37 done
  • Timeouts, Retries & Breakers
  • Multi-Region Design
  • Cells & Shuffle Sharding
  • SLIs, SLOs & Error Budgets
  • Observability
  • Deployment Strategies
  • Backups & Disaster Recovery
Learn/System Design/Reliability & Operations

Multi-Region Design

8 min Advanced 7 steps

Active-passive vs active-active, and what each costs you.

Regions
1
Asia latency
high
Region outage
-
US-EASTUserin Asiafar away: slowUserin the USAppDatabasefastfar
request response data
In this lessonStep 1 of 7
  1. The problem
  2. Multi-AZ first
  3. RTO and RPO
  4. Strategies
  5. Home region
  6. Write anywhere
  7. Recap
Step 1 / 7problem

One region is far away and fails as one

Everything runs in us-east. Users in Asia pay for the distance on every request.

And when the region has an outage, every user is down at once.

Key idea

Teams go multi-region for lower latency, to survive a regional outage, or for data residency. It costs complexity, consistency and money.

←→ step·R replay·Space pause

You'll learn

  • Why go multi-region, and why multi-AZ is usually the right start
  • Read RTO and RPO as downtime and data loss
  • Compare backup and restore, pilot light, warm standby and active-active
  • Choose a write strategy for active-active data
Read the full notesDistributed Systems, 12