Module 4 - Scaling

Vertical vs Horizontal Scaling

Two fundamental approaches to handling more load. Know when to use each.

1The Building Analogy

Simple Analogy
Vertical Scaling (Scale Up): Make your building taller-add more floors to handle more people.

Horizontal Scaling (Scale Out): Build more buildings across the city-spread people across multiple locations.

Eventually, you can't make a building infinitely tall (hardware limits). But you can always build more buildings (add servers).

2Comparison

AspectVerticalHorizontal
ApproachBigger machineMore machines
CostExponentialLinear
LimitHardware ceilingTheoretically unlimited
DowntimeRequired for upgradeZero downtime
ComplexitySimpleDistributed systems

3Vertical Scaling

Vertical scaling means upgrading your existing machine: more CPU, RAM, faster storage.

Pros

  • ✓Simple-no code changes
  • ✓No distributed complexity
  • ✓Good for databases

Cons

  • ✗Hardware limits
  • ✗Single point of failure
  • ✗Expensive at scale

4Horizontal Scaling

Horizontal scaling means adding more machines and distributing load across them.

Pros

  • ✓No hardware ceiling
  • ✓Fault tolerant
  • ✓Cost-effective at scale

Cons

  • ✗Distributed complexity
  • ✗Data consistency challenges
  • ✗Requires load balancing

5When to Use Each

Use Vertical

  • • Early-stage startups
  • • Database servers (initially)
  • • Quick fix for immediate load
  • • Simple stateful apps

Use Horizontal

  • • High-traffic web apps
  • • Stateless services
  • • Need for high availability
  • • Long-term scaling strategy

6Key Takeaways

1Vertical = bigger machine, Horizontal = more machines
2Vertical is simpler but has hardware limits
3Horizontal enables unlimited scaling but adds complexity
4Most systems use both-scale up then scale out
5Stateless services are easier to scale horizontally

?Quiz

1. Your single server is at 90% CPU. Quick fix?

2. Web app needs to handle 10x traffic long-term. Best approach?