Cloud Computing — Introduction & Fundamentals
1. IT Infrastructure
Definition: The combination of hardware, software, networks, and facilities used to develop, test, deliver, and manage IT services.
| Component | Description | Example |
| Hardware | Physical devices: servers, storage, networking | Dell Servers, Cisco Routers |
| Software | OS, apps, management tools | Windows Server, Linux, VMware |
| Networking | Switches, routers, firewalls | Cisco Switches, Juniper Firewalls |
| Data Storage | Enterprise data management solutions | SAN, NAS, Cloud Storage |
| Security | Protection against threats and unauthorized access | Firewalls, Antivirus, IAM |
| Cloud Services | Internet-based infra to reduce physical dependency | AWS, Azure, GCP |
2. Evolution Timeline
| Era | What Existed | Core Problem |
| 80s–2000s | Physical Data Centers | Owned, built, and managed everything yourself |
| 2000s | Virtualization | Fixed hardware waste and isolation issues |
| Post-Virtualization | Remote-managed VDCs | Still needed physical access for control |
| Today | Cloud Computing | Internet-based, on-demand, pay-as-you-go |
3. Problems with Physical Data Centers
| # | Problem | Why It Hurts |
| 1 | High upfront cost | Buy all servers before knowing actual need |
| 2 | Unequal resource needs | Cart (low CPU) and Payment (high CPU) given same machine |
| 3 | No resource sharing | Idle RAM/CPU on Server A cannot help Server B |
| 4 | Uncertain demand | Over-provision = waste; Under-provision = crash |
| 5 | One OS per server | Dependency conflicts (Python 3.12 vs 3.8 on same machine) |
4. Virtualization
Definition: Creating multiple Virtual Machines (VMs) on a single physical machine using a hypervisor.
Benefits
| Benefit | What It Solves |
| Better resource utilization | Multiple VMs on one server reduce hardware waste |
| Improved scalability | Create/remove VMs as needed |
| Increased fault tolerance | VMs can be migrated between servers, reducing downtime |
| Reduced costs | Fewer physical machines = lower operational expenses |
Types of Virtualization
| Type | What Gets Virtualized | Example Software |
| Server | Multiple virtual servers on one physical machine | VMware vSphere, Hyper-V, KVM |
| Desktop | Multiple OS environments on one computer | Citrix Virtual Apps, VMware Horizon |
| Storage | Abstracts physical storage from servers | NetApp ONTAP, IBM Spectrum Virtualize |
| Network | Virtual networks within a physical network | Cisco ACI, VMware NSX |
| Application | Apps run in isolated environments without full install | Microsoft App-V, Citrix XenApp |
Hypervisor (The Engine Behind Virtualization)
Definition: Software/firmware layer that creates and manages VMs.
| Type 1 (Bare Metal) | Type 2 (Hosted) |
| Sits on | Directly on hardware | On top of an OS |
| Stack | Hardware → Hypervisor → VMs | Hardware → OS → Hypervisor → VMs |
| Performance | ✅ High | ❌ Lower |
| Used in | Production / Cloud | Local dev / Testing |
| Examples | AWS Nitro, Xen, VMware ESXi, Hyper-V | VirtualBox, VMware Workstation, Parallels |
Bare Metal = Hardware with no OS installed. Type 1 hypervisors run directly on bare metal.
5. Birth of Cloud
Key innovations that pushed from virtualization → cloud:
- Advancements in Virtualization — Enabled better resource allocation
- Increased Internet Speeds — Allowed seamless access to remote resources
- Pay-As-You-Go Pricing — Eliminated upfront hardware investments
- Need for Global Scalability — Companies needed instantly scalable infrastructure
A management layer (internet-accessible console) was added on top of virtualized data centers → you can now create, delete, and manage resources from anywhere over the internet.
6. Cloud Computing — Official Definition (NIST SP 800-145)
"A model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction."
Simple version: Delivery of computing resources (servers, storage, networking, software) over the internet — on-demand, pay-as-you-go.
7. NIST 5 Essential Characteristics ⭐
| # | Characteristic | What It Means | Example |
| 1 | On-Demand Self-Service | Provision resources anytime, no human approval needed | Launch EC2 at 2 AM instantly |
| 2 | Broad Network Access | Accessible from any device over the internet | Laptop, mobile, anywhere |
| 3 | Resource Pooling (Multi-tenancy) | Provider serves multiple customers from shared hardware dynamically | You don't know which physical server you're on |
| 4 | Rapid Elasticity | Scale up/down instantly and automatically | 2 servers → 100 → back to 2 |
| 5 | Measured Service | Usage tracked and billed (CPU hours, storage, bandwidth) | Pay for exactly what you use |
8. CapEx vs OpEx ⭐
| CapEx (Capital Expenditure) | OpEx (Operational Expenditure) |
| What | Upfront investment in physical assets | Ongoing usage-based spending |
| Example | Buy servers for your data center | Pay AWS monthly bill |
| Risk | High — pay before you know demand | Low — pay as you grow |
| Traditional DC | ✅ CapEx-heavy | — |
| Cloud | — | ✅ OpEx model |
Cloud converts CapEx → OpEx. This is one of the core business reasons companies migrate to cloud.
9. Cloud Deployment Models
| Model | Who Uses It | Infra Owned By | Key Point |
| Public | Anyone | Provider (AWS, Azure, GCP) | No upfront cost, less control |
| Private | Single org only | Organization itself | Full control, high cost |
| Hybrid | One org | Both (private + public) | Sensitive data private, scale on public |
| Community | Group of orgs with shared needs | Shared / Provider | Hospitals, universities |
| Multi-Cloud | One org | Multiple public providers | AWS + Azure together |
Important: Data Center ≠ Private Cloud. Data Center + Cloud Software (e.g., OpenStack, VMware vSphere) = Private Cloud
Hybrid vs Multi-Cloud (Trap Question)
| Hybrid | Multi-Cloud |
| Private + Public combined | Multiple public clouds |
| Mixed model | Same type (all public) |
| Example: On-prem + AWS | Example: AWS + Azure |
10. Cloud Service Models ⭐
| Model | Provider Manages | You Manage | AWS Example |
| IaaS | Hardware, networking, virtualization | OS, runtime, apps, data | EC2 |
| PaaS | Infra + OS + runtime | Only your app/code | Elastic Beanstalk, EKS (managed) |
| SaaS | Everything | Just usage | Gmail, Dropbox, Zoom |
| FaaS | Infra + OS + runtime + scaling | Only function code | AWS Lambda |
IaaS → PaaS → SaaS → FaaS
Less you manage, but also less control.
11. Common Confusions Resolved ✅
| ❌ Wrong Belief | ✅ Correct Answer |
| Virtualization = Cloud | Virtualization is a building block of cloud, not cloud itself |
| Data Center = Cloud | DC + cloud software stack = Private Cloud |
| Docker = PaaS | Docker is a containerization tool, not a service model |
| Raw Kubernetes = PaaS | Raw K8s = IaaS-like; Managed K8s (EKS, GKE) ≈ PaaS |
| Netlify/Vercel = SaaS | They are PaaS — you deploy code, they manage infra |
| Custom software for a client = SaaS | That is software development; SaaS = one product, many users |
| Type 2 hypervisor used in cloud | Type 1 (bare metal) is used in production cloud |
12. Interview Questions Checklist ✅