CAP Theorem in System Design

CAP Theorem in System Design

The CAP theorem states that a distributed data store can only guarantee two of the
following three properties at the same time:

• Consistency (C) — every read receives the most recent write or an error.
• Availability (A) — every request receives a (non-error) response, without the
  guarantee it contains the most recent write.
• Partition tolerance (P) — the system continues to operate despite network
  partitions between nodes.

The part people get wrong

In any real distributed system, network partitions will happen — so P isn't really
optional. The actual choice CAP forces on you is: when a partition happens, do you
sacrifice consistency (stay available, might serve stale data) or availability (refuse
requests until the partition heals, to keep data consistent)?

That's why you'll more often hear systems described as CP (e.g. ZooKeeper, etcd,
HBase) or AP (e.g. Cassandra, DynamoDB, Riak) rather than as satisfying all three.

See CAP Theorem Explained for concrete examples, and
Availability vs Consistency In System Design for how to actually decide
between them for a given feature.