Linux — Why Network is Layered

Different networking model introduction

Introduction

When learning network troubleshooting, the first thing you may have to figure out is the layered model of the network. In our work, we often hear these terms, such as Layer 3 switches, Layer 7 rules, and so on. The concept of network layering is deeply rooted in the hearts of the people.

But have you ever thought about why the network should be layered? Is it indistinguishable? Before answering this question, let’s make an interesting assumption: this is the eve of the birth of the network, no IP protocol or TCP protocol exists, and you are the creator of the network, facing the great task of designing the network. Faced with such a good opportunity, what kind of design would you choose to do?

You basically have two options:

• The app does it all. You application converts the data of the application layer into binary data according to a certain encoding, and then the program controls the network card and sends the binary data to the network. During this period, the connection mode of communication, the reliability of transmission, the guarantee of speed and efficiency, etc., all need this program to be realized. Then next time you develop another application, do the above work again.