The Internet & TCP/IP—The Language of the Digital World

Introduction: Why Do We Need a Global Network?

Imagine living in a world where every country had its own version of the internet, completely isolated from others. A website in the U.S. wouldn’t be accessible in Europe, and sending an email from London to Tokyo would be impossible.

This was the reality before the internet—a time when computers were standalone machines, disconnected from each other. Then came the revolutionary idea: a global network where all computers could communicate, regardless of their location, manufacturer, or operating system.

The internet didn’t appear overnight. It started as a Cold War defence project, grew into a network for scientists, and finally became the backbone of modern communication.

By the end of this lesson, you will fully understand:

1. How the internet was born and why it evolved.
2. How data travels across the internet using the TCP/IP protocol stack.
3. The major internet services and protocols that allow the web, email, and apps to function.

Let’s begin this journey by going back to where it all started—a world on the brink of war, with a need for a network that could survive destruction.

---

The Birth of the Internet: From Military Experiment to Global Revolution

Circuit Switching vs. Packet Switching: The Battle of Two Ideas

Before we dive into the internet’s origin, we need to understand a fundamental question: How do computers send data to each other?

For most of history, networks used circuit switching – a method where a dedicated physical connection was established between two devices before any data could be sent.

Problems with circuit switching:

1. Wasted resources — If no one is talking, the line remains idle but reserved.
2. Vulnerable to failure — If the connection is broken, communication stops immediately.
3. Not scalable — In a growing network, dedicated circuits quickly become inefficient.

A better alternative emerged: packet switching. Instead of maintaining a dedicated link, data is broken into smaller packets, sent independently, and reassembled at the destination.

The ARPANET: The Internet’s First Building Blocks

In the late 1960s, the U.S. Department of Defence (DoD) saw the potential of packet switching for military communications. They funded ARPANET (Advanced Research Project Agency Network) to create a robust, decentralised communication network.

Here’s how it unfolded:

The First Message Ever Sent on the Internet

The first internet message was sent on October 29, 1969, from UCLA to Stanford Research Institute. The goal was to send the word “Login“, but the system crashed after transmitting only the first two letters:

Interface Message Processors (IMPs): The First Internet Routers

Early ARPANET computers weren’t connected directly. Instead, they used specialised minicomputers called Interface Message Processors (IMPs)—the first internet routers.

An experimental network with four nodes was set up in December 1969 at:

1. The University of California, Los Angelos (UCLA).
2. The University of California, Santa Barbara (UCSB).
3. Stanford Research Institute
4. University of Utah

Each IMP had to be connected to at least one other IMP, ensuring that no single failure could take down the network.

---

The TCP/IP Protocol Stack: The Internet’s Universal Language

Why TCP/IP?

As ARPANET grew, it became clear that different networks needed a common way to communicate. Vinton Cerf and Robert Kahn developed TCP/IP (Transmission Control Protocol / Internet Protocol)—the language of the internet.

The Four Layers of TCP/IP

Layer	What It Does	Example Protocols
1. Network Access	Handles physical transmission of data.	Ethernet, Wi-Fi
2. Internet	Routes packets between networks.	IP, ICMP, ARP
3. Transport	Ensures reliable data delivery.	TCP, UDP
4. Application	Provides services to users.	HTTP, FTP, DNS

Think of TCP/IP like a postal system:

• The Transport Layer ensures safe delivery (TCP).
• The Application Layer is the post office counter, where you send and receive mail (websites, email, etc.).

---

Essential Internet Services & Protocols

DNS: The Internet’s Phonebook

Every website (host within the network, i.e., a machine or a node) has an IP address, but humans prefer names like www.google.com. The Domain Name System (DNS) translates domain names into IP addresses.

DNS works as a look-up table, while TCP/IP was introduced to handle communication over different networks, i.e., ARPANET and NSFNET networks. It used to be referred to as the interconnected network, hence the name internet.

DHCP: Assigning IP Addresses Automatically

When you connect to Wi-Fi, your device needs an IP address. Dynamic Host Configuration Protocol (DHCP) assigns one automatically.

Example: Instead of manually entering an IP address, DHCP does it for you, ensuring hassle-free connectivity.

HTTP & HTTPS: The Web’s Foundation

• HTTP (Hypertext Transfer Protocol) allows browsers to fetch webpages.
• HTTPS (Secure HTTP) adds encryption, making websites secure.

---

Summary: What Did We Learn So Far?

1. Circuit switching was replaced by packet switching, enabling more efficient data transfer.
2. ARPANET was the first step, leading to today’s internet.
3. TCP/IP became the universal standard, allowing different networks to communicate.
4. DNS, DHCP, HTTP, and HTPPS are essential for web browsing and online communication.

---

Now that we understand the basics of how the internet was born and what the TCP/IP is, let’s jump into further details.

Why Do We Need a Standard Model?

Imagine you’re sending a parcel to a friend across the world. You don’t just drop it in a mailbox and hope for the best. The postal system follows a structured process:

1. Packing: You put items inside a box.
2. Addressing: You write your friend’s address on the package.
3. Routing: The postal system figures out the best way to get it there.
4. Delivery Confirmation: The recipient acknowledges receiving the parcel.

The internet works the same way. If every device had its own rules, computers wouldn’t be able to talk to each other. That’s why TCP/IP comes in.

Now, let’s break it down layer by layer.

---

The TCP/IP Reference Model: The Four Layers

We’ve established earlier that the TCP/IP model consists of four layers, each responsible for a specific function in the data transmission.

Layer	Purpose	Example Protocols
Application Layer	Provides services to applications	HTTP, FTP, DNS, SMTP, RTP, Telnet
Transport Layer	Ensures reliable or fast data transfer	TCP, UDP
Internet Layer	Handles addressing and routing	IP, ARP, RARP, ICMP, IGMP
Network Access Layer	Deals with physical transmission	Ethernet, Wi-Fi, PPP, DSL

Now, let’s explore each layer in detail and uncover the protocols they host.

---

Network Access Layer: The Foundation of Connectivity

What it Does

This layer is responsible for the physical transmission of data over cables, fibre optics, or wireless signals. It ensures that bits (1s and 0s) are converted into signals that can travel through a network medium.

Key Technologies in This Layer

• Ethernet: The most common wired LAN technology.
• Wi-Fi (802.11): Wireless communication for local networks.
• PPP (Point-to-Point Protocol): Used for direct connections between two devices (e.g., dial-up).
• DSL (Digital Subscriber Line): Internet over telephone lines.

How Data Moves in This Layer

Data at this layer exists as raw electrical, optical, or radio signals that represent binary bits. The Network Interface Card (NIC) in each device transmits and receives these signals.

Example: When you plug in an Ethernet cable, the Network Access Layer handles the physical connection, allowing your device to send and receive data.

---

Internet Layer: The Backbone of Internet Routing

The Internet Layer is where data gets an IP address and finds its way to the correct destination. Think of it as a GPS system for network packets.

Protocols in This Layer

Protocol	Function
IP (Internet Protocol)	Assigns and routes IP addresses
ARP (Address Resolution Protocol)	Maps IP addresses to physical MAC addresses
RARP (Reverse Address Resolution Protocol)	Converts MAC addresses back to IP addresses
ICMP (Internet Control Message Protocol)	Sends error messages (e.g., unreachable hosts).
IGMP (Internet Group Management Protocol)	Manages multicast group memberships

How IP Addresses Work

Every device on network needs a unique IP address to be identified. Think of an IP address like a home address—it tells the network where to send data.

IPv4 vs. IPv6

• IPv4 (e.g., 192.168.1.1): 32-bit address, supports ~4.3 billion devices.
• IPv6 (e.g., 2001:db8::ff00:42:8329): 128-bit address, supports trillions of devices.

The massive scale of IoT (Internet Of Things) devices necessitates the movement towards IPv6. According to IoT analytics, there were 18.8 billion connected IoT devices by the end of 2024 and it is projected to exceed 40 billion by 2030. Given the limitation of the number of addresses supported by IPv4 (4.3 billion devices only), the need to move to IPv6 is imminent.

ARP and RARP: Finding Devices on a Network

• ARP (Address Resolution Protocol) helps find the MAC (Physical) address of a device when you only have its IP address.
• RARP (Reverse ARP) does the opposite—it finds the IP address of a device when you only have its MAC address.

Example: When you type www.google.com, your computer doesn’t immediately know Google’s IP address. It first asks a DNS server for Google’s IP and then uses ARP to find the nearest router’s MAC address.

---

Transport Layer: Ensuring Reliable Communication

The Transport Layer is responsible for end-to-end communication. it ensures that data reaches the correct application on the receiving device.

Protocols in This Layer

Protocol	Reliable?	Speed	Usage
TCP (Transmission Control Protocol)	Yes, ensures all data is received	Slower	Web browsing, emails, file transfers
UDP (User Datagram Protocol)	No, sends data fast without checking	Faster	Streaming, gaming, VoIP

How TCP Works

TCP provides error-checking, sequencing, and retransmission to ensure all packets arrive correctly.

Example: When you download a file, TCP ensures every part of the file arrives in the correct order. If a piece is missing, it requests a retransmission.

How UDP Works

UDP doesn’t check for errors—it just sends data as fast as possible.

Example: When you’re watching a live stream, you don’t want delays. Even if some packets are lost, the video keeps playing. UDP is perfect for this because speed is more important than reliability.

In the OSI Reference Model, the TCP and UDP protocols are referred to as Connection-Oriented and Connectionless services.

---

Application Layer: The Services We Use Every Day

The Application Layer is where user interacts with the network. It provides protocols that enable web browsing, email, file transfers, and more.

Major Protocols in This Layer

Protocol	Purpose
HTTP/HTTPS	Web browsing (HTTP is insecure, HTTPS is secure)
SMTP (Simple Mail Transfer Protocol)	Sending emails
POP3/IMAP	Retrieving emails
DNS (Domain Name System)	Converts domain names to IP addresses
FTP (File Transfer Protocol)	Transfers files between devices
RTP (Real-Time Transport Protocol)	Handles audio and video streaming
Telnet	Provides remote access to another computer

How it Works in Action

Let’s say you send an email using Gmail, here’s what happens:

1. SMTP sends the email to the recipient’s mail server.
2. The recipient retrieves the email using POP3 or IMAP.
3. If the email contains images, they might be fetched using HTTP or HTTPS.

---

IP-Based Protocols and Services

Now, let’s learn more about some of the IP-Based Protocols and Services.

Connection-Oriented vs. Connectionless Communication

Before we dive into IP-based protocols, let’s first understand two fundamental ways that data is transmitted over networks:

Connection-Oriented Communication

This type of communication establishes a dedicated connection between the sender and receiver before transmitting data.

Analogy: Think of a telephone call—before speaking, you first establish a connection by dialling a number.

Key Features:

• Reliable communication (ensures data is received in the correct order).
• Data follows a fixed route.
• Slower due to the connection setup process.

Example Protocols:

• TCP (Transmission Control Protocol): Ensures error-free, in-order data delivery.
• SCTP (Stream Control Transmission Protocol): Used for VoIP and telephony services..

---

Connectionless Communication

In contrast, connectionless communication does not establish a fixed path—each data packet is sent independently.

Analogy: Think of sending a bunch of letters through the postal service—each letter can take a different route, and they might not arrive in order.

Key Features:

• Faster because no prior connection is needed.
• Packets can take different routes.
• Less reliable (packets might arrive out of order).

Example Protocols:

• IP (Internet Protocol): The foundation of the internet, responsible for routing packets.
• UDP (Used Datagram Protocol): Used in streaming, gaming, and real-time communication where speed matters more than reliability.

Why does this matter?

If you need speed over accuracy (like a Zoom call), use UDP.

---

IPv4 Datagram Structure: Anatomy of an IP Packet

In IPv4, data packets are called datagrams. Every IP datagram consists of a header (with control information) and a payload (actual data).

IPv4 Datagram Format

Field	Size (Bits)	Purpose
Version	4	IPv4 or IPv6 identification.
Header Length (HLEN)	4	Specifies the size of the header.
Type of Service (TOS)	8	Specifies priority and quality of service.
Total Length	16	The total size of the packet (header + data).
Identification	16	Controls fragmentation behaviour.
Flags	3	Helps in packet fragmentation and reassembly.
Fragment Offset	13	Specifies where a fragment belongs in a reassembled packet.
Time to Live (TTL)	8	Limit how long a packet can exist.
Protocol	8	Specifies the higher-layer protocol (TCP, UDP, etc).
Header Checksum	16	Error detection mechanism.
Source IP Address	32	The sender’s IP address.
Destination IP Address	32	The receiver’s IP address.
Options	Variable	Additional settings (rarely used).
Payload	Variable	The actual data being transmitted.

How Packets Travel in a Network

1. The Source IP Address identifies who is sending the packet.
2. The Destination IP Address determines where it’s going.
3. The TTL field prevents packets from looping endlessly (each router reduces the TTL by 1; if TTL reaches 0, the packet is discarded).

---

Type of Service (TOS) and Differentiated Services

The Type of Service (TOS) field is an 8-bit field that helps networks prioritise packets. Over time, this field evolved into Differentiated Services (DiffServ) to support modern Quality of Service (QoS) applications.

Breaking Down the TOS Field

Bits	Name	Function
0 0 0 0	Normal (Default)	Standard data transmission.
0 0 0 1	Minimise Cost	Uses cheaper routes (e.g., background updates).
0 0 1 0	Maximise Reliability	Prefers stable connections (e.g., file downloads).
0 1 0 0	Maximise Throughput	Prioritises high-speed data (e.g., video streaming).
1 0 0 0	Minimise Delay	Low-latency routing (e.g., online gaming, VoIP).

Example: A WhatsApp call would use Minimise Delay, while a software update might use Minimise Cost.

Differentiated Services (DiffServ) and Codepoint Subfield

The first six bits of the TOS field are now called the DSCP (Differentiated Services Code Point).

• If the three least significant bits (LSB) are 0, the three most significant bits (MSB) act as precedence bits.
• If the LSB are not 0, the six-bit DSCP value defines 64 service classes, assigned as:

Category	Code Point	Assigned By
Standard Services	XXXXX0	Assigned by IETF (The Internet Engineering Task Force)
Local Services	XXXX11	Assigned by local networks
Experimental Services	XXXX01	Used for testing

Why DiffServ Important? It ensures efficient network performance for critical applications.

---

Total Length Calculation

The Total Length field (16 bits) defines the entire packet size, including both the header and data.

Formula

Total Length = Header Length + Data Length

• The maximum total length is 65,535 bytes.
• If a packet exceeds the Maximum Transmission Unit (MTU) of a network (e.g., 1500 bytes for Ethernet), it is fragmented.

Example: A packet with a 20-byte header and a 1000-byte payload would have a total length of 1020 bytes.

Header Length Calculation

The Header Length (HLEN) field is a 4-bit field that defines the header size in 4-byte words.

HLEN Value	Header Size
5	20 bytes (minimum)
15	60 bytes (maximum)

Note: When the optional field is empty, the size of the header is 20 bytes and the value of the field is five (as 5 · 4 = 20). When the optional field is fully used, the size of the header is 60 bytes and the value of the field is 15 (15 · 4 = 60).

---

ARP and RARP: How Devices Find Each Other on a Network

Let’s now zoom in on one of the most critical -and often misunderstood- parts of networking: the Address Resolution Protocol (ARP) and the Reverse Address Resolution Protocol (RARP). These two protocols sit at the intersection of the Internet Layer and Network Access Layer of the TCP/IP model and ensure that devices can actually find each other on the same local network.

Why Do We Need ARP and RARP?

Imagine you’re at a conference and someone says, “Go give this folder to Thomas.” You know Thomas’ name, but you wouldn’t know what they look like. You need some way to match their name to their face. So, you ask around: “Hey, who is Thomas?”

That’s exactly what ARP does in a computer network.

In a similar way, RARP resolves the opposite problem. Let’s say you have a picture of Thomas, but you forgot their name. You ask: “Who is this person?”.

In networking terms:

• ARP: “I have an IP address. What’s the corresponding MAC address?”
• RARP: “I have a MAC address, What’s the corresponding IP address?”

Let’s break these down in detail.

ARP: Address Resolution Protocol

What Problem Does ARP Solve?

Every device in a local network has:

• A logical address: IP address (e.g., 192.168.1.10)
• A physical address: MAC address (e.g., 00:0a:95:9d:68:16)

The IP address tells us who to send data to, but the actual transmission happens using the MAC address. So before a device can send data to another on the same network, it needs to find out that device’s MAC address.

That’s where ARP comes in.

How Does ARP Work?

Let’s say your computer wants to send a message to 192.168.1.15.

Step-by-step ARP resolution:

1. Your computer checks its ARP cache. if the IP-to-MAC mapping is stored, it uses it.
2. If not found, it broadcasts an ARP Request: “Who has IP address 192.168.1.15? Tell me!”.
3. The device with that IP replies with an ARP Reply: “I’m 192.168.1.15, and my MAC address is AA:BB:CC:DD:EE:FF”.
4. Your computer stores this info in its ARP cache and sends the data to the MAC address.

ARP Packet Structure

Field	Purpose
Hardware type	Specifies network type (e.g., Ethernet = 1)
Protocol type	Specifies the protocol (e.g., IP = 0x0800)
Hardware size	Length of MAC address (typically 6)
Protocol size	Length of IP address (typically 4).
Opcode	1 = ARP Request, 2 = ARP Reply
Sender MAC address	MAC of requesting device
Sender IP address	IP of requesting device
Target MAC address	Empty (in request), filled in (in repl)
Target IP address	IP being queried

Key point: ARP only works within a local subnet. If the destination IP is on a different network, ARP finds the MAC address of the default gateway (router) instead.

Types of ARP

• Static ARP: Manually configured IP-MAC pairs.
• Dynamic ARP: Automatically learned via ARP Requests/Replies.
• Gratuitous ARP: A device sends an ARP request for its own IP to check if someone else is using it (IP conflict detection).
• Proxy ARP: A router responds to an ARP request on behalf of another device, ofter across subnets.

---

RARP: Reverse Address Resolution Protocol

What Problem Does RARP Solve?

Let’s rewind to a time before DHCP existed. Devices (especially diskless workstations) didn’t know their IP address at boot. But they did know their MAC address – because it has hardwired into their network interface card (NIC).

They needed a way to say: “Here’s my MAC address. Can someone tell me what IP I should use?”.

That’s what RARP was designed for.

How Does RARP Work?

1. A device sends a RARP Request to a RARP server: “This is my MAC address. What is my IP address?”.
2. The RARP server checks a pre-configured table and replies with the matching IP address.

Limitations of RARP

• RARP requires a RARP server with static mappings.
• It can only supply the IP address, not the subnet mask, gateway, or DNS servers.
• It operates only at the data link layer, so it cannot use routing.

Because of these limitations, RARP is now considered obselete and has been replaced by:

• BOOTP (Bootstrap Protocol), which added more configuration options.
• DHCP (Dynamic Host Configuration Protocol), which dynamically assigns full IP settings.

Comparison: ARP vs. RARP

Feature	ARP	RARP
Purpose	Find MAC from IP	Find IP from MAC
Layer	Internet Layer	Internet Layer
Packet Type	Broadcast (Request) + Unicast (Reply)	Broadcast
Still used today?	Yes, widely used	No, replaced by DHCP
Use Case	Local communication	Early device boot-up

Bonus: Why These Protocols Matter

Thought-provoking question:

If a network can route global traffic using IP addresses, why bother with a MAC address at all?

Answer:

Because the final delivery of packets still relies on hardware-level addressing. IP routing gets the packet to the correct network, but it’s the MAC address that delivers it to the exact device.

Without ARP, devices on a local network would be blind, unable to complete the journey from logical to physical delivery.