# Chapter 7: The Cloud Scales ![The cloud scales](Postgres/assets/chap_7_cloud_scales.png) For a long time, the [[Postgres/Chapter 1 - The Building Blocks|Lazy Elephant]] lived in a single warehouse. He had one desk, one filing cabinet, and one diary. If the warehouse got too crowded, the elephant simply got overwhelmed. But then, the world changed. The horizon expanded. The elephant realized that if he could master the art of **[[Postgres/Chapter 4/4.1 - The Pocket Diary (WAL & fsync)|Communication]]**, he could effectively be in many places at once. In this chapter, we explore how the foundational safety of the **Pocket Diary (WAL)** and the shouting of the **Town Crier (Pub/Sub)** allow Postgres to transcend the physical limits of a single machine. We will look at how an army of **Read Replicas** can share the burden of work, and how modern **Cloud Architectures** (like Aurora and Neon) divorce the elephant from his filing cabinet entirely, shipping his thoughts into an infinite, glowing archive in the sky. This is the ultimate evolution of laziness: a warehouse that never fills up, an elephant that never loses its memory, and a world where scaling is just a matter of shouting a little louder. Giant tech companies have realized that if you intercept the elephant's frantic diary scribblings (the WAL) and send them across the network instead of to a local disk, you can perform miracles. [[Postgres/Architecture/Cluster/Aurora|Amazon Aurora]] scatters those scribbles across 6 different storage nodes across 3 availability zones for unkillable resilience. [[Postgres/Architecture/Cluster/AlloyDB|Google AlloyDB]] takes those scribbles and magically transforms them into lightning-fast, columnar analytical nodes so the elephant can answer huge math questions instantly. And [[Postgres/Architecture/Cluster/Neon]] intercepts the scribbles to allow instant, serverless branching, letting you clone the mammoth database in seconds just like git branching code. The elephant has successfully reached nirvana. It reclines gracefully on a cloud, processing queries and making plans, while a sprawling army of invisible network storage nodes violently handles the frozen, heavy labor far below.