Service Models, Elasticity & Deployment Types

The Core Shift

On-demand computing resources delivered over the internet from a provider's data centers, instead of an organization owning/maintaining physical hardware — the fundamental characteristic distinguishing cloud from on-premises infrastructure.

IaaS / PaaS / SaaS

IaaS (AWS EC2) — raw infrastructure you configure yourself. PaaS (Heroku, App Engine) — a managed platform for deploying apps, no infrastructure management. SaaS (Gmail, Salesforce) — a complete, ready-to-use application. Each step up trades control for less management responsibility.

Public / Private / Hybrid

Public cloud: shared, provider-owned (AWS/Azure/GCP). Private cloud: dedicated to one organization. Hybrid: combines both, often keeping sensitive data on private infrastructure while using public cloud elasticity elsewhere for regulatory/compliance reasons.

Elasticity & Scaling Strategy

Auto-scaling adjusts resources to actual demand — scale up for a traffic spike, back down when quiet — without permanently provisioning for peak load. Horizontal scaling (more instances) is what cloud elasticity particularly enables; vertical scaling (a bigger single instance) eventually hits a hard ceiling.

Regions & Availability Zones

A region is a broad geographic area; an availability zone is an isolated data center within it. Spreading resources across zones is standard practice for fault tolerance — one zone's outage shouldn't take the whole app down.