The CAP theorem in HLD - Architecture Diagram

System Overview - The CAP theorem

The CAP theorem explains the trade-offs in distributed systems between Consistency, Availability, and Partition tolerance. It states that a system can only guarantee two of these three properties at the same time. This helps architects decide which properties to prioritize based on system needs.

Architecture Diagram

          +-------------------+          +-------------------+
          |    Client A       |          |    Client B       |
          +---------+---------+          +---------+---------+
                    |                              |
                    |                              |
          +---------v---------+          +---------v---------+
          |   Load Balancer   |          |   Load Balancer   |
          +---------+---------+          +---------+---------+
                    |                              |
                    +--------------+---------------+
                                   |
                         +---------v---------+
                         |   Distributed     |
                         |   Database Cluster|
                         +---------+---------+
                                   |
                  +----------------+----------------+
                  |                                 |
          +-------v-------+                 +-------v-------+
          | Partition 1   |                 | Partition 2   |
          | (Subset of DB)|                 | (Subset of DB)|
          +---------------+                 +---------------+

Components

Client A

client

User or application sending requests to the system

Client B

client

Another user or application sending requests to the system

Load Balancer

load_balancer

Distributes incoming requests evenly to database nodes

Distributed Database Cluster

database_cluster

Stores data across multiple nodes to provide redundancy and scalability

Partition 1

database_node

One part of the database cluster holding a subset of data

Partition 2

database_node

Another part of the database cluster holding a subset of data

Request Flow - 6 Hops

Client A → Load Balancer

Load Balancer → Distributed Database Cluster

Distributed Database Cluster → Partition 1 or Partition 2

Partition 1 or Partition 2 → Distributed Database Cluster

Distributed Database Cluster → Load Balancer

Load Balancer → Client A

Failure Scenario

Component Fails:Network Partition between Partition 1 and Partition 2

Impact:Partitions cannot communicate, causing potential inconsistency or unavailability depending on system choice

Mitigation:System must choose to prioritize either Consistency (reject some requests) or Availability (serve stale data) during partition

Architecture Quiz - 3 Questions

According to the CAP theorem, which of the following is impossible for a distributed system to guarantee simultaneously?

A All three: Consistency, Availability, and Partition tolerance

B Consistency and Availability

C Consistency and Partition tolerance

D Availability and Partition tolerance

Design Principle

The CAP theorem guides system architects to understand that in distributed systems, network failures force a choice between consistency and availability. Designing systems requires prioritizing which two properties best fit the application's needs.