CAP Theorem
The CAP theorem, proposed by Eric Brewer, states that in a distributed system, you can achieve at most two out of the three properties: Consistency (C), Availability (A), and Partition Tolerance (P).
According to the CAP theorem, when a network partition happens, you need to make a trade-off between Consistency and Availability.
Network Partition
A network partition is a situation where communication between different nodes or components of the distributed system is either slow or completely lost. It might be due to network failures, latency, or other issues.
Centralized System
In a centralized systems(e.g., RDBMS), there is no concept of network partitions. All components of the system are within a single network, and communication between different parts of the system is assumed to be reliable and instantaneous. Hence, you effectively get both Availability and Consistency.
Distributed System
In a distributed system, network partitions can occur. This means that in the face of a network partition, you have to choose whether to prioritize consistency of data (making sure all nodes see the same data) or availability of the system (ensuring that every request receives a response without guaranteeing that it contains the most recent version of the data).
If you choose Consistency, it might mean sacrificing Availability during a partition.
If you choose Availability, it might mean sacrificing Consistency, allowing nodes to diverge in their views of the data during a partition.
In CAP terms:
- CA: Consistent and Available, but not Partition Tolerant (not suitable for distributed systems).
- CP: Consistent and Partition Tolerant, sacrificing Availability during a partition.
- AP: Available and Partition Tolerant, sacrificing Consistency during a partition.
Consistency Patterns: Consistency Models
Availability Patterns: Replication