Server Virtualization Architecture

1. Defining Server-Level Virtualization

At its core, server virtualization is the process of partitioning a single physical server into multiple, isolated virtual servers. Each virtual server, known as a Virtual Machine (VM), runs its own operating system and applications, functioning as a complete, independent server.

This concept is the practical application of the server consolidation benefit we discussed previously.

To get a precise, exam-ready definition, please read the following brief section from the GeeksforGeeks article.

As the article explains, server virtualization allows you to run, for instance, a Linux web server, a Windows database server, and another Linux mail server all on the same physical machine, with each being completely isolated from the others.

The following diagram illustrates this high-level concept:

2. The Core Architectural Components

The architecture of server virtualization is best understood as a stack of layers. There are three fundamental components you must know for your exams:

1. Physical Hardware (The Host): This is the underlying physical server, which provides the actual computing resources: CPU, memory (RAM), storage (hard disks), and network interfaces.
2. Hypervisor (or Virtual Machine Monitor - VMM): This is a specialized software layer that sits between the physical hardware and the virtual machines. Its job is to create, run, and manage the VMs. The hypervisor is the core engine of virtualization; it abstracts the physical hardware and allocates resources to each VM.
3. Virtual Machine (The Guest): A VM is a software-based computer. It is a self-contained package of a guest operating system (e.g., Windows, Linux) and the application(s) running on it. From the perspective of the software inside the VM, it appears to be a dedicated physical machine with its own virtual hardware.

3. Architectural Models: Type 1 vs. Type 2 Hypervisors

The specific arrangement of these components—and thus the server virtualization architecture—depends on the type of hypervisor used. There are two primary architectural models, defined by Type 1 and Type 2 hypervisors. Understanding this distinction is critical.

The following video gives an excellent animated overview of virtualization and then builds a virtualized server step-by-step, illustrating the Type 1 architecture in action.

Now, let's formally define and illustrate the two architectures.

Architecture 1: Type 1 Hypervisor (Bare-Metal)

As the name "bare-metal" suggests, a Type 1 hypervisor is installed and runs directly on the host's physical hardware. There is no traditional operating system in between. This architecture is the standard for enterprise data centers and cloud providers because it offers the best performance and security.

• Architectural Stack:

  1. Physical Hardware
  2. Hypervisor
  3. Virtual Machines (each with a Guest OS and Apps)

• Key Characteristics:

  • High Performance: The hypervisor has direct access to hardware, leading to minimal overhead.
  • High Security & Isolation: With no underlying host OS, there is a smaller attack surface. Issues in one VM are completely isolated from others.
  • Examples: VMware ESXi, Microsoft Hyper-V, KVM (Kernel-based Virtual Machine).

Architecture 2: Type 2 Hypervisor (Hosted)

A Type 2 hypervisor runs as a software application on top of a conventional host operating system (like Windows, macOS, or Linux). The hypervisor then creates and manages VMs within this application. This architecture is common for personal use, development, and testing.

• Architectural Stack:

  1. Physical Hardware
  2. Host Operating System (e.g., Windows 10)
  3. Hypervisor Application
  4. Virtual Machines (each with a Guest OS and Apps)

• Key Characteristics:

  • Lower Performance: The hypervisor's requests for hardware resources must pass through the host OS, adding latency and overhead. The host OS also consumes resources (CPU, RAM).
  • Easier to Manage: It's as simple to install and run as any other desktop application.
  • Examples: Oracle VirtualBox, VMware Workstation, Parallels Desktop.

The following diagram is essential for your exams as it clearly illustrates the architectural difference between the two types.

4. Summary and Comparison

For exam preparation, being able to compare these two architectures in a tabular format is extremely valuable. The following resource from AWS provides an excellent, detailed breakdown.

Based on the resource and our discussion, here is a consolidated comparison table for your notes:

Conclusion

In this lesson, we dissected the architecture of server-level virtualization, the technology that enables the benefits we previously discussed. We defined it as the partitioning of one physical server into multiple isolated virtual machines.

Key Takeaways:

• Server Virtualization creates multiple independent virtual servers (VMs) on a single physical host.
• The architecture consists of three main layers: Physical Hardware, the Hypervisor, and the Virtual Machines.
• The specific architecture is determined by the hypervisor type.
• Type 1 (Bare-Metal) Architecture: The hypervisor runs directly on the hardware, offering high performance and security. This is the standard for enterprise and cloud environments.
• Type 2 (Hosted) Architecture: The hypervisor runs as an application on a host OS, offering ease of use but with a performance penalty. This is used for desktop-level virtualization.
• Being able to draw and explain the layered diagrams for both Type 1 and Type 2 architectures is crucial for exams.

Preview of the Next Lesson:

Now that we have a solid understanding of server virtualization architecture, we will broaden our scope. The next lesson will cover the learning outcome: Explain the concepts of hardware virtualization and operating-system-level virtualization, noting key differences. We will dive deeper into different methods and levels of abstraction used in virtualization.