Hypervisors: Types, Functions & Architecture

What is a Hypervisor?

A hypervisor (or Virtual Machine Monitor, VMM) is software/firmware that creates, manages, and runs virtual machines (VMs) on a single physical host. It abstracts hardware resources (CPU, RAM, storage, network) and allocates them to VMs while ensuring isolation and security.

Why Use Hypervisors?

• Resource Efficiency: Run multiple OSes on one machine (e.g., Windows + Linux on a single server).
• Isolation: Crash one VM without affecting others (critical for cloud services).
• Scalability: Dynamically allocate resources (e.g., auto-scaling in AWS).
• Cost Savings: Reduce hardware costs by consolidating workloads.

Types of Hypervisors

Hypervisors are classified based on their placement in the system architecture. The two main types are:

1. Type-1 (Bare-Metal) Hypervisor

• Runs directly on hardware (no host OS).
• Faster performance (lower overhead).
• Used in enterprise/data centers (e.g., VMware ESXi, Microsoft Hyper-V, Xen).
• Example: A cloud provider like Ncell’s data center uses Type-1 hypervisors to host thousands of VMs for their services.

graph TD
    A["Physical Hardware"] --> B["Type-1 Hypervisor"]
    B --> C["VM 1 (OS: Ubuntu)"]
    B --> D["VM 2 (OS: Windows Server)"]
    B --> E["VM 3 (OS: Linux)"]

2. Type-2 (Hosted) Hypervisor

• Runs on top of a host OS (e.g., Windows, Linux).
• Slower (extra layer of abstraction).
• Used for development/testing (e.g., Oracle VirtualBox, VMware Workstation).
• Example: A Nepalese developer testing a web app on both Windows and Linux VMs using VirtualBox.

graph TD
    A["Host OS (Windows/Linux)"] --> B["Type-2 Hypervisor"]
    B --> C["VM 1 (OS: Ubuntu)"]
    B --> D["VM 2 (OS: macOS)"]

How Hypervisors Work: Core Functions

A hypervisor performs three key tasks:

1. Hardware Abstraction

• Virtualizes hardware (CPU, RAM, storage, network) so VMs think they have dedicated resources.
• Uses emulation (e.g., running ARM code on x86 via QEMU) or paravirtualization (guest OS modified for efficiency).
• Example: A Daraz server running multiple e-commerce VMs on a single physical machine.

2. Resource Allocation & Scheduling

• CPU Scheduling: Uses algorithms like Round Robin, Priority Scheduling to share CPU time.
• Memory Management: Allocates RAM dynamically (e.g., ballooning in VMware to reclaim memory).
• Storage Virtualization: Uses LUN masking, SAN/NAS to provide storage to VMs.
• Network Virtualization: Assigns virtual NICs (vNICs) and manages traffic via vSwitches.

stateDiagram-v2
    [*] --> Idle
    Idle --> Running: "CPU Allocated"
    Running --> Idle: "Time Slice Ends"
    Running --> Paused: "Low Priority"
    Paused --> Running: "High Priority"

3. Isolation & Security

• Memory Isolation: Prevents VMs from accessing each other’s RAM.
• CPU Pinning: Binds VMs to specific cores for performance.
• Security Features:
  • Sandboxing (e.g., Docker containers).
  • Encryption (e.g., VMware’s vTPM for secure boot).
• Example: Nepal Rastra Bank uses hypervisor isolation to secure financial transactions across VMs.

Hypervisor Architecture

Hypervisors can be designed in two ways:

1. Microkernel Architecture

• Minimal core (only essential functions like scheduling, memory management).
• Device drivers and services run as separate modules.
• Pros: More stable (crash in one module doesn’t affect others).
• Cons: Slightly slower due to inter-module communication.
• Example: Xen (used in AWS for high availability).

classDiagram
    class Hypervisor {
        +Core Scheduling
        +Memory Management
    }
    class Module1 {
        +Device Driver
    }
    class Module2 {
        +Network Service
    }
    Hypervisor --> Module1
    Hypervisor --> Module2

2. Monolithic Architecture

• All components (scheduling, I/O, memory) in a single kernel.
• Pros: Faster performance (no inter-module overhead).
• Cons: Less stable (crash in one component affects others).
• Example: VMware ESXi (used in enterprise data centers).

graph TD
    A["Monolithic Hypervisor"] --> B["CPU Scheduler"]
    A --> C["Memory Manager"]
    A --> D["I/O Subsystem"]

Real-World Applications of Hypervisors

1. Cloud Computing (AWS, Google Cloud, NTC’s Cloud Services)

• How it works: AWS uses Xen and KVM hypervisors to run thousands of VMs for services like EC2 (Elastic Compute Cloud).
• Example: When you use eSewa to pay bills online, the backend servers are likely running on hypervisor-managed VMs for scalability.

2. Data Centers & Enterprise IT (Ncell, Nabil Bank, NEPSE)

• How it works: Banks like Nabil Bank use VMware ESXi to host multiple VMs for different departments (HR, Finance, Customer Service) on a single physical server.
• Example: During NEPSE trading hours, stock exchange servers use hypervisors to handle high traffic without downtime.

3. Development & Testing (Local Machines, Universities)

• How it works: Developers use VirtualBox to test software on different OSes (e.g., a Pathao driver app tested on Android + iOS VMs).
• Example: A PU student testing a Python app on both Windows and Linux VMs without buying separate hardware.

Worked Example: Hypervisor in a Bank’s Loan Processing System

Scenario: A bank like Global IME Bank processes loan applications 24/7. They use a Type-1 hypervisor (VMware ESXi) to run:

• VM 1: Customer portal (Windows Server).
• VM 2: Loan processing (Linux).
• VM 3: Database (SQL Server).

Steps:

1. Resource Allocation:

   • VM 1 (Customer Portal) gets 4 CPU cores, 8GB RAM.
   • VM 2 (Loan Processing) gets 8 CPU cores, 16GB RAM (higher priority during peak hours).
   • VM 3 (Database) gets SSD storage, 32GB RAM.

2. Isolation:

   • If the customer portal crashes, loan processing continues unaffected.

3. Scalability:

   • During Diwali season, the bank clones VM 1 to handle extra traffic.

Mermaid Diagram:

sequenceDiagram
    participant User
    participant VM1 as Customer Portal (VM)
    participant VM2 as Loan Processing (VM)
    participant Database as SQL Server (VM)
    User->>VM1: Logs in
    VM1->>VM2: Validates loan
    VM2->>Database: Queries data
    Database-->>VM2: Returns records
    VM2-->>VM1: Approves/Rejects
    VM1-->>User: Displays result

Popular Hypervisor Software

Advantages & Disadvantages of Hypervisors

Advantages

✅ Cost Savings: Fewer physical servers needed. ✅ High Availability: Live migration (e.g., VMware vMotion) for zero downtime. ✅ Disaster Recovery: Snapshots and backups of VMs. ✅ Flexibility: Run different OSes on one machine. ✅ Security: Isolation prevents malware from spreading across VMs.

Disadvantages

❌ Complexity: Managing multiple VMs requires expertise. ❌ Performance Overhead: Type-2 hypervisors slow down the host. ❌ Licensing Costs: Enterprise hypervisors (e.g., VMware) are expensive. ❌ Single Point of Failure: If the hypervisor crashes, all VMs go down.

Exam Tip

What to Focus On

1. Differentiate Type-1 vs. Type-2 hypervisors (performance, use cases, examples).
2. Explain how hypervisors allocate resources (CPU, RAM, storage, network).
3. Compare microkernel vs. monolithic architecture (stability vs. speed).
4. Real-world examples:
   • Cloud providers (AWS, NTC) use Type-1 hypervisors.
   • Banks (Nabil, Global IME) use VMware ESXi for security.
   • Developers use VirtualBox for testing.
5. Diagrams:
   • Draw hypervisor architecture (layers).
   • Show VM lifecycle (creation, suspension, deletion).
   • Illustrate resource allocation (CPU, RAM sharing).

Common Exam Questions

• "How does a Type-1 hypervisor differ from a Type-2?"
• "Explain the role of a hypervisor in cloud computing."
• "What are the security benefits of hypervisor isolation?"
• "Compare Xen and VMware ESXi in terms of architecture."

In the Real World

1. eSewa & Khalti (Digital Payments)

   • Idea Used: Type-1 Hypervisors (KVM/Xen)
   • How: eSewa’s backend servers run on scalable VMs managed by hypervisors to handle millions of transactions during Dashain/Tihar. The hypervisor ensures isolation so a payment failure in one VM doesn’t crash the entire system.

2. Ncell & NTC (Telecom Services)

   • Idea Used: Live Migration (VMware vMotion)
   • How: During peak hours, Ncell dynamically moves VMs between physical servers using hypervisor-based live migration to prevent overload. This ensures 99.99% uptime for calls and internet.

3. Daraz & Pathao (E-Commerce & Ride-Hailing)

   • Idea Used: Containerization + Hypervisors
   • How: Daraz uses Docker containers (lightweight VMs) on hypervisors to deploy microservices (e.g., order processing, payment gateway). Pathao uses hypervisors to scale ride-matching servers during festival traffic surges.

Key Takeaways

• Hypervisors are the engine of virtualization, enabling cloud computing and resource efficiency.
• Type-1 (bare-metal) is for enterprise, Type-2 (hosted) is for development.
• Resource allocation, isolation, and scheduling are core functions.
• AWS, banks, and telecoms rely on hypervisors for scalability and security.
• Exam focus: Compare types, explain architecture, and give real-world cloud examples.