Horizontal vs. Vertical Scaling

One-Liner

The two primary strategies for increasing a system’s capacity: adding more resources to an existing machine (vertical) or adding more machines (horizontal).

What It Is

Vertical Scaling (Scaling Up): Increasing the capacity of a single server by adding more CPU, RAM, or faster storage.
Horizontal Scaling (Scaling Out): Increasing capacity by adding more servers to a distributed system.

Why It Exists

To provide different approaches to handle increased load and demand on a system. Each has its own benefits and limitations.

How It Works

Vertical: Upgrade the hardware of a single machine.
Horizontal: Distribute the workload across multiple machines, often using a load balancer.

Tradeoffs

Vertical Scaling

Pros: Simpler to implement (no distributed system complexities).
Cons: Limited by the maximum capacity of a single machine, often more expensive per unit of capacity at higher levels, introduces a single point of failure.

Horizontal Scaling

Pros: Virtually limitless scalability, high availability (if one server fails, others can take over), often cheaper at scale (using commodity hardware).
Cons: Significantly increases system complexity (distributed consensus, data consistency, load balancing, service discovery), requires careful design for statelessness.

Failure Modes

Vertical Scaling: Hardware failure means the entire system goes down.
Horizontal Scaling: Complexities of distributed systems can lead to new types of failures (e.g., network partitions, data inconsistencies).

Interview Traps

Not being able to explain the core difference.
Suggesting vertical scaling as a long-term solution for high-growth systems.

Real-World Usage

Most modern web applications and microservices architectures rely heavily on horizontal scaling.

Anti-Patterns

Trying to vertically scale a service indefinitely, leading to a massive, expensive, and fragile single server.

Load Balancing
Statelessness
Database Sharding
Service Discovery