Docker 기초 : 가상화 개념Docker Basics: Virtualization Concepts

Physical Machine

A physical machine is an actual computer you can see and touch. It’s made of physical hardware such as a CPU, memory and storage, and usually runs a single operating system.

> A physical machine that actually exists

Limits of Physical Machines

Scaling physical machines means thinking not only about the cost of buying hardware but also about infrastructure like floor space and electricity. Physical machines also lack elasticity: once you’ve added resources, it’s hard to grow them further or scale them back down.

> Physical machines carry the burden of expanding infrastructure and have poor elasticity

Cloud Computing

The graph below compares how computing capacity is managed over time with the traditional hardware approach versus the automated cloud approach. The limits of the traditional hardware approach it points out are:

• Large capital expenditure: capacity has to grow in steps
• Opportunity cost: losses from the gap between actual demand and provided capacity
• Fixed infrastructure: no immediate response to changes in demand

> Source: Cloud Computing - Benefits to the Business

To overcome these limits, cloud computing, which lets you use resources elastically, keeps evolving. And the core technology behind cloud computing is ‘virtualization’.

Virtual Machine

Virtualization is a technology that logically partitions a physical machine’s resources. A virtual computer built on top of a physical machine is called a ‘virtual machine’, and virtualization software runs and manages virtual machines.

Virtualization lets you run several virtual environments on one physical system. Each virtual machine has its own operating system (OS) and resources and runs in a fully isolated environment.

> You can build virtual machines on top of a physical machine.

Virtualization Software

There are a few kinds of virtualization software.

• Host-based virtualization
• Hypervisor-based virtualization
• Container-based virtualization

Host OS / Guest OS

Before getting into virtualization software, let’s sort out the basic kinds of OS.

• Host OS : the OS running directly on the physical machine
• Guest OS : the OS running on a virtual machine

Host-Based Virtualization

Host-based virtualization software is installed on the host OS and manages guest OSes. Just like installing Chrome on your operating system, you install the virtualization software and create new OSes inside it.

> Host-based virtualization

Pros

The host OS and guest OS coexist, so you can work on the host and the guest at the same time.

Cons

To control hardware, the guest OS has to go through the host OS, so hardware-related performance suffers.

Representative Products

• Oracle VM VirtualBox
• VMware Fusion

Hypervisor-Based Virtualization

Hypervisor-based virtualization software is installed directly on the physical machine and manages guest OSes.

> Hypervisor-based virtualization

Pros

Since there’s no host OS, more resources can go to the guest OSes, and direct access to the hardware delivers high performance.

Cons

Unlike host-based virtualization, it can’t coexist with a host OS.

Representative Products

• VMware ESXi
• Microsoft Hyper-V
• KVM (Kernel-based Virtual Machine)
• Xen

Container Virtualization

Container-based virtualization shares the host OS’s kernel and uses process isolation to manage applications as containers, without booting a guest OS.

> Container virtualization

Pros

Because no guest OS is booted, it uses fewer resources and starts quickly. It can also coexist with the host OS.

Cons

With no guest OS, running Linux containers means borrowing a Linux kernel from the host OS. So the downside is that the host OS has to provide a Linux kernel by some additional means.

Representative Products

• Docker
• Podman

Wrapping Up

Before learning Docker, we looked at virtualization technologies. In the next post, let’s look at the Docker Daemon, which manages containers!

Finally, let’s recap each virtualization technology once more. 😊

What’s Next

In the next post, Docker Basics: Docker Engine, we’ll cover the basic concepts of containers and images!