CAP Theorem

PublishedMarch 25, 2024

•2 min read•View as Markdown

Tiger Abrodi

Just a guy who loves to write code and watch anime.

Comments

Join the discussion

No comments yet. Be the first to comment.

More from this blog

Become a cracked product engineer today

Qualities that make outstanding builders.

Jun 1, 20264 min read

Bipedal, humanoid, and the words for creature shapes in games

Bipedal vs quadrupedal: how many legs This is about the number of legs the creature walks on. Bipedal. Two legs. Humans, ostriches, T-rex, kangaroos, most fantasy humanoids. Quadrupedal. Four legs. Do

May 18, 20263 min read

Bipedal, humanoid, and the words for creature shapes in games

Lerp and smoothstep, what they actually do

Intro You hear "lerp" and "smoothstep" everywhere in game dev. They sound like math jargon. They're not. Both are small tools that do the same job: smoothly move from one value to another. The problem

May 18, 20264 min read

Lerp and smoothstep, what they actually do

IK and FK, what they actually are

What is Kinematics Kinematics is the math field. It's the study of how things move without worrying about forces (which would be dynamics). FK and IK are the two branches: forward kinematics and inver

May 10, 20267 min read

How to make your textures fast on the GPU

Intro Textures are usually the biggest cost in a 3D scene. Memory, bandwidth, and load time all get eaten by them. Resizing them is the obvious lever. There's more. This post is about the less obvious

May 10, 20268 min read

How to make your textures fast on the GPU

Tiger's Place

497 posts

Introduction

CAP theorem is a concept that a distributed system can only have two of the following three properties:

Consistency: All nodes see the same data at the same time.

Availability: Every request gets a response on success/failure.

Partition Tolerance: The system continues to operate despite network partitions.

It's important to mention that the CAP theorem doesn't apply to single-node databases. It only applies to distributed systems, systems that run on multiple nodes.

It doesn't apply to single-node databases because they don't have to deal with network partitions. With a single-node database, you can have consistency and availability simultaneously.

What can we pick?

You may think we can pick any two of the three properties, but that's not the case.

Partition tolerance is a must in a distributed system because network partitions are inevitable. By network partition, I mean that some nodes can't communicate with other nodes. Imagine a system where we have a leader node and some follower nodes. If the leader node can't communicate with the followers, we have a network partition. So, you can't pick Consistency and Availability (CA).

This means you can choose either Consistency and Partition Tolerance (CP) or Availability and Partition Tolerance (AP).

Consistency

Consistency means that every single read on any node returns the most recent write. This is the strongest consistency model. If you have a consistent system, you can be sure that all nodes see the same data simultaneously.

The upside of consistency is that your users will always see the most recent data. The downside is that it can't serve requests if a node can't communicate with the other nodes. This means that if you have a consistent system, you can't be available when there is a network partition (a node can't communicate with the other nodes).

Availability

Availability means that every request gets a response on success/failure. This is the strongest availability model. If you have an available system, you can ensure that every request receives a response.

The upside of availability is that your system can serve requests even when there is a network partition. The downside is that your users may see stale data. This is because if a node can't communicate with the other nodes, it can't be sure it has the most recent data.

Consistency or Availability?

If the system must always return the most recent data, such as in banking systems, you should pick Consistency over Availability. If the system must always be available, such as in social networks, you should pick Availability over Consistency.

It's different for every system, and you must decide what is more important for your system.

Command Palette