Cloud Computing: The Operating Model of Modern IT
Cloud computing is the on-demand delivery of compute, storage, networking, platforms, and software over the internet with measurable, elastic usage and shared infrastructure. It is less a technology and more an operating model: automate everything, scale instantly, and pay only for what you consume.
NIST Characteristics
Self-service, broad access, resource pooling, rapid elasticity, measured service.
Why It Matters
Transforms CapEx-heavy infrastructure into flexible OpEx with automation and global reach.
Design Mindset
Everything is an API. Everything is a service. Everything is automatable.
Essential Cloud Behaviors
Provision resources through APIs, auto-scale based on demand, and expose consumption metrics for governance and cost control.
Shared Responsibility
Providers secure the cloud infrastructure; customers secure their data, identities, and configurations.
| Layer | Provider Manages | Customer Manages |
|---|---|---|
| SaaS | App, runtime, OS, virtualization, hardware | Data, access policies |
| PaaS | Runtime, OS, virtualization, hardware | App code, data, access policies |
| IaaS | Virtualization, hardware, facilities | OS, runtime, apps, data, access policies |
1.2 Roots of Cloud Computing
Why History Matters
Every wave of computing reduces the cost of resource sharing. Cloud is the culmination: automation and elasticity at global scale.
1.3 Cloud Service Models (SPI)
IaaS
Provision raw compute, storage, and network resources. Maximum control, maximum responsibility.
Examples: EC2, Compute Engine, Azure VMs
PaaS
Managed runtimes for code deployment without managing OS or runtime patches.
Examples: App Engine, Azure App Service, Heroku
SaaS
Full applications delivered over the web. Ideal for rapid productivity.
Examples: Office 365, Salesforce, Slack
Deployment Blueprint
Architects map workloads to service models based on data sensitivity, operational maturity, and speed-to-market requirements.
1.4 Cloud Layers & Deployment Types
Public Cloud
Shared infrastructure operated by a provider. Best for elasticity and global reach.
Private Cloud
Single-tenant environment with high governance and regulatory control.
Hybrid Cloud
Mix of on-prem and cloud with portability, data residency, and burst capacity.
Community Cloud
Shared by organizations with common compliance or security requirements.
Reference Architecture Layers
1.5 - 1.7 Features, Benefits, and Applications
Desired Features
High availability, automated recovery, secure multi-tenancy, and policy-driven governance.
Benefits vs Challenges
Benefits: agility, elasticity, global scale. Challenges: cost sprawl, shared responsibility, vendor lock-in.
Exam Focus: Cloud Economics
Know when to choose OpEx over CapEx, how elasticity saves cost, and how reserved capacity lowers steady-state spend.
Typical Cloud Applications
Disaster recovery, streaming media, analytics pipelines, Dev/Test environments, and global e-commerce platforms.
Virtualization Fundamentals
2.1 Introduction to Virtualization
Virtualization abstracts physical resources into logical units. It allows consolidation, isolation, and rapid provisioning while maximizing hardware utilization.
2.2 Core Characteristics
- Partitioning: multiple OS instances on one host.
- Isolation: one VM cannot impact another.
- Encapsulation: VM state stored as files.
- Hardware Independence: move workloads freely.
Virtualization vs Containerization
VMs emulate hardware; containers share the host kernel. Containers are lighter and faster but rely on OS-level isolation.
2.3 Virtualization Methods
| Method | Mechanism | Typical Use |
|---|---|---|
| Full Virtualization | Binary translation, unmodified guest OS | Legacy workloads |
| Para-virtualization | Hypercalls, modified guest OS | High-performance workloads |
| OS-level Virtualization | Namespaces and cgroups | Containers, microservices |
Performance Insight
CPU virtualization is efficient; I/O virtualization is the bottleneck. Use SR-IOV or paravirtual drivers for high throughput.
2.4 Types of Virtualization
Compute Virtualization
VMs and containers delivering logical compute. Supports live migration and autoscaling.
Network Virtualization
SDN and overlay networks (VXLAN) decouple network services from hardware.
Storage Virtualization
Aggregates multiple disks into logical volumes. Enables thin provisioning and snapshots.
Desktop/Application
VDI and app containers provide secure, centralized application delivery.
2.5 Hypervisors and VM Lifecycle
Type 1 Hypervisor
Runs directly on hardware. Preferred for data centers and cloud providers.
Examples: ESXi, Hyper-V, KVM
Type 2 Hypervisor
Runs on top of a host OS. Common for development and testing environments.
Examples: VirtualBox, VMware Workstation
VM Lifecycle
2.6 Clustering and Automation
Cluster Types
High-availability clusters provide failover; load-balancing clusters distribute traffic; HPC clusters maximize throughput.
Auto-scaling Triggers
Scale based on CPU, queue depth, response latency, or custom business metrics.
Infrastructure as Code
Declarative templates (Terraform, CloudFormation) make infrastructure versioned, repeatable, and auditable.