Architecture & System Design › System Design Fundamentals

Vertical Scaling

Scaling up with a bigger machine.

Also known as: scaling up, scale up

Vertical scaling means handling more load by giving one machine more resources: more CPU, more memory or faster disks. It’s also called scaling up. It needs no change to the application, which is why it’s the simplest first step.

Illustratively, a machine goes from a small size to a larger one, with the same code and the same architecture:

Before: small instance   -> resize the instance
After:  larger instance  (same code, same architecture)

The limits are practical ones. A single machine has a maximum size, the cost per unit of capacity often rises at the top end, and the machine remains a single point of failure. Some resizes also need a restart, which means downtime, so check what your provider requires.

The classic mistake is resizing before finding the bottleneck. If the slowdown is a missing database index or a query running in a loop, a bigger machine only hides it, and the cost keeps growing. Profile first, then decide. When one machine is no longer enough, horizontal scaling across several machines is usually the next step, which is covered by the load balancer concept.