A must-read for anyone working in IT!
Easily understand common sense about IT infrastructure through pictures!

This book contains the core of infrastructure technologies that the authors have personally acquired in various environments, and it reflects the authors' thoughts on how to make so-called basic technologies such as architecture, networks, servers, and processes easily understandable to readers.
Therefore, you will be able to correctly understand the common principles (structure) of each part of the system, and thereby develop the basic ability to immediately adapt to new technologies.
Additionally, the abundant illustrations explaining the principles will help even readers with little practical experience easily understand the overall structure.

You might think of things we use at home, such as electricity, water, and gas, or infrastructure for public purposes, such as subways and buses.
In Korean, infrastructure means 'foundation', meaning the foundation or basis that supports your life.
Although the infrastructure itself is complex, it is managed by experts, so users can easily use it without having to understand the structure.
---p.3

In an OS, the kernel is the heart, the brain, and the spinal cord.
It is no exaggeration to say that the kernel is the essence of the OS, and the rest is just extras.
You can think of the kernel itself as the 'infrastructure' of the OS.
The kernel has many roles, but the most important one is to provide a convenient interface while hiding what is happening behind the scenes.
Because the kernel exists, developers can create applications without being conscious of their impact on hardware or other applications.

---p.50

In the computer world, exclusive control is not necessary in serial processing, but it is necessary in parallel processing.
The part that has exclusive control is prone to bottlenecks.
Being able to visualize parallel processing and exclusive control can help with problem solving and performance tuning.
For example, consider the use of a conference room (Figure 4.17).

---p.97

Interruption is an 'event-driven' structure that contacts when something happens.
Alternatively, the CPU can periodically poll the I/O device (see Section 5.3, "Polling") to check the status of the device, but if the polling interval is long, it may not immediately notice that (for example) disk I/O has completed.
Conversely, a shorter interval results in frequent polling, which wastes a lot of CPU power. The frequency of I/O device status updates is low compared to CPU speed.
Therefore, we are adopting interruption because it is more efficient to control using interruption than polling.

---p.137

TCP manages data in units called segments.
For this reason, TCP segments are created by attaching a TCP header to the application data.
The header contains a lot of information to describe TCP functionality, including the destination port number.
The maximum data size that can be transmitted in one TCP segment is called MSS (Maximum Segment Size).
Because data is ultimately transmitted using the link layer, MSS depends on the maximum size that can be transmitted on the link layer, and varies depending on the environment or settings.
As we will see again in Section 6.8, the maximum data size that can be transmitted at the link layer is called the Maximum Transfer Unit (MTU).

---p.197