Understand the core concepts, models, and services that power modern cloud infrastructure.
Cloud computing offers three primary service models that abstract different layers of infrastructure. Understanding where responsibility lies in each model is critical for IT support professionals.
Infrastructure as a Service provides fundamental computing resources: servers, storage, and networking on demand.
You manage: OS, runtime, data, applications
Provider manages: Physical hardware, virtualization, networking, storage
Examples: Amazon EC2, Azure Virtual Machines, Google Compute Engine
Best for: Organizations needing full control over their infrastructure stack without maintaining physical hardware.
Platform as a Service provides a managed environment for developing, running, and managing applications without infrastructure concerns.
You manage: Data, applications, runtime configuration
Provider manages: OS, middleware, runtime, servers, storage
Examples: Azure App Service, AWS Elastic Beanstalk, Google App Engine
Best for: Developers who want to focus on code without worrying about server maintenance.
Software as a Service delivers fully managed applications over the internet on a subscription basis.
You manage: Your data and user access settings only
Provider manages: Everything -- infrastructure, platform, application, updates, security patches
Examples: Microsoft 365, Google Workspace, Salesforce, Dropbox Business
Best for: End-users and organizations wanting ready-to-use software with zero infrastructure management.
IaaS = You order ingredients and bake the pizza yourself (you get the kitchen). PaaS = You order a homemade-style pizza from a delivery service (they provide the oven). SaaS = You go to a restaurant and eat someone else's pizza (everything is handled for you).
Organizations choose deployment models based on their security requirements, regulatory needs, cost constraints, and operational flexibility preferences.
| Model | Description | Best For | Cost | Control Level |
|---|---|---|---|---|
| Public Cloud | Resources owned and operated by a third-party provider, delivered over the public internet | Startups, teams needing quick scaling, non-sensitive workloads | Lowest Pay-as-you-go | Low to medium |
| Private Cloud | Exclusive cloud resources used by a single organization, hosted on-premises or by a third party | Government, healthcare, finance -- regulated industries | Higher Capital investment | High |
| Hybrid Cloud | Combination of public and private clouds, allowing data and applications to share between them | Organizations with mixed workloads and compliance requirements | Medium-High Flexible mix | Medium to high |
| Multi-Cloud | Using services from multiple cloud providers to avoid vendor lock-in and optimize capabilities | Large enterprises seeking best-of-breed services across providers | Variable Provider-dependent | Medium |
Q1: Do you have strict data residency or compliance requirements? → Yes = Private Cloud | No = continue
Q2: Do you need to handle variable traffic spikes? → Yes = Public Cloud | No = continue
Q3: Do you have sensitive workloads AND need burst capacity? → Yes = Hybrid Cloud | No = Multi-Cloud
Market position: ~31% market share, the largest cloud provider globally
Key strengths:
Core compute: EC2, Lambda, ECS, EKS
Core storage: S3, EBS, EFS, Glacier
Market position: ~25% market share, second largest, fastest-growing in enterprise
Key strengths:
Core compute: Virtual Machines, Azure Functions, AKS
Core storage: Blob Storage, File Shares, Disk Storage
Market position: ~11% market share, strong in AI/ML and data analytics
Key strengths:
Core compute: Compute Engine, Cloud Run, GKE
Core storage: Cloud Storage, Persistent Disks, Cloud SQL
As an IT support professional, you will most commonly encounter Azure (due to Microsoft ecosystem dominance in enterprises) and AWS (most common general-purpose cloud). Focus your learning on these two first, then explore GCP if needed for your organization.
Cloud pricing operates on an operational expenditure (OpEx) model rather than capital expenditure (CapEx). Understanding cost structures helps you advocate for budget decisions.
| Pricing Model | How It Works | Best Use Case | Risk |
|---|---|---|---|
| On-Demand | Pay per second/hour for resources as consumed, no commitment | Short-term workloads, testing, unpredictable traffic | Highest unit cost |
| Reserved Instances | Commit to 1-3 years for significant discounts (up to 72%) | Steady-state workloads with predictable usage | Obligation even if unused |
| Spot Instances | Bid on unused capacity at steep discounts (up to 90% off) | Fault-tolerant, interruptible workloads (batch processing) | Can be reclaimed with short notice |
| Savings Plans | Commit to consistent spend ($/hour) for 1-3 years | Flexible commitment across multiple service types | Must meet minimum spend |
| Free Tier | Limited free resources permanently or for 12 months | Learning, experimentation, small projects | Usage limits apply strictly |
Always check for active instances before shutting down a VM. A common rookie mistake is terminating a running production instance. Always verify resource ownership and labels before making changes in any cloud console.
The 6 Rs framework (introduced by AWS) helps organizations decide how to migrate workloads from on-premises to the cloud.
| Strategy | What Happens | Effort | Business Value | Example |
|---|---|---|---|---|
| Rehost (Lift & Shift) | Move applications as-is to cloud VMs | Low | Moderate | Migrate on-prem VM to EC2/Azure VM |
| Replatform | Minor optimizations during migration (e.g., switch DB engine) | Medium | Moderate-High | Migrate MySQL to Amazon RDS |
| Repurchase (Drop & Replace) | Replace existing app with a cloud SaaS alternative | Medium | High (new features) | Switch from on-prem Exchange to Microsoft 365 |
| Refactor (Re-architect) | Rewrite application to use cloud-native features | High | Highest | Convert monolith to microservices on Kubernetes |
| Retain (Hold) | Keep workloads on-premises (not ready to migrate) | None | N/A | Legacy system with compliance restrictions |
| Retire | Decommission unused or redundant applications | Low | Cost savings | Remove duplicate systems identified in audit |
Every cloud provider uses a shared responsibility model. Understanding what YOU are responsible for versus what the PROVIDER handles is essential for security compliance.
| Responsibility Area | Provider Handles | You (Customer) Handle |
|---|---|---|
| Physical security | Data centers, hardware, environmental controls | -- |
| Infrastructure | Compute, storage, networking hardware | -- |
| Virtualization layer | Hypervisor, guest OS isolation | -- |
| Guest operating system | -- | Patching, configuration, hardening |
| Data protection | -- | Encryption at rest, backups, classification |
| Identity & Access | IAM platform availability | User policies, MFA, permissions |
| Application security | -- | Secure coding, vulnerability management |
| Network firewall | Physical/network perimeter | Security groups, NACLs, WAF rules |
In IaaS, you manage the most (OS, apps, data). In PaaS, the provider manages more (OS, runtime). In SaaS, the provider manages almost everything except your data and user access. As you move up the service model ladder, your security responsibility decreases but your data governance responsibility remains constant.
The ability to handle increased load by adding resources. Can be vertical (scale up) or horizontal (scale out).
Vertical scaling (Scale Up): Adding more CPU/RAM to an existing VM. Limited by maximum instance size.
Horizontal scaling (Scale Out): Adding more VM instances behind a load balancer. More flexible and resilient.
When to choose: Vertical for databases and single-instance apps. Horizontal for web services and stateless applications.
The ability to automatically scale resources UP and DOWN based on real-time demand. This dynamic adjustment is what makes cloud economically attractive.
Auto-scaling example: An e-commerce site automatically adds 10 servers during Black Friday and removes them at midnight.
Key benefit: You only pay for what you actually use at any given moment.
When to choose: Any workload with variable or unpredictable traffic patterns.
Redundancy means having backup components ready to take over if primary ones fail. High availability ensures services remain accessible with minimal downtime.
Physically separate data centers within a cloud region. Each AZ has independent power, cooling, and networking. Distributing instances across multiple AZs protects against single-point failures.
A region is a geographic area containing multiple AZs. For disaster recovery, replicate data to a secondary region. Cross-region replication ensures business continuity during regional outages.
Get practical experience by setting up a free account and launching your first cloud resource.
my-first-vm, select Amazon Linux 2023 AMI (free-tier eligible)t2.micro or t3.micro instance type (both free-tier eligible)ssh -i "your-key.pem" ec2-user@your-public-ipsudo dnf update -y and hostname -f inside the terminalNow that you understand cloud fundamentals, move on to hands-on exploration of Azure and AWS. The next module (Azure & AWS Basics for IT Support) will guide you through actual portal navigation, VM creation, and storage configuration in both platforms. After completing that, proceed to PowerShell scripting for Windows automation.