Cloud Deployment Models: Public, Private, Hybrid, Community

TAKEAWAYS:

• Cloud deployment models define where and how cloud services are hosted, accessed, and managed (public vs. private vs. hybrid).
• Public clouds (e.g., AWS, Google Cloud) offer shared, scalable resources but require strict security controls.
• Private clouds (e.g., corporate data centers) provide full control and isolation but demand high upfront costs.
• Hybrid clouds (e.g., Ncell’s backend + AWS for peak traffic) combine public and private for flexibility and compliance.
• Community clouds (e.g., NTC’s shared infrastructure for telecom providers) serve specific groups with shared needs.
• Cost, security, scalability, and compliance are the key factors when choosing a deployment model.

1. Introduction to Cloud Deployment Models

Cloud deployment models determine how cloud resources are allocated, secured, and accessed. Unlike service models (IaaS/PaaS/SaaS), which define what is delivered, deployment models define where and how the cloud operates.

Why Does It Matter?

• Security: Private clouds isolate sensitive data (e.g., bank transactions).
• Cost: Public clouds reduce capital expenditure (e.g., Daraz’s seasonal sales).
• Compliance: Government clouds (e.g., Nepal’s e-Governance) require strict data sovereignty.
• Scalability: Hybrid clouds handle unpredictable loads (e.g., Pathao during Diwali).

2. The Four Cloud Deployment Models

Use this table to compare them at a glance:

3. Deep Dive: How Each Model Works

A. Public Cloud

Definition: A third-party provider (e.g., AWS, Azure) hosts resources in a shared, multi-tenant environment. Users pay for what they use (pay-as-you-go).

How It Works:

1. Resource Pooling: Physical servers are virtualized into shared instances.
2. Self-Service Portal: Users provision resources via dashboards (e.g., AWS Console).
3. Automatic Scaling: Resources scale up/down based on demand (e.g., WhatsApp’s message queues).

Visual: Public Cloud Architecture

flowchart TD
    A["Users (eSewa, Daraz)"] -->|"HTTPS"| B["Load Balancer"]
    B --> C["Public Cloud Provider (AWS/Azure)"]
    C --> D["Shared Virtual Machines"]
    C --> E["Shared Storage (S3/Blob Storage)"]
    C --> F["Shared Network (VPC/Subnets)"]
    D -->|"Isolated"| G["User A: Web App"]
    D -->|"Isolated"| H["User B: Database"]

Real-World Example: eSewa’s Public Cloud

• Why? eSewa uses AWS to handle millions of transactions/day without investing in physical servers.
• How?
  • Compute: EC2 instances for APIs.
  • Storage: S3 for user data backups.
  • Database: RDS (PostgreSQL) for transaction logs.
• Challenge: Ensures PCI-DSS compliance (payment security) in a shared environment.

B. Private Cloud

Definition: Dedicated cloud infrastructure for a single organization, hosted on-premises or by a third party. Offers full control over security and customization.

How It Works:

1. Dedicated Hardware: No sharing with other tenants.
2. Custom Security: Firewalls, encryption, and access controls tailored to the org.
3. Internal Management: IT teams handle updates, patches, and scaling.

Visual: Private Cloud vs. Public Cloud

flowchart LR
    subgraph Private Cloud ["Nabil Bank's Private Cloud"]
        A["Dedicated Servers"] --> B["Internal Firewall"]
        B --> C["Core Banking App"]
        B --> D["Customer Data DB"]
    end
    subgraph Public Cloud ["AWS"]
        E["Shared Servers"] --> F["Multi-Tenant VMs"]
        F --> G["eSewa App"]
    end

Real-World Example: Nabil Bank’s Private Cloud

• Why? Handles sensitive financial data (loans, accounts) requiring Nepal Rastra Bank (NRB) compliance.
• How?
  • On-Premises Servers: IBM Power Systems for transaction processing.
  • Custom Encryption: AES-256 for data at rest.
  • Disaster Recovery: Mirrored data centers in Kathmandu and Pokhara.
• Cost: High upfront investment (~$500K+) but zero shared-risk exposure.

C. Hybrid Cloud

Definition: A combination of public and private clouds, connected via secure channels (e.g., VPN, API gateways). Data/missions move between environments based on needs.

How It Works:

1. Workload Segmentation:
   • Private Cloud: Sensitive data (e.g., customer PII).
   • Public Cloud: Scalable services (e.g., seasonal promotions).
2. Orchestration: Tools like AWS Outposts or Azure Arc manage hybrid workflows.
3. Data Sync: APIs or cloud bursting (e.g., Ncell routing overflow traffic to AWS).

Visual: Hybrid Cloud Workflow (Ncell Example)

sequenceDiagram
    participant User as Customer
    participant NcellCore as Ncell Private Cloud
    participant AWS as AWS Public Cloud
    participant Database as Private DB

    User->>NcellCore: Calls 9800000000
    NcellCore->>AWS: Checks if traffic spike (e.g., Diwali)
    AWS-->>NcellCore: Yes, route to AWS
    NcellCore->>AWS: Spin up EC2 for IVR
    AWS->>User: Handles call
    AWS->>Database: Writes logs to private DB via VPN

Real-World Example: Ncell’s Hybrid Cloud

• Why? Avoids public cloud latency for core services but uses AWS for peak-hour call routing.
• How?
  • Private Cloud: Core billing system (on-prem).
  • Public Cloud: AWS Lambda for real-time fraud detection.
  • Connection: Direct Connect (dedicated 10Gbps link).
• Result: 99.99% uptime during festivals (vs. 99.9% with pure public cloud).

D. Community Cloud

Definition: Shared infrastructure for a specific community (e.g., government agencies, healthcare providers, or telecom operators). Governed by a joint steering committee.

How It Works:

1. Shared Costs: Members split infrastructure costs.
2. Custom Policies: Security/compliance tailored to the community (e.g., HIPAA for hospitals).
3. Limited Access: Only authorized members can use resources.

Visual: Community Cloud Topology

flowchart TD
    subgraph Community Cloud ["NTC Telecom Cloud"]
        A["ISP 1: Ncell"] --> B["Shared Core Network"]
        A["ISP 2: NTC"] --> B
        A["ISP 3: Smart"] --> B
        B --> C["Shared SDN Controller"]
        B --> D["Community Storage"]
    end

Real-World Example: NTC’s Community Cloud for ISPs

• Why? Telecom providers (Ncell, Smart, NTC) share backbone infrastructure to reduce costs.
• How?
  • Shared Peering: Reduces inter-ISP latency.
  • Joint Security: DDoS protection via community firewalls.
  • Cost Savings: Each ISP pays ~30% less than building private infrastructure.
• Challenge: Neutrality rules ensure no ISP gets preferential treatment.

4. Key Factors in Choosing a Deployment Model

Use this decision tree to pick the right model:

flowchart TD
    A["Need for Scalability?"] -->|"Yes"| B["Public or Hybrid"]
    A -->|"No"| C["Private or Community"]
    B --> D["Need for Security?"] -->|"High"| E["Hybrid (Public for scale, Private for data)"]
    B --> D -->|"Low"| F["Public Cloud"]
    C --> G["Budget?"] -->|"High"| H["Private Cloud"]
    C --> G -->|"Low"| I["Community Cloud"]

Decision Criteria Table

5. Worked Example: Daraz’s Hybrid Cloud Strategy

Scenario: Daraz Nepal faces spikes in traffic during sales (e.g., Dashain, Tihar) but needs secure inventory management.

Solution:

• Private Cloud: On-premises ERP (SAP) for inventory and orders.
• Public Cloud (AWS): Auto-scaling web servers and CDN for global users.
• Connection: AWS Direct Connect for low-latency data sync.

Step-by-Step Trace:

1. Normal Traffic:
   • All requests handled by private cloud servers.
2. Peak Traffic (e.g., Dashain Sale):
   • CloudWatch detects traffic surge.
   • Auto Scaling spins up 100+ EC2 instances in AWS.
   • Route 53 directs users to the nearest AWS region (e.g., Singapore).
3. Post-Sale:
   • AWS instances scale down to zero.
   • Inventory updates sync back to private SAP database.

Visual: Daraz’s Hybrid Workflow

sequenceDiagram
    participant User as Customer
    participant DarazPrivate as Daraz Private Cloud
    participant AWS as AWS Public Cloud
    participant CDN as CloudFront CDN

    User->>CDN: Requests product page
    CDN->>AWS: Fetches static content
    User->>DarazPrivate: Places order
    DarazPrivate->>AWS: Triggers scaling if needed
    AWS->>DarazPrivate: Syncs order to ERP

Outcome:

• Cost Savings: ~40% lower than pure public cloud.
• Performance: <200ms latency even during sales.
• Security: Inventory data never leaves private cloud.

6. Advantages and Disadvantages

7. Real-World Applications in Nepal

8. Common Misconceptions

1. "All clouds are the same."

   • Reality: Public clouds (AWS) ≠ private clouds (on-prem). Security, cost, and compliance vary wildly.

2. "Hybrid clouds are just ‘public + private.’"

   • Reality: They require orchestration tools (e.g., Kubernetes, Terraform) to manage workloads seamlessly.

3. "Community clouds are only for governments."

   • Reality: Any group with shared needs can use them (e.g., university research clusters).

9. Exam Tip: How This Unit Is Tested

This unit is heavily tested in TU/PU exams via:

1. Definitions & Comparisons (3–5 marks):

   • "Differentiate between public and private clouds with examples."
   • Key: Use the table above and real-world examples (e.g., eSewa vs. Nabil Bank).

2. Scenario-Based Questions (5–8 marks):

   • "A hospital wants to store patient records securely but needs to scale its telemedicine app. Recommend a deployment model and justify."
   • Key: Hybrid cloud (private for records, public for app scaling). Draw a sequence diagram like the Ncell example.

3. Diagram-Based Questions (4–6 marks):

   • "Draw and label the architecture of a hybrid cloud."
   • Key: Use the Daraz hybrid workflow or Ncell example as a template.

4. Short Answer (2–3 marks):

   • "What is a community cloud? Give one Nepalese example."
   • Answer: "A shared cloud for a specific group. Example: NTC’s community cloud for ISPs."

Pro Tip:

• Memorize the 4 models and one real-world example each.
• Practice drawing hybrid/public/private architectures.
• Relate to Nepal: Examiners love eSewa, Ncell, Nabil Bank, and NTC as examples.

10. Summary Checklist

Before the exam, ensure you can: ✅ Define all 4 deployment models and their key characteristics. ✅ Compare them using a table (cost, security, scalability). ✅ Explain how hybrid clouds work with a sequence diagram. ✅ Give 2 Nepalese examples for each model (e.g., eSewa = public). ✅ Describe one real-world scenario (e.g., Daraz’s hybrid strategy). ✅ Draw architecture diagrams for public/private/hybrid clouds.