Cloud Computing Basics: Definitions, Models, Services & Real-World Impact

Core Concepts: What is Cloud Computing?

Cloud computing is the delivery of on-demand computing resources (servers, storage, databases, networking, software, analytics, intelligence) over the internet to offer faster innovation, flexible resources, and economies of scale. Unlike traditional IT, where organizations purchase and maintain physical hardware, cloud computing leverages shared, virtualized resources managed by third-party providers.

How Cloud Computing Works

Cloud computing operates on three fundamental principles:

1. Resource Pooling: Multi-tenant model where physical resources (servers, storage, bandwidth) are dynamically allocated to multiple users.
2. Rapid Elasticity: Resources can be scaled up or down automatically based on demand (e.g., handling sudden traffic spikes).
3. Measured Service: Usage is monitored, controlled, and reported, providing transparency and billing efficiency.

Visual: Cloud Computing Architecture Layers

classDiagram
    class Frontend {
        +User Interface (UI)
        +Applications
        +Web/Mobile Clients
    }
    class Backend {
        +Servers
        +Storage
        +Virtualization Layer
        +Data Centers
    }
    class DeliveryModel {
        +SaaS (Software as a Service)
        +PaaS (Platform as a Service)
        +IaaS (Infrastructure as a Service)
    }
    class DeploymentModel {
        +Public Cloud
        +Private Cloud
        +Hybrid Cloud
        +Community Cloud
    }
    Frontend --> Backend : "Uses"
    Backend --> DeliveryModel : "Provides"
    Backend --> DeploymentModel : "Deployed as"

Service Models: How Cloud Provides Value

Cloud computing is categorized into three primary service models, each offering different levels of control and management:

classDiagram
    class SaaS {
        +Delivered via browser
        +No installation needed
        +Examples: eSewa, Google Workspace
    }
    class PaaS {
        +Develop/test/deploy apps
        +Managed runtime
        +Examples: Heroku, Azure App Service
    }
    class IaaS {
        +Virtualized resources
        +User-managed OS/apps
        +Examples: AWS EC2, Google Compute
    }
    SaaS --> PaaS : "Built on"
    PaaS --> IaaS : "Runs on"
    IaaS --> "Physical Hardware" : "Hosted by"
    note for SaaS "eSewa’s payment app is SaaS"
    note for IaaS "NTC’s cloud servers are IaaS"

Worked Example: eSewa’s Cloud-Based Service (SaaS)

Scenario: eSewa, Nepal’s leading digital payment platform, relies on SaaS to provide its services. Users access eSewa via a web or mobile app without needing to install or maintain any software. The backend infrastructure (servers, databases, security) is managed by cloud providers like AWS or Azure.

Key Cloud Components Used:

1. Frontend: Mobile/web apps for users.
2. Backend: Cloud-hosted servers running payment processing logic.
3. Database: Managed cloud databases (e.g., AWS RDS) storing transaction records.
4. Security: Cloud-based encryption and fraud detection (e.g., AWS KMS).

Why Cloud?

• Scalability: Handles millions of transactions during festivals (e.g., Dashain, Tihar).
• Reliability: 99.9% uptime guaranteed by cloud SLAs.
• Cost-Efficiency: No need to invest in physical data centers.

Deployment Models: Where Cloud Lives

Cloud environments can be deployed in four primary models, each suited to different organizational needs:

Visual: Hybrid Cloud Deployment (Ncell Example)

flowchart TD
    A["Ncell Customer Data"] -->|"Private Cloud"| B["On-Premise Servers"]
    B -->|"Secure Connection"| C["AWS Public Cloud"]
    C --> D["Scalable Services\n(e.g., Call Routing, SMS Gateway)"]
    E["Nepal Telecom Authority\n(Regulatory Compliance)"] -->|"Data Residency"| B

Real-World Tie-In: Ncell uses a hybrid cloud to balance security (private cloud for customer data) and scalability (public cloud for handling peak call volumes during events like elections or festivals). This ensures compliance with Nepal’s data localization laws while leveraging cloud elasticity.

Advantages and Disadvantages of Cloud Computing

Advantages

1. Cost Savings: No upfront capital expenditure on hardware; pay-as-you-go model.
2. Scalability: Instantly scale resources up or down (e.g., Daraz during sales events).
3. Reliability: High availability and disaster recovery (e.g., Google Cloud’s 99.95% uptime SLA).
4. Global Reach: Deploy applications closer to users via multi-region cloud deployments (e.g., AWS regions in Asia-Pacific).
5. Security: Cloud providers offer enterprise-grade security (e.g., AWS Shield, Azure Sentinel).
6. Collaboration: Real-time data sharing (e.g., Google Docs, Microsoft Teams).

Disadvantages

1. Dependency on Internet: Downtime or poor connectivity can disrupt services (e.g., Pathao app crashes during load shedding).
2. Vendor Lock-in: Difficulty migrating between providers due to proprietary formats.
3. Data Privacy Concerns: Storing sensitive data (e.g., bank records) in foreign clouds may violate laws (e.g., Nepal’s Electronic Transactions Act).
4. Limited Customization: Public cloud offerings may not meet niche requirements.
5. Compliance Risks: Some industries (e.g., healthcare, defense) require on-premise solutions.

Comparison: Cloud vs. Traditional Computing

In the Real World

Cloud computing is the backbone of modern digital services. Here’s how Nepalese and global companies leverage it:

1. eSewa (SaaS)

   • Idea Used: Multi-tenancy (shared infrastructure serving millions of users).
   • How: eSewa’s backend runs on AWS, allowing simultaneous transactions for bill payments, remittances, and government services without performance degradation.
   • Nepal Impact: Enabled cashless transactions during COVID-19 lockdowns.

2. Pathao (Hybrid Cloud + IaaS)

   • Idea Used: Auto-scaling and geolocation services.
   • How: Pathao uses Google Cloud’s Kubernetes Engine to dynamically scale ride-matching servers during peak hours (e.g., 7–9 PM in Kathmandu). Driver locations are processed via Google Maps API (cloud-based).
   • Nepal Impact: Reduced wait times for rides by 40% during festivals.

3. NTC (Public Cloud for Network Management)

   • Idea Used: IaaS for network monitoring.
   • How: Nepal Telecom Corporation (NTC) uses AWS IoT Core to monitor fiber-optic networks across Nepal. Sensors in exchange buildings send real-time data to cloud dashboards, enabling predictive maintenance.
   • Nepal Impact: Reduced fiber outages by 30% in remote areas like Darchula and Solukhumbu.

4. Daraz (PaaS for E-Commerce)

   • Idea Used: Serverless computing (AWS Lambda) for order processing.
   • How: During Daraz’s 11.11 Global Shopping Festival, the platform uses AWS Lambda to handle 10x the usual traffic without over-provisioning servers. Orders are processed in milliseconds.
   • Nepal Impact: Processed 500,000+ orders in a single day without crashes.

5. Nepal Rastra Bank (Private Cloud for Financial Data)

   • Idea Used: Private cloud for compliance.
   • How: NRB uses a private cloud hosted by IBM to store sensitive financial data (e.g., bank audits, forex records). This ensures data sovereignty (stored within Nepal) while benefiting from cloud security.
   • Nepal Impact: Complies with Banking Act 2020 and Electronic Transactions Act.

Cloud Computing vs. Traditional IT: A Worked Example

Scenario: A Kathmandu-based online tuition platform (e.g., Srijan Academy) needs to handle 10,000 concurrent students during exam seasons.

Traditional IT Approach

• Hardware: Purchase 50 physical servers (cost: NPR 50,00,000).
• Scaling: Manual addition of servers during peak times (slow, costly).
• Downtime: Risk of crashes if traffic spikes unexpectedly.
• Maintenance: Hire 5 IT staff for server management.

Cloud Computing Approach (AWS)

• Service Model: IaaS (EC2) + SaaS (Zoom for live classes).
• Scaling: Auto-scaling group adds/removes VMs based on demand (cost: ~NPR 15,00,000/month).
• Database: Amazon RDS for student records (auto-backups).
• CDN: CloudFront to deliver videos faster across Nepal.
• Cost: Pay only for actual usage (no idle server costs).

Outcome:

• 99.9% uptime during exams.
• No IT staff needed (AWS manages security, patches).
• Cost savings of 70% compared to traditional IT.

Exam Tip

This unit is heavily tested in TU exams through:

1. Definitions: Expect short-answer questions on terms like:
   • What is cloud computing?
   • Differentiate SaaS, PaaS, and IaaS.
   • Explain the hybrid cloud model with an example.
2. Comparisons: Table-based questions comparing cloud vs. traditional computing or public vs. private clouds.
3. Scenario-Based Questions: Case studies (like the eSewa or Pathao examples above) asking:
   • Which cloud service model does [X] use? Why?
   • How would you deploy [Y] in a hybrid cloud?
4. Advantages/Disadvantages: Descriptive questions requiring 3–5 points for each.
5. Diagrams: Draw and label cloud architecture layers or deployment models.

High-Scoring Strategy:

• Memorize the 3 service models and 4 deployment models with one Nepalese example each.
• Practice drawing the cloud architecture diagram (frontend/backend/delivery/deployment).
• Relate every answer to Nepal (e.g., "NTC uses public cloud for...").
• Use bullet points for comparisons (examiners love structured answers).

IMAGE: AWS Data Center Interior | Real-world cloud infrastructure: servers, cooling systems, and networking hardware in an AWS facility.

IMAGE: Hybrid Cloud Architecture Diagram | Visual representation of private + public cloud integration (like Ncell’s setup).

In the real world

• eSewa (SaaS): Uses AWS’s scalability to handle 100,000+ transactions/minute during Dashain, auto-scaling servers based on demand.
• Ncell (Hybrid Cloud): Stores customer data in private cloud (compliance with Nepal’s data laws) while using AWS for public cloud to manage call routing during peak hours (e.g., elections).
• Pathao (Public Cloud): Leverages Google Cloud’s global reach to match drivers/rides in real-time across Kathmandu/Lalitpur, reducing latency via multi-region deployments.