Understanding CAP Theorem, Consistency Levels, and BASE Theory in Distributed Systems

Ensuring data consistency without sacrificing system performance is a core challenge for any distributed system, which leads to the definition of consistency levels.

1. Strong Consistency – The system returns exactly what was written; it offers the best user experience but often incurs high performance costs.

2. Weak Consistency – After a write, the system does not guarantee an immediate read of the new value; it only promises that consistency will be achieved within a certain time window (e.g., seconds).

3. Eventual Consistency – A special case of weak consistency where the system guarantees that, after some time, all replicas will converge to the same state. This model is widely adopted in large‑scale distributed systems.

What is the CAP Theorem?

In 2000, Professor Eric Brewer proposed the CAP conjecture, which was formally proved by Seth Gilbert and Nancy Lynch two years later. The theorem states that a distributed system cannot simultaneously guarantee Consistency (C), Availability (A), and Partition tolerance (P); at most two of these can be satisfied.

C – Consistency : All replicas see the same data at the same time (strict consistency).

A – Availability : Every request receives a non‑error response, though the data may be stale.

P – Partition tolerance : The system continues to operate despite network partitions.

A partition occurs when nodes are split into isolated subnetworks that cannot communicate, effectively dividing the system.

Why Only Two of the Three Can Be Achieved

Consider a two‑node system where node A must update node B atomically. If a network partition separates A and B, the system cannot simultaneously maintain consistency, availability, and partition tolerance. It must sacrifice one property:

CA – Consistency and Availability, giving up Partition tolerance.

CP – Consistency and Partition tolerance, giving up Availability.

AP – Availability and Partition tolerance, giving up Consistency.

Can the 3‑choose‑2 Dilemma Be Resolved?

While the CAP theorem is proven, practical systems can mitigate its impact. If partitions are rare, a system may aim for all three properties most of the time. When partitions occur, strategies such as relaxing consistency to eventual consistency can preserve availability.

What Is the BASE Theory?

BASE stands for Basically Available, Soft state, Eventually consistent. It was introduced by eBay architects as a pragmatic response to CAP, emphasizing high availability and tolerating temporary inconsistency.

Basically Available – The system remains operational, possibly with degraded performance or functionality.

Soft state – Data may exist in intermediate states; the system tolerates temporary divergence among replicas.

Eventually consistent – After a period without new updates, all replicas converge to the same value.

Types of Eventual Consistency

In practice, eventual consistency manifests in several models:

1. Causal consistency – Operations that are causally related are seen in the same order.

2. Read‑your‑writes – A client always reads its own writes.

3. Session consistency – Guarantees read‑your‑writes within a single session.

4. Monotonic reads – Once a client has seen a value, it will never see an older one.

5. Monotonic writes – Writes from a single client are applied in order.

These models are often combined to build distributed systems that provide high availability while eventually reaching a consistent state, similar to how relational databases handle replication lag.

Conclusion

The CAP theorem shows that a distributed system cannot simultaneously satisfy consistency, availability, and partition tolerance; partition tolerance is mandatory, so a trade‑off between consistency and availability is required. The BASE theory offers a practical approach for large‑scale, highly available systems by sacrificing strong consistency in favor of eventual consistency.

Compiled by: Souyunku Technical Team