CAP Theorem

CAP Theorem Explained (with Examples)

The CAP Theorem is a key idea in distributed systems, introduced by Eric Brewer. It states that a distributed system can provide only two out of three guarantees at any time:

• Consistency (C): Every read gets the latest write (or an error).
• Availability (A): Every request receives a response, even if it doesn't contain the latest write.
• Partition Tolerance (P): The system works even if network partitions split communication between nodes.

Due to the reality of network failures, partition tolerance is a must for distributed systems. This means you have to choose between consistency and availability during a partition.

---

Understanding the Properties

• Consistency: All nodes show the same data at the same time.

  Example: Updating your profile photo on a social network and seeing the change everywhere instantly.

• Availability: The system always responds to requests, but data may not be the latest.

  Example: DNS servers always reply, even if a recent update hasn't reached every server yet.

• Partition Tolerance: The system stays operational even if parts can’t communicate.

  Example: A global database keeps running, even if two data centers can't talk to each other.

---

CAP Combinations with Real-World Examples

• CP (Consistency + Partition Tolerance):
  System favors correct data, but some requests may be denied during a partition.
  Example: HBase, MongoDB (in CP mode). Used in banking or finance, where correct data is critical.

• AP (Availability + Partition Tolerance):
  System always responds, but some data may be outdated during a partition.
  Example: DynamoDB, Cassandra. Used in shopping carts or social feeds where being up is more important than always being up-to-date.

• CA (Consistency + Availability):
  System always shows the latest data and responds to every request—but only as long as there’s no partition.
  Example: Single-node MySQL or PostgreSQL. Suitable for small, local systems.

---

Summary Table

---

Interview Tips

• Partition tolerance is essential in distributed systems.
• You must choose: Consistency or Availability during a partition—never all three.
• Use concrete examples to explain CAP trade-offs, tailored to the use case.