DEV Community: AkashKola

Building a Simple TCP Client in C: A Ground-Up Networking Exercise

AkashKola — Wed, 30 Apr 2025 22:24:10 +0000

🔌 Part 1: Building a Simple TCP Client in C

As engineers, we often work several layers above the network stack—abstracted by frameworks, SDKs, and libraries. Recently, I decided to go back to basics and explore socket programming in C by building a simple TCP client from scratch.

The goal? Connect to a server send a raw HTTP request, and read the response—all using low-level system calls.

🧠 What I Learned (Client Side)

Creating a socket

Used socket(AF_INET, SOCK_STREAM, 0) to create a TCP socket with IPv4.

Preparing the server address

inet_addr() to convert IP string to binary
htons() to convert port to network byte order

Connecting to the server

Used connect() with a typecast to struct sockaddr*
Passed in a sockaddr_in struct with Server’s IP and port 80

Sending and receiving data

Used write() to send an HTTP GET request
Used read() to fetch the response from the server

Man pages were essential

Especially man 2 socket, man 2 connect, man 2 read, etc.

🧪 Sample Code

/* tcp_client.c */
#include <stdio.h>

#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>

#include <unistd.h>

#include <string.h>

#define IP "142.250.70.100"
#define PORT 80

int main()
{
        int s;
        struct sockaddr_in sserver;
        char* data;
        char buf[10240];

        data = "GET /webhp HTTP/1.0\n\n";

        s = socket(AF_INET, SOCK_STREAM, 0);
        if(s < 0)
        {
                printf("socket()\n");
                return -1;
        }

        sserver.sin_addr.s_addr = inet_addr(IP);
        sserver.sin_port = htons(PORT);
        sserver.sin_family = AF_INET;

        if(connect(s, (struct sockaddr*)&sserver, sizeof(sserver)) != 0)
        {
                printf("connect()\n");
                close(s);
                return -1;
        }

        write(s, data, strlen(data));
        read(s, buf, 10239);
        close(s);

        printf("\n%s\n", buf);
        return 0;
}

🖥️ Part 2: Writing a Simple TCP Server in C

In Part 1 (link to Part 1), I shared how I built a simple TCP client in C to connect to a server, send an HTTP request, and read the response. This time, I flipped the role—writing a TCP server from scratch that listens on port 8181, accepts client connections, and responds with a minimal HTTP reply.

🔄 What the Server Does

This TCP server:

Listens on 0.0.0.0:8181
Accepts incoming TCP connections
Reads the client’s request
Responds with a static HTTP response (hello)
Sleeps for 2 minutes (120s) before replying—to simulate a long-running process

🧠 What I Learned (Server Side)

Creating and binding a server socket

Used socket(AF_INET, SOCK_STREAM, 0) to create a TCP socket
Used a sockaddr_in struct to bind the server to port 8181 on all interfaces using bind()

Listening for incoming connections

Called listen(s, 1) to queue one pending connection
Interestingly, the backlog queue didn’t behave as expected—still exploring why that might be OS or implementation dependent

Accepting clients using accept()

Accepts a new connection and returns a client socket descriptor
accept() also fills in the client’s socket address.

Simple client-server interaction

Read the incoming request using read()
Delayed for 2 minutes to simulate latency
Replied with a static HTTP/1.1 200 OK response using write()

🧪 Sample Code

/* tcp_server.c */
#include<stdio.h>

#include <sys/socket.h>
#include <arpa/inet.h>
#include <unistd.h>

#include <string.h>

#define PORT 8181

int main()
{
        int c, s;
        struct sockaddr_in cli, srv;
        char buf[1024];
        char* data;
        socklen_t addrlen;

        addrlen = 0;

        data = "HTTP/1.1 200 OK\nContent-Type: text/html\nContent-Length: 5\n\nhello";

        memset(&cli, 0, sizeof(cli));
        memset(&srv, 0, sizeof(srv));

        s = socket(AF_INET, SOCK_STREAM, 0);

        if(s == -1)
        {
                printf("socket()\n");
                return -1;
        }

        srv.sin_family = AF_INET;
        srv.sin_port = htons(PORT);
        srv.sin_addr.s_addr = 0;

        if(bind(s, (struct sockaddr*)&srv, sizeof(srv)))
        {
                printf("bind()\n");
                close(s);
                return -1;
        }

        if(listen(s, 1))
        {
                printf("listen()\n");
                close(s);
                return -1;
        }

        printf("Listening on 0.0.0.0:%d\n", PORT);

        while(1)
        {
                addrlen = 0;
                c = accept(s, (struct sockaddr *)&cli, &addrlen);
                if(c == -1)
                {
                        printf("accept()\n");
                        close(s);
                        return -1;
                }

                printf("client connected\n");

                read(c, buf, 1023);

                sleep(120);

                write(c, data, strlen(data));

                close(c);
        }

        close(s);
        return 0;
}

Thanks to Jonas Birch for his excellent instructional videos. His ability to explain system-level concepts clearly was a huge help during this learning phase.

Network Address

AkashKola — Mon, 28 Feb 2022 05:57:15 +0000

Before going to our main topic Network Address, let's discuss about converting ip address dotted notation to a single decimal number.
This is the second post of networking I am sharing
Here is the first post link : Post1

Let's see how to convert ip address dotted notation to a single decimal number.
I think we all know how to convert binary number into a decimal number same process applies here also.
Example : let's convert binary number 1010 into decimal number.



  1      0       1       0

 2^3    2^2     2^1     2^0

= 1*8 + 0*4 + 1*2 + 0*1 

= 10

Here we use our base as 2 because binary Language has only two options either 0 or 1
But in Ip address, we can have 0 to 255 so our base is 256.
Let's convert 192.168.31.1 into single decimal.

  192      168      31     1
256^3     256^2   256^1  256^0

192*256^3 + 168*256^2 + 31*256^1 + 1*256^0 = 3232243457

192.168.31.1 = 3232243457
We can use ip address as 192.168.31.1 or 3232243457 both are same.

Soo lets go to our main topic Network Address
The laptops, desktops, routers, switches etc.. are called network devices

Network devices having same network address fall into same network named as local area network (LAN)
We identify the network address using netmask
Ip : 192.168.31.1
Netmask : 255.255.255.0
In binary
Ip : 11000000 10101000 00011111 00000001
Netmask : 11111111 11111111 11111111 00000000

So we see continuous 1's in netmask, it's corresponding ip address bits are known as Network address.

In our example
11000000 10101000 00011111
11111111 11111111 11111111
These are the corresponding bits

Place 0's in last remaining bits, So our network address is
11000000 10101000 00011111 00000000
192.168.31.0

Thank you for reading the post
Follow me in LinkedIn and dev.to and share this post to your friends....

Computer Networking

AkashKola — Tue, 01 Feb 2022 05:54:26 +0000

Today let's discuss about networking which is still essential for any technology we learn.

What is Computer Networking ?

We all know Computer Networking is all about creating communication between computers to share resources.

Let's see what google definition of computer networking is

A computer network is a set of computers sharing resources located on or provided by network nodes.

So here network nodes means computers, switches, routers etc..
Let's see what are switches and routers in soon.

How Networking is possible ?

So to get our network establish, we have to make sure of having some steps

Connection between computers (using wired/wireless).
Unique IP Address to computers.
Same Network Address for basic communication (we can also have different network addresses but we have to use router to communicate, let's see that in future posts).

We all know about creating connection between two computers using WIFI / Ethernet Lan
So Let's see about IP Address
So IP address is used to uniquely identify the computer in a
network. For example we can imagine our phone numbers as IP
addresses which are unique.
What is IP address format?
Actually an IP address is any positive integer having 4 bytes of
size.
So for example 100 is an valid IP address and 212021241 is also
an valid IP address.
But In real life we won't see IP addresses as 212021241 and 100
because we use some format which is easy to remember our IP
addresses.
So let's see how 212021241 is formatted into an IP address
format.
convert 212021241 into 32 bit binary number
00001100101000110010111111111001
Here is the binary representation.
Let's make it four parts
00001100
10100011
00101111
11111001
Convert each part into decimal
00001100 -> 12
10100011 -> 163
00101111 -> 47
11111001 -> 249
so we will use 12.163.47.249 instead of 212021241 this large
number.
so what is the proof right ?
open your chrome browser and type http://212021241 in search bar

it will automatically converts that into http://12.163.47.249

So Let's see about Network Address in next post...