CAP Theorem

▶

The CAP Theorem is a fundamental principle in distributed systems that states:

A distributed system cannot simultaneously guarantee all three properties: Consistency (C), Availability (A), and Partition Tolerance (P). It can guarantee only two of these three properties at any given time.

When designing a distributed system, architects must decide which two properties are most important based on the application's requirements.

The Three Properties of CAP Theorem

Understanding the Three Properties

1. Consistency (C)

Consistency ensures that every client always sees the most recent version of the data.

Once a write operation is completed, every subsequent read returns the updated value.

Example

Suppose a user's account balance is updated from:

₹10,000
↓
₹12,000

Every user reading the balance should immediately see:

₹12,000

No user should receive the old value.

2. Availability (A)

Availability ensures that every request receives a response, regardless of the current state of the system.

The response may contain:

• Latest data
• Older (stale) data
• An empty result

However, the system must always respond.

Example

Client Request
↓
Server Responds

✓ Always

Even if one server fails, another server should continue serving requests.

3. Partition Tolerance (P)

Partition Tolerance means the system continues functioning even when communication between distributed nodes is interrupted.

Network failures should not completely stop the system.

Example

Node A

──── Network Failure ────

Node B

Despite the communication failure, both nodes continue operating independently.

Why Can't We Have All Three?

In a distributed system, data is stored across multiple nodes.

Whenever data changes:

1. One node updates its data.
2. The updated data must be synchronized with other nodes.
3. Synchronization requires time.

During this synchronization period, the system must make a decision.

• If we show the data (even if outdated) → Availability is maintained, but Consistency is compromised.
• If we block the request until sync completes → Consistency is maintained, but Availability is compromised.

This unavoidable trade-off forms the basis of the CAP Theorem.

Single Database
↓
Always Available
↓
Always Consistent

User Request
↓
Wait Until Data Sync Completes
↓
Latest Data Returned

User Request
↓
Immediate Response
↓
Latest Data Synchronizes Later

Advantages of Understanding CAP Theorem

• Helps design reliable distributed systems.
• Clarifies the trade-offs between consistency and availability.
• Guides database and architecture decisions.
• Improves fault tolerance planning.
• Enables systems to meet business requirements effectively.

Summary

• CAP Theorem states that a distributed system can guarantee only two out of three properties at a time: Consistency (C), Availability (A), and Partition Tolerance (P).
• During network partitions or communication failures, systems must choose between maintaining Consistency (always returning the latest data) or Availability (always responding to requests).
• CP systems prioritize data accuracy and consistency over availability, making them suitable for critical applications such as banking and financial systems, while AP systems prioritize availability and fault tolerance, even if some data becomes temporarily stale.
• The CAP Theorem highlights the fundamental trade-offs in distributed system design, requiring architects to select the most appropriate balance based on business and application requirements.

Written By: Muskan Garg