On-demand access to compute, storage, and networking resources over the internet — pay only for what you use.
Five Key Characteristics
Characteristic
Meaning
On-demand self-service
Provision instantly, no human interaction
Broad network access
Available from anywhere
Resource pooling
Shared infrastructure (multi-tenant)
Rapid elasticity
Scale up/down automatically
Measured service
Metered billing — pay per use
Cloud Deployment Models
PUBLIC PRIVATE HYBRID
────────────── ────────────── ──────────────────────
Azure (cloud) Your datacenter Both, connected
Anyone can use Internal only Flexible split
Least control Full control Mixed control
Cheapest Most expensive Middle cost
When to Use Each
Model
Use When
Public
Normal workloads, variable scale, no special compliance
Private
Regulatory, sensitive data, full control required
Hybrid
Some data must stay on-premises, burst to cloud for scale
Exam: Hybrid cloud is for organisations that can’t fully move to public cloud (compliance, legacy systems).
Cloud Service Models
YOU MANAGE ←────────────────────────────────→ PROVIDER MANAGES
SaaS PaaS IaaS
(Use it) (Build on it) (Rent the hardware)
────────── ───────────── ──────────────────
Office 365 App Service Azure VMs
Dynamics SQL Database Azure Storage
Teams Functions Virtual Networks
Shared Responsibility
Layer
IaaS
PaaS
SaaS
Application
You
You
Provider
Data
You
You
Provider
Runtime
You
Provider
Provider
OS
You
Provider
Provider
Virtualisation
Provider
Provider
Provider
Hardware / Physical
Provider
Provider
Provider
Exam: With IaaS you’re responsible for OS patches. With PaaS and SaaS you’re not.
CapEx vs OpEx
CapEx (Capital Expenditure)
OpEx (Operating Expenditure)
What
Buy hardware upfront
Pay monthly/hourly
Example
On-premises server
Azure subscription
Cloud model
Private cloud (you own hardware)
Public cloud
Tax treatment
Depreciated over time
Deducted immediately
Exam: Cloud = OpEx. On-premises = CapEx.
Key Benefits of Cloud
Benefit
Meaning
High Availability
Stays up despite failures (SLA guarantees)
Scalability
Add resources as demand grows
Elasticity
Auto-scale up AND down with demand
Agility
Deploy in minutes, not months
Geo-distribution
Data centres worldwide
Disaster recovery
Backup and failover built in
Scalability vs Elasticity:
Scalability = can grow (add more VMs over time)
Elasticity = automatically grows and shrinks in real time
Fault Tolerance & Disaster Recovery
Term
Meaning
Fault tolerance
System keeps running when a component fails
High Availability
Minimal downtime — measured by SLA (e.g. 99.95%)
Disaster Recovery
Recover from major outage
RTO
Recovery Time Objective — max acceptable downtime
RPO
Recovery Point Objective — max acceptable data loss