Computer Networking Basics

What Is a Computer Network?

A computer network is a group of interconnected devices that communicate with each other to share data, resources, and services. These devices — called nodes — can be computers, smartphones, printers, servers, routers, or any other device capable of sending and receiving data.

Every time you send an email, stream a video, or open a website, your device is exchanging data with one or more remote devices over a network. The connection might travel through a copper cable, a fiber-optic line, or a wireless radio signal, but in every case the same basic idea applies: two or more devices agree on a way to exchange information, and a physical or wireless medium carries that information between them.

A computer network is built from three core components:

• Nodes — the devices connected to the network, such as computers, printers, servers, and mobile phones.
• Communication links — the medium that carries data between nodes, such as Ethernet cables, fiber-optic cables, or Wi-Fi radio signals.
• Protocols — the agreed-upon rules that define how devices format, send, and interpret data. Without a shared protocol, two connected devices would have a physical link but no way to understand each other — similar to two people sharing a phone line but speaking different languages.

Why Do We Need Computer Networks?

Before networks were common, each computer worked in isolation: files had to be copied manually between machines, and every device that needed to print required its own printer. Networks solve this by letting devices share data and hardware directly, which makes organizations faster, more efficient, and easier to manage. The main benefits fall into a few categories.

Data Sharing

A network lets multiple users access and work with the same files, applications, and databases instead of keeping separate copies on each machine.

Example: In an office, employees can open and edit documents stored on a shared company server rather than emailing copies back and forth. This also avoids the problem of different people editing outdated versions of the same file.

Resource Sharing

Networks allow expensive or specialized hardware to be shared across many users instead of duplicated for each one.

Example: A single network printer or a shared scanner can serve an entire office, rather than every desk needing its own device.

Communication

Networks are the foundation for nearly every modern communication tool, including:

• Email
• Instant messaging
• Video conferencing
• Voice calls (VoIP)
• Social media

Applications like video calls and messaging apps only work because the underlying network can deliver small packets of data between participants with low delay.

Centralized Data Management

Instead of scattering data across individual machines, organizations store it centrally on network-connected servers or cloud platforms. Centralizing data this way makes several tasks easier:

• Backup — a single system can back up data for the whole organization.
• Security — access can be monitored and controlled from one place.
• Consistency — everyone works from the same up-to-date copy of the data.

Remote Access

Networks let authorized users connect to systems from a different physical location, rather than requiring them to be physically present at the machine they need.

Example: An employee working from home can connect to their company's internal servers or cloud applications over the internet, often through a VPN (covered later in this lesson).

Controlled Access and Security

Networks include mechanisms that decide who can see or use which resources, which is essential once many people share the same infrastructure.

Example: A bank's online banking system lets a customer view and manage their own account, but the network's access controls prevent that customer from reaching the bank's internal systems or other customers' data.

Distributed Processing

Modern networks rarely rely on a single computer to do all the work. Instead, they use distributed processing, where a task is split across multiple computers that each handle part of the job and combine their results.

This matters because a single machine has limited processing power and represents a single point of failure — if it goes down, everything relying on it goes down too. Splitting the work across many machines improves both performance and reliability: if one server fails, others can continue serving requests.

Example: A large website rarely runs on one server. Incoming traffic is typically spread across many machines — one group might handle user logins, another might process payments, and another might store and retrieve data — all coordinated over the network so the system behaves as a single service to the end user.

Network Models: Why Layering Matters

Real networks involve many moving parts at once: physical cabling, electrical signaling, addressing schemes, routing decisions, and the applications that actually use the data. Trying to describe all of that as one single process would be difficult to design, troubleshoot, or teach.

To manage this complexity, networking is broken into layers, where each layer is responsible for one specific job and communicates only with the layers directly above and below it. This has two major practical benefits:

• Modularity — a change to one layer (for example, replacing Wi-Fi with a wired connection) doesn't require changing how applications are written, because the layers above don't need to know how the layer below does its job.
• Easier troubleshooting — when something goes wrong, you can isolate the problem to a specific layer (for example, "the cable is unplugged" is a Physical Layer problem, while "the website won't load" is more likely an Application Layer or DNS problem).

The two major layered models used to describe networking are the OSI Model and the TCP/IP Model.

The OSI Model (Open Systems Interconnection)

The OSI model is a conceptual reference model published by the International Organization for Standardization. It describes networking as seven distinct layers, each handling a specific part of getting data from one device to another. It's important to understand that OSI is a teaching and design reference — real-world protocols (especially TCP/IP) don't always map cleanly onto exactly seven layers, but the model remains the standard way to reason about and discuss network behavior.

Layer	Name	Main Responsibility	Example Protocol / Technology
7	Application	Provides network services directly to end-user applications	HTTP, FTP, SMTP, DNS
6	Presentation	Formats, encrypts, and compresses data for the application	TLS/SSL, data encoding
5	Session	Establishes, manages, and ends communication sessions	Session tokens, API sessions
4	Transport	Provides end-to-end delivery between applications	TCP, UDP
3	Network	Handles logical addressing and routing between networks	IP, routers
2	Data Link	Handles addressing and reliable delivery between directly connected devices	Ethernet, MAC addresses, switches
1	Physical	Transmits raw bits as electrical, optical, or radio signals	Cables, connectors, Wi-Fi radios

Data flows down through these layers on the sending device (each layer adds its own header of control information — a process called encapsulation), travels across the physical medium, and flows back up through the same layers in reverse on the receiving device, where each layer's header is read and stripped off (decapsulation).

Layer 1: Physical Layer

The Physical Layer is responsible for transmitting raw bits — 0s and 1s — across a physical medium. It doesn't understand addresses, errors, or meaning; it only cares about turning a bit into a signal (an electrical voltage, a pulse of light, or a radio wave) and sending it.

This layer defines:

• The type of cables and connectors used
• How electrical or optical signals represent 0s and 1s
• The data rate (how many bits per second can be transmitted)
• The physical layout of connections between devices

Example: An Ethernet cable connecting a laptop to a network switch, and the electrical signaling used on that cable, both belong to the Physical Layer.

Network Topology

Topology describes how devices are physically or logically arranged and connected within a network. The choice of topology affects cost, reliability, and how easy the network is to expand or troubleshoot.

Topology	Layout	Key Trade-off
Bus	All devices share a single cable	Cheap to install, but one cable break can disable the whole network
Star	All devices connect to a central switch or hub	If the central device fails, the whole network goes down, but a single faulty cable only affects one device
Ring	Each device connects to exactly two neighbors, forming a loop	Data travels predictably, but one broken link can disrupt the ring unless it supports dual paths
Mesh	Devices connect directly to many (or all) other devices	Highly fault-tolerant, but expensive and complex to wire
Tree	Star networks connected together in a hierarchy	Scales well for larger organizations, but depends heavily on the top-level connections

Example: Most modern office and home LANs use a star topology, where every device connects to a central switch or Wi-Fi router. This makes it easy to add or remove a device without disrupting the rest of the network.

Transmission Media

Transmission media are the physical or wireless paths that actually carry signals between devices.

Guided media — signals travel through a physical, enclosed path:

• Twisted pair cable — the most common cable in office and home LANs (used in standard Ethernet cabling)
• Coaxial cable — used historically for cable television and early Ethernet, still used for some cable internet connections
• Fiber-optic cable — carries data as pulses of light, offering much higher speed and distance than copper cables, at higher cost

Unguided media — signals travel through open space, without a physical enclosure:

• Wi-Fi — wireless LAN connectivity based on radio waves
• Bluetooth — short-range wireless connections between nearby devices
• Satellite communication — long-distance wireless links, often used where cabling is impractical

Multiplexing

Multiplexing allows multiple independent signals to share a single physical communication channel, instead of requiring a separate physical link for every conversation. Common techniques include splitting the channel by frequency (Frequency Division Multiplexing), by time slot (Time Division Multiplexing), or, for fiber, by light wavelength (Wavelength Division Multiplexing).

Example: A single fiber-optic cable between two cities can carry thousands of simultaneous phone calls and data streams at once because multiplexing divides the available capacity among them.

Switching

Switching determines how data actually moves from a source to a destination across a network made up of many interconnected links. The main approaches are:

• Circuit switching — a dedicated physical path is reserved for the entire duration of a communication session (used in traditional telephone networks). This guarantees consistent performance but wastes capacity when the link is idle.
• Packet switching — data is broken into small units called packets, and each packet is routed independently, potentially through different paths, then reassembled at the destination. This uses available bandwidth far more efficiently.
• Message switching — an entire message is sent as a single unit and stored temporarily at each intermediate device before being forwarded. This approach is largely historical and is rarely used in modern networks.

The internet is built on packet switching, which is why it can efficiently share the same physical links among millions of unrelated conversations at once.

Layer 2: Data Link Layer

The Data Link Layer is responsible for reliable communication between two devices that are directly connected on the same physical link (for example, a computer and the switch it plugs into). It organizes raw bits from the Physical Layer into structured units called frames, and it makes sure those frames reach the correct device on the local network.

Main functions:

• Framing — grouping bits into frames with a defined start and end
• MAC addressing — identifying devices on the local network using a hardware address burned into the network interface
• Error detection and correction — catching (and sometimes fixing) corrupted data
• Flow control — preventing a fast sender from overwhelming a slow receiver

Example: Ethernet is the most common Data Link Layer technology in wired LANs. Every network interface card has a unique MAC (Media Access Control) address, a 48-bit hardware identifier (commonly written as six pairs of hex digits, e.g., 3C:5A:B4:12:9F:01) used to deliver frames to the correct device on the local network segment.

Error Detection

Because electrical and radio signals can be distorted by interference, the Data Link Layer includes techniques to detect whether a frame arrived correctly:

• Parity check — adds a single bit that indicates whether the number of 1-bits in the data is even or odd, catching simple single-bit errors
• Checksum — a calculated value sent along with the data, recalculated by the receiver to verify the data wasn't altered
• Cyclic Redundancy Check (CRC) — a more robust mathematical check widely used in Ethernet and other protocols, able to catch most common transmission errors

Error Correction

Detecting an error is not the same as fixing it. Some systems use error correction so the receiver can recover the original data without needing a retransmission:

• Forward Error Correction (FEC) — extra redundant data is sent alongside the original data, allowing the receiver to reconstruct small errors on its own
• Hamming code — a classic error-correcting code that can detect and correct single-bit errors, still used as a teaching example and in some low-level hardware protocols

Flow Control

Flow control prevents a fast sender from transmitting data faster than a slow receiver can process it, which would otherwise cause the receiver to drop data.

Example: When a smartphone downloads a large file from a high-speed server, flow control mechanisms let the phone signal how much data it can currently accept, so the server doesn't flood it faster than it can keep up.

Sliding Window Protocol: Rather than sending one frame and waiting for an acknowledgment before sending the next (which wastes time), the sliding window protocol allows a sender to transmit several frames before it needs to receive an acknowledgment. This improves throughput, reduces idle waiting time, and makes better use of available bandwidth — while still allowing the receiver to signal when it needs the sender to slow down.

Layer 3: Network Layer

The Network Layer is responsible for logical addressing and routing — that is, giving every device a network-wide address and figuring out the best path to move data from one network to another. While the Data Link Layer only handles delivery within a single local network, the Network Layer handles delivery across multiple interconnected networks, which is exactly what happens every time data travels across the internet.

Main functions:

• Logical (IP) addressing
• Routing — selecting a path between source and destination
• Packet forwarding — moving each packet toward its destination, one router at a time
• Congestion management — reducing packet loss and delay when a link becomes overloaded

IP Addressing

Every device on a network needs a unique logical address, called an IP address, so that other devices know where to send data. Two versions of the Internet Protocol are in use today:

IPv4

• Uses 32-bit addresses, written as four decimal numbers separated by dots (e.g., 192.168.1.10)
• Provides roughly 4.3 billion possible addresses, which is no longer enough for the number of internet-connected devices in use today

IPv6

• Uses 128-bit addresses, written as eight groups of hexadecimal digits (e.g., 2001:0db8:85a3::8a2e:0370:7334)
• Provides an effectively unlimited address space, which removes much of the practical need for the address-sharing workarounds IPv4 requires
• Was designed with native support for IPsec (network-layer encryption and authentication), though this support is no longer a strict requirement of the specification — IPv6 does not automatically make a network "more secure" by itself

Public vs. Private IP Addresses

Public IP addresses are globally unique and directly reachable over the internet. Example: 203.0.113.10.

Private IP addresses are reserved for use inside local networks (homes, offices, data centers) and are not routable on the public internet. The reserved private ranges are:

• 10.0.0.0 – 10.255.255.255
• 172.16.0.0 – 172.31.255.255
• 192.168.0.0 – 192.168.255.255

Because private addresses aren't unique across the whole internet, a device using one can't be reached directly from outside. To let devices on a private network access the internet, a router performs Network Address Translation (NAT), rewriting each outgoing packet's private source address to the router's own public address (and reversing the translation for replies). This is why a home network can have dozens of devices, each with a private address like 192.168.1.x, while sharing a single public IP address issued by an internet service provider.

Routing Algorithms

Routers use routing algorithms to decide the best path for forwarding packets toward their destination. "Best" usually means fewest hops, lowest delay, or highest available bandwidth, depending on how the algorithm is configured. Efficient routing directly affects a network's speed, reliability, and overall performance.

Distance Vector Routing: Each router shares its own routing table with its directly connected neighbors, and routers gradually build up knowledge of the wider network by exchanging this information periodically.

• Example protocol: RIP (Routing Information Protocol)
• Advantage: Simple to configure and understand
• Disadvantage: Slow to converge on a stable set of routes after a network change, and prone to routing loops in larger networks

Link State Routing: Each router builds a complete map of the network's topology by exchanging link-state information with all other routers, then independently calculates the best path using that full picture.

• Example protocol: OSPF (Open Shortest Path First)
• Advantages: Converges faster after a network change and scales better to larger, more complex networks than distance vector routing

Layer 4: Transport Layer

The Transport Layer provides end-to-end communication between applications running on different devices, regardless of how many networks the data has to cross to get there. While the Network Layer only gets packets from one device to another, the Transport Layer makes sure the right application on the receiving device gets the right data, in the right order (if required).

Main functions:

• Segmentation — breaking data from an application into smaller segments for transmission, and reassembling them on arrival
• Error recovery — detecting lost or corrupted segments and requesting retransmission when needed
• Flow control — matching the sending rate to what the receiver can handle
• Congestion control — reducing the sending rate when the network itself is overloaded, to avoid making congestion worse

TCP (Transmission Control Protocol)

TCP provides reliable, ordered delivery between two applications.

Key features:

• Connection-oriented — a connection is explicitly established before data is sent
• Error checking — corrupted segments are detected and retransmitted
• Ordered delivery — segments are reassembled in the correct order even if they arrive out of sequence

Connection establishment (three-way handshake): Before exchanging data, TCP establishes a connection using three messages: the client sends a SYN (synchronize) to request a connection, the server responds with SYN-ACK to accept it, and the client replies with ACK to confirm. Only after this handshake completes do the two sides start exchanging actual data.

Connection termination: TCP closes a connection using a four-step exchange, since either side can close its half of the connection independently:

1. The side finishing first sends FIN.
2. The other side responds with ACK.
3. That side then sends its own FIN when it is also done.
4. The first side responds with a final ACK.

This two-way closing process ensures that both sides have finished sending data before the connection is fully torn down.

Example: Web browsing, file downloads, and online banking all rely on TCP, because losing or misordering even a small part of the data would break the page, file, or transaction.

UDP (User Datagram Protocol)

UDP is a simpler, faster alternative to TCP. It sends data without establishing a connection first and without guaranteeing delivery or order — if a packet is lost, UDP does not automatically resend it.

Key features:

• Connectionless — no handshake before sending data
• Low latency — less overhead means data can be sent with less delay
• No retransmission — lost packets are simply not resent by UDP itself

This isn't a flaw — it's a deliberate trade-off. For applications like live video or voice calls, an old, retransmitted packet arriving late is more disruptive than simply dropping it and moving on, so UDP's speed is more valuable than TCP's reliability.

Example uses: online gaming, live video streaming, and voice calls (VoIP), where speed and low delay matter more than guaranteeing every single packet arrives.

TCP vs. UDP at a Glance

	TCP	UDP
Connection	Connection-oriented (handshake required)	Connectionless
Reliability	Guarantees delivery and order	No delivery or order guarantee
Speed	Slower, due to error checking and acknowledgments	Faster, minimal overhead
Typical use	Web browsing, email, file transfer	Streaming, gaming, VoIP

Layer 5: Session Layer

The Session Layer manages the lifecycle of a communication session between two applications — establishing it, keeping it active while data is exchanged, and closing it cleanly when finished. In practice, this responsibility is often absorbed into the application itself rather than implemented as a separate protocol layer, which is one reason the TCP/IP model (discussed below) doesn't include a distinct session layer.

Main functions:

• Session establishment
• Session maintenance (keeping the conversation active and synchronized)
• Session termination

Example: A video conference call needs to stay logically connected for its full duration even though the underlying data is constantly being broken into packets and reassembled — that continuity is the kind of job the Session Layer describes.

Layer 6: Presentation Layer

The Presentation Layer is responsible for making sure data is represented in a format both the sender and receiver can understand, regardless of differences in how each device internally stores data. This includes translating between data formats, compressing data to reduce its size, and encrypting it for security.

Main functions:

• Data translation (e.g., between character encodings)
• Compression
• Encryption

Example: Converting text into a standard character encoding before transmission so the receiving device interprets it correctly, regardless of what operating system or application generated it.

SSL/TLS: Transport Layer Security (TLS), and its older, now-deprecated predecessor SSL, provide encryption, authentication, and data integrity for data traveling across a network. In the strict OSI model, encryption is often associated with this layer, though in real implementations TLS actually operates on top of TCP (that is, functionally between the Transport and Application layers) — a good example of why OSI is best treated as a conceptual reference rather than a literal description of how modern protocols are built.

Example: HTTPS websites use TLS to encrypt the data exchanged between a browser and a web server, which is why sensitive information like passwords or payment details shouldn't be sent over plain HTTP.

Layer 7: Application Layer

The Application Layer is the layer closest to the end user. It doesn't refer to a specific "application" like a web browser, but to the network protocols that applications use to communicate — things like requesting a web page, sending an email, or resolving a domain name.

Main functions:

• Providing network services that applications rely on directly, such as web access, email, file transfer, and name resolution

Several protocols operate at this layer, each solving a specific communication need:

Protocol	Purpose	Common Port
HTTP	Transfers web pages between browsers and web servers	80
HTTPS	Encrypted version of HTTP, using TLS	443
FTP	Transfers files between computers over a network	20 (data), 21 (control)
SSH	Encrypted remote command-line access	22
Telnet	Unencrypted remote command-line access (largely obsolete)	23
SMTP	Sends email between mail servers	25
DNS	Resolves domain names to IP addresses	53
SNMP	Monitors and manages network devices	161/162

A few of these are worth calling out individually:

• DNS (Domain Name System) works like the internet's phone book: it's much easier for a person to remember www.google.com than an IP address like 142.250.190.14, so DNS translates the domain name into the correct IP address before the browser connects to the server.
• Telnet provides remote command-line access but sends everything — including passwords — as plain, unencrypted text. It has largely been replaced by SSH, which encrypts the entire session.
• SNMP lets administrators monitor and manage network hardware such as routers, switches, and servers from a central console, often used for detecting failures and tracking performance.

The TCP/IP Model

While OSI is a conceptual reference model, the TCP/IP model describes how the internet actually works in practice. It groups networking into four layers instead of seven:

1. Application Layer — combines the roles of OSI's Application, Presentation, and Session layers
2. Transport Layer — same role as OSI's Transport Layer (TCP, UDP)
3. Internet Layer — corresponds to OSI's Network Layer (IP addressing and routing)
4. Network Access Layer — combines OSI's Data Link and Physical Layers

OSI Layer	TCP/IP Layer
Application, Presentation, Session	Application
Transport	Transport
Network	Internet
Data Link, Physical	Network Access

The TCP/IP model is the practical foundation of the modern internet — it's what protocol suites are actually built and documented against — while OSI remains the more detailed model used for teaching and troubleshooting.

The Client-Server Model

Most network applications are built around the client-server model, where one program requests a service and another program provides it.

• Client — requests a service. Examples: a web browser, a mobile app.
• Server — fulfills the request. Examples: a web server, a database server.

Example: When you open a website, your browser (the client) sends a request to a web server, which responds with the page content.

An alternative architecture is peer-to-peer (P2P), where devices communicate directly with each other as equals instead of relying on a central server — used by technologies like some file-sharing protocols. Client-server remains far more common because centralizing services makes them easier to secure, update, and manage.

Cloud Networking and Virtualization

Cloud networking applies the same networking principles covered above — addressing, routing, protocols — to infrastructure that runs in a cloud provider's data center rather than on hardware an organization owns and maintains directly. This lets organizations deploy applications and scale capacity up or down without buying and managing physical servers themselves.

Types of Cloud Services

• Infrastructure as a Service (IaaS) — provides virtualized computing infrastructure, such as virtual machines, storage, and networking, that customers configure themselves.
• Platform as a Service (PaaS) — provides a ready-to-use platform for building and hosting applications, without managing the underlying servers.
• Software as a Service (SaaS) — provides complete, ready-to-use software over the network. Example: web-based email and collaboration tools.

Virtual Private Network (VPN)

A VPN creates an encrypted tunnel between a user's device and a remote network, so traffic between them can't be read or tampered with by anyone else on the networks it passes through.

Example: An employee working from home can use a VPN to securely reach internal company resources as if their laptop were physically connected to the office network — the connection is encrypted, but it's worth noting that a VPN protects data in transit; it doesn't make a user anonymous or automatically secure the endpoint device itself.

Network Function Virtualization (NFV)

NFV replaces dedicated physical networking hardware — routers, firewalls, load balancers — with equivalent functions running as software, typically on standard servers. This reduces hardware costs, allows new network functions to be deployed in software rather than by installing new equipment, and makes networks more flexible to reconfigure.

Emerging Networking Trends

5G Networks

5G is the current generation of mobile network technology, offering substantially higher speeds, lower latency, and the ability to support far more connected devices per cell than earlier generations. This capacity is what makes it practical for applications like smart city infrastructure, autonomous vehicles, and industrial automation, which depend on large numbers of devices communicating with minimal delay.

Network Slicing

Network slicing allows a single physical network infrastructure to be partitioned into multiple isolated virtual networks, each tuned for a different purpose — for example, one slice optimized for low-latency traffic and another for high-throughput traffic — without needing separate physical hardware for each. This improves resource utilization and lets a provider offer customized service levels to different customers on the same infrastructure.

Computer Network Security

As networks connect more devices and carry more sensitive data, protecting that data becomes as important as delivering it. Network security covers the tools and practices that protect systems and communications from unauthorized access or interference.

Common security mechanisms include:

• Firewalls — filter incoming and outgoing traffic based on defined rules
• Antivirus software — detects and removes malicious software
• Intrusion Detection Systems (IDS) — monitor network traffic for suspicious activity
• Access control policies — restrict which users or devices can reach specific resources

Privacy, closely related to security, ensures that personal and organizational data stays confidential even when it's stored or transmitted across shared infrastructure. This is achieved through encryption, secure authentication, and access control.

Digital signatures verify two things about a piece of data: that it genuinely came from the claimed sender (authenticity), and that it wasn't altered after being signed (integrity). They're commonly used in online transactions, secure email, and digital certificates.

Pretty Good Privacy (PGP) is an encryption system used to secure the contents of emails and files, combining encryption, authentication, and digital signatures so that only the intended recipient can read a message and verify who sent it.

Common Mistakes to Avoid

• Confusing MAC addresses with IP addresses. A MAC address identifies a device's network interface on the local network (Data Link Layer) and generally doesn't change; an IP address identifies a device's location on a larger network (Network Layer) and can change depending on which network it connects to.
• Assuming a private IP address is reachable from the internet. Private addresses (like 192.168.1.5) only work within their local network — reaching them from outside requires NAT, port forwarding, or a VPN.
• Treating HTTP as safe for sensitive data. Plain HTTP is not encrypted; only HTTPS (HTTP over TLS) protects data such as passwords or payment details in transit.
• Assuming UDP is simply "worse" than TCP. UDP's lack of reliability guarantees is an intentional trade-off for speed, not a defect — the right choice depends on whether the application needs guaranteed delivery (TCP) or minimal delay (UDP).
• Believing a VPN provides full anonymity. A VPN encrypts traffic between your device and the VPN server, but it doesn't hide your activity from the VPN provider itself or protect a compromised device.
• Memorizing the OSI layers without understanding what each one does. The value of the model is in understanding why the layers are separated — it makes troubleshooting and protocol design far easier than treating it as a list to recite.