01_Introduction_to_Virtualization

Loading views...

Introduction to Virtualization

Complete Study Notes for Cloud Computing

What is Virtualization?
Virtualization Classification
Key Technical Distinctions
Evolution Timeline

What is Virtualization?

Virtualization is a technology that creates an abstraction layer over physical computer hardware, allowing resources such as processors, memory, storage, and networks to be divided into multiple virtual environments. These environments—known as Virtual Machines (VMs) or Containers—can run their own operating systems and applications as if they were independent physical machines.

Core Benefits

Resource Optimization: Run multiple workloads on single physical hardware
Isolation: Separate environments prevent interference between workloads
Flexibility: Easy provisioning, scaling, and migration of workloads
Cost Efficiency: Reduced hardware requirements and energy consumption
High Availability: Enable live migration and disaster recovery

Virtualization Classification

1. Based on Implementation Layer

A. Hardware-Level Virtualization

Full Virtualization (Hardware-Assisted)

Utilizes CPU extensions: Intel VT-x, AMD-V
Examples: VMware ESXi, KVM with hardware support
Guest OS runs completely unmodified
Binary translation eliminated through hardware support
Best performance for unmodified guest operating systems

Para-virtualization

Examples: Xen (classic mode)
Guest OS kernel modified to use hypercalls instead of privileged instructions
Better performance than pure software virtualization
Trade-off: Requires OS modification (less portable)
Guest is "aware" it's running in a virtualized environment

B. Software-Level Virtualization

Operating System-Level Virtualization (Containers)

Examples: Docker, LXC, OpenVZ, Podman
All containers share the host kernel
Extremely lightweight with fast startup times
Uses namespaces for process isolation
Uses cgroups for resource control

Application-Level Virtualization

Examples: Java Virtual Machine (JVM), .NET CLR, Wine
Application runs in isolated sandbox
Provides portable runtime environment
Wine provides Windows API translation on Linux

2. Based on Architecture Approach

Type-1 Hypervisor (Bare-Metal/Native)

┌───────────────────────────────────────────┐
│        VM 1        VM 2        VM 3       │
│       ┌────┐      ┌────┐      ┌────┐      │
│       │ OS │      │ OS │      │ OS │      │
│       │Apps│      │Apps│      │Apps│      │
│       └────┘      └────┘      └────┘      │
├───────────────────────────────────────────┤
│           Hypervisor (VMM)                │
│   Scheduler │ Memory Mgr │ I/O Handler    │
├───────────────────────────────────────────┤
│          Physical Hardware                │
│    CPU │ Memory │ NIC │ Storage │ GPU     │
└───────────────────────────────────────────┘

Key Characteristics:

Runs directly on hardware (no host OS layer)
Operates in Ring 0 (highest privilege level)
Direct management of all hardware resources
Lower overhead due to minimal abstraction layers
Production-grade for data centers and cloud infrastructure

Major Type-1 Hypervisors:

Hypervisor	Architecture	Footprint	Management
VMware ESXi	Microkernel	~150MB	vCenter
Xen	Microkernel	~1MB	xl/xm/XAPI
KVM	Linux-integrated	Kernel module	libvirt/virsh
Hyper-V	Microkernel	~20MB	SCVMM/PowerShell

Type-2 Hypervisor (Hosted)

┌───────────────────────────────────────────┐
│        VM 1        VM 2        VM 3       │
│       ┌────┐      ┌────┐      ┌────┐      │
│       │ OS │      │ OS │      │ OS │      │
│       │Apps│      │Apps│      │Apps│      │
│       └────┘      └────┘      └────┘      │
├───────────────────────────────────────────┤
│   Hypervisor (User-space application)     │
│   VM Manager │ Device Emulation           │
├───────────────────────────────────────────┤
│       Host Operating System               │
│    (Windows, Linux, macOS)                │
├───────────────────────────────────────────┤
│          Physical Hardware                │
└───────────────────────────────────────────┘

Key Characteristics:

Runs as an application on host OS
Relies on host OS for device drivers
Higher overhead due to extra OS layer
Easier installation and management
Ideal for development and testing environments

Major Type-2 Hypervisors:

Hypervisor	Host OS	Performance	Use Case
VMware Workstation	Win/Linux	Excellent	Development
VirtualBox	Win/Linux/Mac	Good	Testing
QEMU (no KVM)	Any	Poor (emulation)	Cross-architecture
Parallels	macOS only	Excellent	Mac users

3. Based on Resource Virtualization

Compute Virtualization

CPU Virtualization: Hardware-assisted (VT-x/AMD-V), Binary translation, Paravirtualization
vCPUs mapped to physical CPU cores/threads
CPU scheduling algorithms manage time-sharing

Memory Virtualization

Shadow page tables (legacy approach)
Extended Page Tables (EPT) / Nested Page Tables (NPT) - Hardware-assisted
Memory ballooning - Dynamic memory reclamation
Kernel Same-page Merging (KSM) - Deduplication of identical pages

Storage Virtualization

Virtual disks: QCOW2, VMDK, VHD formats
Storage pools and volumes
Thin provisioning - Allocate on demand
Snapshot and cloning mechanisms

Network Virtualization

Virtual switches: Linux Bridge, Open vSwitch (OVS)
Virtual NICs: virtio-net, e1000
Network namespaces for isolation
Software-Defined Networking (SDN)
Tunneling protocols: VXLAN, GRE

4. Based on Isolation Mechanism

Full Isolation (Strong)

Traditional VMs with separate kernels
Each VM has complete OS stack
Hardware-enforced boundaries via CPU virtualization extensions
Highest security but highest overhead

Namespace Isolation (Lightweight)

Linux Namespaces provide isolation for:

Namespace	Purpose	Kernel Version
PID	Process isolation	2.6.24 (2008)
Network	Network stack isolation	2.6.29 (2009)
Mount	Filesystem isolation	2.4.19 (2002)
UTS	Hostname isolation	2.6.19 (2006)
IPC	Inter-process communication	2.6.19 (2006)
User	UID/GID mapping	3.8 (2013)
Cgroup	Control group visibility	4.6 (2016)
Time	Time isolation	5.6 (2020)

Hybrid Isolation

Kata Containers: Combines container speed with VM security
gVisor: User-space kernel for containers
Firecracker: MicroVM for serverless workloads

Key Technical Distinctions

Containers vs Virtual Machines

Aspect	Containers	VMs
Kernel	Shared with host	Separate per VM
Startup Time	Seconds	Minutes
Size	Megabytes	Gigabytes
Isolation	Namespace-based	Hardware-enforced
Density	High (thousands per host)	Low (tens per host)
Overhead	Minimal (~1-2%)	Moderate (~5-15%)

Hardware vs Software Virtualization

Aspect	Hardware Virtualization	Software Virtualization
Requirements	CPU extensions (VT-x/AMD-V)	None
Method	Hardware trap and emulate	Binary translation
Performance	Near-native (95%+)	Slower (70-90%)
Guest OS	Unmodified	May need modification
Use Case	Production workloads	Cross-architecture emulation

Evolution Timeline

1998: VMware founded (first x86 virtualization)
2003: Xen hypervisor released
2005: Intel VT-x introduced
2006: AMD-V introduced, Intel VT-d (IOMMU)
2007: KVM merged into Linux kernel 2.6.20
2008: Intel EPT, AMD NPT (memory virtualization)
2008: cgroups v1 merged into kernel
2010: SR-IOV widespread adoption
2013: Docker released (container revolution)
2014: Kubernetes released by Google
2015: NVIDIA vGPU for GPU virtualization
2016: cgroups v2 in kernel 4.5
2017: AMD SEV (encrypted virtualization)
       AWS Nitro System (KVM-based)
2018: Firecracker released by AWS
2020: Intel TDX announced
2021: DPU/SmartNIC mainstream adoption

Summary

Virtualization forms the foundation of modern cloud computing, enabling:

Multi-tenancy: Multiple customers sharing physical infrastructure
Elasticity: Dynamic resource allocation
Portability: Workloads can move between physical hosts
Efficiency: Better hardware utilization

The choice of virtualization technology depends on:

Security requirements: Full VMs for strong isolation
Performance needs: Containers for minimal overhead
Workload characteristics: Stateless vs stateful
Operational constraints: Skill sets, tooling, ecosystem

Continue to: 02HardwareBasedVirtualization.md

← Previous Next →