Scaling and Elasticity in Cloud Computing

Overview

Elasticity is the capability to automatically adjust compute resources to match demand.
This includes:

• Scaling up/down: More or fewer resources per instance.
• Scaling out/in: More or fewer instances serving traffic.

Effective elasticity improves:

• Performance
• Cost efficiency
• Resilience

Scaling Types

Vertical Scaling

• Definition: Adjusting the resources of a single instance (CPU, RAM, instance size/type).

• Terminology:

  • Scale up: Add more resources.
  • Scale down: Reduce resources.

• Notes:

  • Typically incurs downtime (instance must be stopped/restarted or replaced).
  • Limited by the largest available instance type.

Horizontal Scaling

• Definition: Adjusting the number of instances serving traffic.

• Terminology:

  • Scale out: Add instances.
  • Scale in: Remove instances.

• Notes:

  • Can be performed without downtime when combined with a load balancer and health checks.
  • Best suited for stateless or externally state-managed services.