Course contentsShow
← System Design
Lesson 523 of 1,91915. Consistency ModelsFree lesson

Linearizability Defined

The strongest consistency model where operations appear instantaneous and respect real-time ordering.

Linearizability Defined

What You'll Learn

You'll learn what linearizability means and why it's considered the gold standard for strong consistency in distributed systems.

The Strongest Consistency Guarantee

Linearizability is the strictest consistency model possible in distributed systems. It guarantees that every operation appears to take effect instantaneously at some point between when it starts and when it completes, and all operations respect real-time ordering.

Think of it like a single, magical box with a slot: you drop your request in, and instantly (from everyone's perspective) the system state changes. No one can ever observe an "old" value after a "new" one has been written, and if operation A completes before operation B starts, everyone must see A's effect before B's.

What Makes Linearizability Special

In a linearizable system:

  1. Total order: All operations appear to happen in a single, global sequence
  2. Real-time constraint: If operation A completes before operation B begins (in wall-clock time), A must appear before B in that sequence
  3. Instantaneous effect: Each operation takes effect at a single, atomic point in time during its execution

For example, if you write x=5 and that write completes at 10:00:00, then any read starting at 10:00:01 anywhere in the system must see x=5 or a newer value—never the old value.

Why It Matters

Linearizability makes distributed systems behave like a single machine with no replication. There's no confusion about "which value is current" or "did my write succeed?" It's the consistency model that matches human intuition about how storage should work.

Key Takeaway: Linearizability is the strongest consistency model where every operation appears to happen atomically at a single instant in real-time order, making a distributed system behave like a single, reliable machine.