Chapter 18

Network Programming and Automation

New protocols, technologies, delivery, and operation and maintenance models

continue to emerge in the field of network engineering. Traditional networks are
challenged by new connectivity needs such as cloud computing and artificial
intelligence. Meanwhile, enterprises are also constantly seeking convenience, flex-
ibility and resiliency in their business. Against this backdrop, network automation is
becoming increasingly imperative.
Network programming and automation aim to simplify network engineers’ net-
work configuration, management, monitoring, operations and other tasks, and
improve the efficiency of their deployment, operation and maintenance. This chapter
is designed to guide network engineers through an initial understanding of how to
use Python programming to enable network automation.

18.1 Introduction to Network Programming

and Automation

The following are some classic network operation and maintenance scenarios that
you may have encountered in your work.
1. Device upgrade: there are thousands of network devices in the network, and you
need to upgrade the devices in a regular basis and in bulk.
2. Configuration audits: enterprises need to conduct configuration audits on devices,
for example, requiring all devices to be enabled with STelnet function and all
Ethernet switches to be configured with Spanning Tree security. You need to
quickly identify devices that do not meet the requirements.
3. Configuration changes: because of the network security requirements, the
accounts and passwords of devices need to be changed every 3 months. You
need to delete the original accounts and create new ones on thousands of network
devices.

© The Author(s) 2023 537

Huawei Technologies Co., Ltd., Data Communications and Network Technologies,
https://doi.org/10.1007/978-981-19-3029-4_18
538 18 Network Programming and Automation

Traditional network operation and maintenance require network engineers to man-

ually login to the network devices, manually view and execute configuration com-
mands, and screen the configuration results with their bare eyes. This working
method that heavily relies on “humans” is an enduring and inefficient operation
process, and the operation process is not audited.
Network automation, that is, using tools for automatic network deployment,
operation and maintenance, can gradually reduce the reliance on “humans”. This
can be a desirable solution to the traditional network operation and maintenance
problems. There are also many open-source tools for network automation, such as
Ansible, SaltStack, Puppet, and Chef.
From the perspective of building network engineering capabilities, it is more
recommended that engineers have programming skills that focus on network pro-
gramming. Network programming, in a broad sense, is the development of programs
to send and receive information over the network. The main task of network
programming is to assemble the information on the sending end through a defined
protocol, and to parse the packets according to the specified protocols at the
receiving end to extract the corresponding information for communication purposes.
The most important work in the process is packet assembly, packet filtering, packet
capture and packet analysis, and of course some processing is needed at the end. We
may come into contact with five parts, that is, code, development tools, database,
server setup and web design.
In recent years, with the rise of network automation technology, the abilities to
program, of which Python is a major one, become a new skill requirement for
network engineers. Automation scripts written in Python can perform repetitive,
time-consuming and regular operations excellently.
What can network automation do? One of the most intuitive examples is the
automated configuration of devices. We can break this process into two steps:
writing a configuration file and writing Python code to push the configuration file
to the device. This approach is easier for network engineers who are new to network
programming and automation to understand. This chapter focuses on how this
approach enables network automation. As shown in Fig. 18.1, first, write the script
using command lines, and then pass it to the device to run via Telnet/SSH.

18.2 Overview of Programming Languages

A computer programming language is the most important tool for programming. It is

a language with certain syntax rules that a computer can accept and process. Since
the birth of computers, computer languages have gone through several stages,
including machine languages, assembly languages and high-level languages.
18.2 Overview of Programming Languages 539

Fig. 18.1 Write a Python script to pass the configuration to the device for automated configuration

Fig. 18.2 High-level

language classification

18.2.1 Classification of High-Level Programming Languages

According to whether the language needs to be compiled before execution, high-

level languages can be classified into compiled languages that require compilation,
and interpreted languages that do not need compilation, as shown in Fig. 18.2.
1. Compiled language
Compiled languages have separate compilation and execution processes.
Before a program is executed, a compilation process is needed to compile the
source code into a binary file of machine language. The executable programs
generated by compiled languages do not need recompilation to run and they
directly use the compiled results, which makes them more efficient. However,
executable programs are platform-specific and cannot be executed in other
platforms. C, C++ and Go languages are typical compiled languages.
Figure 18.3 shows the process of compiled language from source code to
program: the source code needs to be compiled into machine instructions by
compiler and assembler, and then the machine language program is generated by
540 18 Network Programming and Automation

link library functions of the linker. The machine language program has to match
the instruction set of the CPU, and is loaded into memory by the loader when it is
run. Then the CPU executes the instructions. The source code of a compiled
language is compiled and converted into a format that the computer can execute,
such as .exe, .dll, and .ocx.
2. Interpreted languages
Interpreted languages do not require prior compilation, and the source code is
directly interpreted into machine code, so the program can be run as long as
the platform provides the corresponding interpreter. Interpreted languages require
the source code to be interpreted into machine code and executed every time the
program is run, which is inefficient. Both Python and Perl are typical interpreted
languages.
Figure 18.4 shows the process of interpreted languages from source code to
program: the source code file (Python file) is converted into a byte code file (.pyc
file) by an interpreter and then run on a Python virtual machine (Python
VM, PVM).

18.2.2 Computing Technology Stack and Program Execution

Process

As mentioned earlier, programming languages can be classified as machine lan-

guages, assembly languages and high-level languages. Machine languages consist of
binary machine code that can be directly recognized by computers. Since machine
language is obscure and difficult to understand, people have made a simple

Fig. 18.3 The process of compiling a language from source code to program

Fig. 18.4 The process of

interpreted languages from
source code to program
18.2 Overview of Programming Languages 541

Fig. 18.5 Computing technology stack and program execution process

encapsulation of hardware instructions consisting of 0s and 1s to make them easier to

be recognized and memorized by programmers (e.g., MOV, ADD), which is the
assembly language. Both of these languages are low-level languages, while lan-
guages such as C, C++, Java, Python, Pascal, Lisp, Prolog and Fortran are high-level
languages.
Among all programming languages, only the source programs written in machine
language can be directly understood and executed by computers; programs written in
other programming languages must first be “translated” into machine language
programs that computers can recognize using language processing programs.
Figure 18.5 illustrates the computational technology stack and the program
execution process.
In the computing stack shown in Fig. 18.5a, we can see the bottom layer of
hardware, consisting of physics and transistors for implementing gate circuits and
registers, which then constitute the CPU micro architecture, whose instruction set is
the interface of the hardware and the software. Applications use the instructions
defined in the instruction set to drive the hardware to perform the computation.
Applications are usually developed using high-level languages such as C, C++, Java,
Go and Python.
In the execution of a program shown in Fig. 18.5b, the high-level language first
needs to be compiled into assembly language and then converted into binary
machine code by the assembler according to the CPU instruction set. A program
exists on a disk in the form of a stack of binary machine code consisting of
instructions and data, which is also commonly referred to as a binary file.
542 18 Network Programming and Automation

18.3 Python Language

Python is a high-level object-oriented programming language. Programs written in

Python language are cross-platform, and Python can be seen everywhere from
clients, servers to Webs and mobile.

18.3.1 What Is Python

Python is an object-oriented interpreted computer programming language invented

by Dutchman Guido van Rossum in 1989, and the first publicly available version of
Python was released in 1991.
Python is a purely free software, whose source code and interpreter are under the
GPL (GNU General Public License) protocol. Its syntax is clean and clear, with the
mandatory use of White Space as a statement indent as one of its features.
Python has a rich and powerful library. It is often referred to as a glue language
that can easily integrate various modules made in other languages (especially C/C+
+). A common application scenario is to use Python to quickly generate a prototype
of a program (and sometimes even the final interface of the program) and then
rewrite parts of it that have special requirements in a more appropriate programming
language. For example, a graphics rendering module in a 3D game with particularly
high-performance requirements can be rewritten in C/C++ and then wrapped into an
extended class library that Python can call. It is important to note that you may need
to consider platform issues when using extended class libraries, and some libraries
may not provide cross-platform implementations.
Although Python source code files (.py) can be directly executed using Python
commands, Python does not actually interpret Python source code straightly;
instead, Python source code is compiled to generate Python byte code (the exten-
sions of byte code files are generally .pyc), which is then executed by the Python
Virtual Machine (PVM). The Python here is an interpreted language, meaning that it
interprets the Python byte code, rather than the Python source code. The basic idea of
this mechanism is the same as Java and .NET.
Although Python also has its own virtual machine, unlike the virtual machines of
Java or .NET, the Python virtual machine here is a more advanced virtual machine.
The “advanced” here is unlike the advanced in the usual sense, and it does not mean
the Python virtual machine is more powerful than Java or .NET virtual machines.
Instead, it means that Python virtual machine is farther away from the real machine
than Java or .NET. Or, you could say that the Python virtual machine is a more
abstract virtual machine.
The execution of the Python program source code is shown in Fig. 18.6.
1. Install Python and the runtime environment on the operating system.
2. Write the Python source code.
18.3 Python Language 543

Fig. 18.6 Python source code execution process

3. The interpreter runs Python source code and interprets it to generate a .pyc file
(byte code).
4. The Python virtual machine converts the byte code into the machine language.
5. The hardware executes the machine language.

18.3.2 Application Areas of Python

Python is a cross-platform programming language, and in theory, it can run on any

operating system platform. The current common operating system platforms are
Windows, Mac OS X and Linux.
Python is easy to learn, has various third-party libraries and a fast operation
speed, so it has an exceptionally wide range of applications. Python’s areas of
application mainly include the following.
• Linux/Unix operations and maintenance.
• Command-line program development.
• GUI program development (PyQt, Kivy, etc.).
• Web program development (Django and other frameworks).
• Mobile App development (PyQt, Kivy, etc.).
• Server program development (based on protocols such as Socket)
• Web crawlers (data sources for search engines, deep learning, etc.)
• Data analysis.
• Deep learning.
• Scientific computing.
Although not all of Python’s application areas are listed here, these listed areas alone
cover the vast majority of the development scenarios. Readers who have used Mac
OS X or Linux will find that the Python interpreter is already built into these two
operating systems, which means that Python programs can be run directly on Mac
OS X and Linux. Therefore, many operations and maintenance engineers are
accustomed to using Python for many automated operations. For the currently
popular deep learning, Python has become its first language. Therefore, from all
angles, whether you are a student, a programmer, a data analyst or a scientist, you
cannot live without Python. Python has become the world language of programming
languages.
This chapter focuses on how to use Python for network programming and to
manage network devices.
544 18 Network Programming and Automation

18.3.3 How Python Code Works

Python can run codes in two ways: interactive mode and scripted mode.
Programming in the interactive mode (i.e., interactive programming) does not
require the creation of a script file, and code can be written by the interactive mode of
the Python interpreter. Figure 18.7 illustrates the process of interactive programming
on a Windows system. [Note: in Fig. 18.7, Print( ) is a built-in Python function that
serves to output the contents in the parentheses.]
Code written in the scripted mode of programming (i.e., scripted programming)
can be run on a variety of Python interpreters or integrated development environ-
ments. For example, IDLE, Atom, Visual Studio, Pycharm, and Anaconda that come
with Python. A typical script-based programming process is shown in Fig. 18.8.
First, write a Python script using Notepad software, save the script file and change its
extension to .py, and then execute the script file in the Python interpreter.

Fig. 18.7 The interactive programming process

Fig. 18.8 The scripted programming process

18.3 Python Language 545

18.3.4 Coding Specifications of Python

Coding specifications are the naming rules, code indentation, code and statement
splitting and so on that should be followed when writing code in Python. Good
coding specifications help improve the readability of code and make it easier to
maintain and modify.
1. Suggestions for the use of semicolons, blank lines, parentheses and spaces.
Semicolons: Python programs allow semicolons at the end of lines, but it is not
suggested to use semicolons to isolate statements. It is recommended that each
statement occupy a separate line. If there are multiple statements on a line, they
can be separated by semicolons.
Blank lines: blank lines can be used to separate different functions or statement
blocks so as to distinguish between two pieces of code and improve the readabil-
ity of the code.
Parentheses: parentheses can be used to continue long statements, and unnec-
essary brackets are generally eliminated.
Spaces: it is not recommended to use spaces within the parentheses. You can
decide whether to add spaces on both sides of an operator according to your
personal habits.
2. Identifier naming conventions.
Python identifiers are commonly used to represent the names of constants,
variables, functions, and other objects. Identifiers usually consist of letters,
numbers and underlines, but they cannot begin with a number. Identifiers are
case-sensitive and do not allow renaming. If an identifier does not conform to the
naming convention, the compiler will output a SyntaxError error message when it
runs the code. As shown in Fig. 18.9, the fifth identifier starts with a number,
which is an incorrect identifier.
3. Code indentation.
When writing conditional and looping statements in Python, you need to use
the concept of code blocks. A code block is a set of statements that are executed
when certain conditions are met. In Python programs, code indentation can be
used to delimit a code block. If a code block contains two or more statements,
they must have the same amount of indentation, and the Python language uses
code indentation and colons to distinguish between levels of code. For Python,
code indentation is a syntax rule.

Fig. 18.9 Identifier naming specification

546 18 Network Programming and Automation

Fig. 18.10 Code blocks and indentations

When writing code, it is recommended to use four spaces for indentation. If the
wrong indentation is used in the code, an IndentationError message will be
returned when the program is run. The judgment statements shown in
Fig. 18.10 list the beginning and end of code blocks, as well as examples of
correct and incorrect indentation. The if line and the else line in the figure belong
to the same code block and have the same indentation. The last line print
(a) belongs to the same block as the if and else lines, so the indentation should
be the same.
4. Use comments
Comments are explanatory notes added to the program that can enhance the
readability of the program. As shown in Fig. 18.11, in Python programs, com-
ments are divided into single-line comments and multi-line comments. Single-
line comments begin with # and end at the end of the line. Multi-line comments
can contain multiple lines of contents that are contained within a pair of triple
quotes (“‘. . .”’ or “““. . .”””).
5. The structure of code file.
A complete Python source code file generally contains an interpreter and
encoding format declarations, documentation strings, module imports and
runtime code.
If you need to call classes from the standard library or other third-party
libraries in your program, you need to first use the “import” or “from . . . import”
statement to import the relevant module. The import statement is always located
at the top of the file after the module comment or documentation string. Fig-
ure 18.12 is an example of the structure of a source code file.
• The interpreter declaration is used to specify the path to the compiler that runs
this file (the compiler is installed via a non-default path or there are multiple
18.3 Python Language 547

Fig. 18.11 Comments

Fig. 18.12 Structure of

source code file

Python compilers). The interpreter declaration in the first line of this example
can be omitted on Windows operating systems.
• The encoding format declaration is used to specify the encoding type used by
this program and to read the source code in the specified encoding type.
Python 2 uses ASCII encoding by default (Chinese is not supported), and
Python 3 supports UTF-8 encoding by default (Chinese is supported).
• The documentation string serves as a general introduction to the program’s
functionality.
• The module import section imports the time module, which is a built-in Python
module that provides functions for handling time-related issues.
548 18 Network Programming and Automation

Fig. 18.13 Functions and modules

Fig. 18.14 Classes and methods

18.3.5 Basic Concepts of Python

1. Data types of Python

The most basic data types in Python are boolean (True/False), integers, floating-
point numbers and strings, and all data in Python (boolean, integers, floating point
numbers, strings and even large data structures, functions and programs) exists as
objects. This makes the Python language extremely uniform.
2. Functions and modules of Python
A function is an organized, reusable piece of code that improves the modu-
larity and code utilization of a program. Functions can be defined using the
keyword def.
A module is a saved Python file that can be composed of functions or classes.
The only difference between a module and a regular Python program is that they
serve different purposes. Modules can be called by other programs and therefore
usually do not have a main function. The definition and call of functions and the
import and call of modules are shown in Fig. 18.13.
18.4 Manage Network Devices with Python 549

3. Classes and methods of Python

A class is a description of a collection that has the same properties and
methods. Classes can be defined using the keyword class. The functions of an
instantiated class are called methods. When defining methods, the class must
carry the keyword self and it represents the created class instance itself. Fig-
ure 18.14 illustrates how a defining class and instantiable class call the instance.

18.4 Manage Network Devices with Python

18.4.1 Introduction to Telnet

Telnet defines the Network Virtual Terminal (NVT). It describes a standard repre-
sentation of data and command sequences transmitted over the Internet, thus
shielding differences across platforms and operating systems, such as different
commands for line breaks on different platforms. In order to distinguish Telnet
commands from ordinary data, Telnet uses escape sequences. Each escape sequence
consists of two bytes. The first byte is 0xFF, called IAC (Interpret as Command,
which is an escape character indicating that the byte following the character is the
command code); and the second byte is the code of the command to be executed.
Telnet can be used on Windows or Linux systems to remotely configure network
devices such as Huawei routers and switches.
The telnetlib is a module in the Python standard library. It provides a class
telnetlib.Telnet that implements the functions of Telnet. Table 18.1 shows the

Table 18.1 Methods of telnetlib module

Methods Functions
Telnet.open(host,port ¼ 0 [, Connect to a host. The optional second argument is the port
timeout]) number, which defaults to the standard Telnet port (23).
The optional timeout parameter specifies a timeout in seconds for
blocking operations like the connection attempt (if not specified,
the global default timeout setting will be used).
Telnet.read_until(expected, Read until a given byte string (b”) expected is encountered or
timeout ¼ None) until timeout seconds have passed.
When no match is found, return whatever is available instead,
possibly empty bytes. Raise EOFError if the connection is closed
and no cooked data is available.
Telnet.read_all( ) Read all data until EOF as bytes; block until connection closed.
Telnet.read_very_eager( ) Read everything that can be without blocking in I/O (eager),
return the byte string
Raise EOFError if connection closed and no cooked data
available.
Telnet.write(buffer) Write a byte string to the socket, doubling any IAC characters.
This can block if the connection is blocked. May raise OSError if
the connection is closed.
Telnet.close( ) Close the connection.
550 18 Network Programming and Automation

methods defined by the telnetlib module, where different functions can be enabled
by calling different methods of the Telnet class of the telnetlib module.

18.4.2 Manage Huawei Routers with Telnet

This case will show how to import the telnetlib module using Python script file, how
to configure Huawei router via Telnet and how to change the device name, create
VLAN, and set the interface IP address of a Huawei router.
1. First, configure the interface IP address of the Huawei router.

Enter system view, return user view with Ctrl+Z.

[Huawei]interface GigabitEthernet 0/0/0
[Huawei-GigabitEthernet0/0/0]ip address 192.168.80.99 24
[Huawei-GigabitEthernet0/0/0]quit
2. Configure the router to allow Telnet.

[Huawei]user-interface vty 0 4
[Huawei-ui-vty0-4]authentication-mode password
Please configure the login password (maximum length 16):huawei@123
[Huawei-ui-vty0-4]user privilege level 15
[Huawei-ui-vty0-4]quit
3. Use Telnet to login to the router on Windows, and observe the interaction
process.

C:\Users\hanlg>telnet 192.168.80.99
Login authentication
Password:
<Huawei>system-view
Enter system view, return user view with Ctrl+Z.
[Huawei]quit
4. Based on the Telnet interaction in the previous step, write a Python script to read
the Telnet output using the telnetlib module, and then enter the Telnet command
to configure the network device.

import telnetlib #Import telnetlib module

host = '192.168.80.99' #Specify the IP address to login to the
device
password = 'huawei@123' #Specify the password to login to the
device
tn = telnetlib.Telnet(host) #Login to the device via Telnet
tn.read_until(b'Password:') #Read until the echo message (the
message returned by the device) is ”Password:”
18.5 Exercises 551

tn.write(password.encode('ascii')+b'\n') #Enter the password of

code ASCII and enter a new line
#Enter the system view, and change the device name
tn.read_until(b'<Huawei>') #Output and read until the message
“<Huawei>”
tn.write(b'system-view'+b'\n') #Enter the command system and
enter a new line
tn.read_until(b'[Huawei]') #Output and read until the message
“[Huawei]”
tn.write(b'sysname R1'+b'\n') #Change the router name to R1
tn.read_until(b'[R1]') #Read until the echo message “[R1]”
#Create VLAN 2
tn.write(b'vlan 2'+b'\n') #Enter the command to create vlan 2
tn.read_until(b'vlan2]') #Output and read until the message
“vlan2]”
tn.write(b'quit'+b'\n') #Enter the command to exit the vlan2
view
tn.read_until(b'[R1]') #Output and read until the message
“[R1]”
#Enter the interface view, and configure the interface IP address
tn.write(b'interface GigabitEthernet 0/0/1'+b'\n') #Enter the
command and enter the interface configuration mode
tn.read_until(b'1]') #Output and read until the message “1]”
tn.write(b'ip address 10.1.1.1 24'+b'\n') #Enter the command to
configure the interface IP address and subnet mask
tn.read_until(b'1]') #Output and read until the message “1]”
tn.close() #Close the Telnet connection

The encode() and decode() functions of Python are used to encode and decode
strings in the specified way. In this case, password.encode(’ ascii’) means to convert
the string huawei@123 to ASCII. Here the encoding format follows the official
requirements of the telnetlib module.
Python prefixes the string with a b, such as b‘string’, to convert the string to bytes.
In this case, b‘Password:’ indicates that the string Password: is converted to a byte-
type string. The encoding format here follows the official requirements of the
telnetlib module.

18.5 Exercises

1. Python is a compiled language. ( )

A. Correct
B. Incorrect
2. Briefly describe the naming specification for Python identifiers.
552 18 Network Programming and Automation

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivatives 4.0 International License (http://creativecommons.org/licenses/by-
nc-nd/4.0/), which permits any noncommercial use, sharing, distribution and reproduction in any
medium or format, as long as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons license and indicate if you modified the licensed material.
You do not have permission under this license to share adapted material derived from this chapter or
parts of it.
The images or other third party material in this chapter are included in the chapter’s Creative
Commons license, unless indicated otherwise in a credit line to the material. If material is not
included in the chapter’s Creative Commons license and your intended use is not permitted by
statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder.