Skip to content
System Design

Consistency and Availability

Scope

Models and guarantees around how and when data changes become visible to different parts of a distributed system.

Why This Topic Exists

Most large-scale system failures are caused by misunderstood consistency assumptions.

Core Tradeoffs

  • Strong consistency vs latency
  • Availability vs correctness
  • Operational complexity vs developer ergonomics

Common Failure Modes

  • Split-brain scenarios
  • Serving stale data unknowingly
  • Over-constraining systems unnecessarily

Interview Signals

Strong candidates can articulate what guarantees the system actually needs—not what sounds safest.

  • Databases
  • Caching
  • Distributed consensus