Consistency

• Every read receives the most recent write or an error.
• If you update a piece of data in the system, any subsequent access to that data will always return the updated value, ensuring that all users see the same data at the same time.

Availability

• Every request receives a (non-error) response, without the guarantee that it contains the most recent write.
• The system is always up and running, and every request is answered, but you might not always get the latest data.

Partition (Outage) Tolerance

• The system continues to operate despite an arbitrary number of messages being dropped (or delayed) by the network between nodes.
• Even if there's a problem in the network and some parts of the system can't talk to each other, the system as a whole still works.

You can pick two

In distributed systems, where the system should work even if a part of it fails (must-P), you can peek between availability and consistency. And you normally would take availiability for normal web

Examples for CAP:

• Domain Name System (DNS): A, P
  • Availability and Patition Tolerance very good, but can be inconsistent
• Relational database systems: C, A
  • Partitioning is tolerated e.g. due to transaction journals
• Banking applications: C, P
  • Consistency is the most important aspect, even with severe communication failures