What Is the CAP Theorem? How to Apply It in System Design?

CAP Theorem

Three Properties

In a distributed system, you can guarantee at most two of:

• C — Consistency: All nodes see the same data at the same time
• A — Availability: Every request receives a response (not necessarily the latest data)
• P — Partition Tolerance: System continues operating despite network partitions

Why Must We Choose P?

Network partitions are unavoidable in real distributed systems (failures are the norm). So the real choice is CP vs AP.

CP Systems (Consistency + Partition Tolerance)

• Reject requests during partitions to guarantee consistency
• Examples: HBase, ZooKeeper, traditional RDBMS

AP Systems (Availability + Partition Tolerance)

• Still respond during partitions (may return stale data); eventual consistency
• Examples: Cassandra, CouchDB, DynamoDB

Extension: PACELC Theorem

CAP only addresses partition scenarios; PACELC is more complete:

• With partition: A vs C
• Without partition (normal operation): Latency vs Consistency