Internet and Networking

A computer network is a set of interconnected devices that exchange data and share resources through agreed communication protocols. The Internet is the global packet-switched network of interconnected networks that uses the TCP/IP protocol suite. The Internet is infrastructure; the World Wide Web, email, file transfer and DNS are services that operate over it.

Network communication requires a source and destination, a transmission medium, addressing, rules for formatting and exchanging data, and intermediate devices that forward traffic. Scope, topology, protocol layer and service role answer different questions and must not be used interchangeably.

Network Scope: PAN, LAN, MAN and WAN

Networks are classified partly by geographical reach and partly by administrative ownership. The boundaries are descriptive rather than exact distance limits.

Type	Typical coverage	Ownership and technology	Example and operating implication
PAN	A person’s immediate area	One user; Bluetooth, USB or short-range wireless	Phone linked to a headset; very short reach and low infrastructure overhead
LAN	Room, building or campus	Usually one home or organization; Ethernet and Wi-Fi	PCs and printers in an office; high local capacity and direct local administration
MAN	Town or metropolitan region	Municipality, operator or consortium; Metro Ethernet and fiber rings	Fiber linking exchanges across Kathmandu; routing and resilient metropolitan paths are needed
WAN	Regional, national or global	Multiple organizations, telecom operators and ISPs; MPLS/IP, leased lines, microwave, submarine fiber and satellite	Operator backbone connecting provinces or the Internet; more hops, carrier agreements and greater delay/error exposure

PAN, LAN, MAN and WAN compared.

In the usual comparison, LAN delay and error exposure are lowest, MAN is intermediate, and WAN is highest because distance and the number of intermediate systems increase. This is a tendency, not a guarantee: a congested wireless LAN can perform worse than a well-engineered WAN.

Note

Example — Classifying an operator network

PCs and printers inside one Nepal Telecom office form a LAN. Linking several exchanges across one city forms a MAN. Carrying that traffic between provinces and onward to international networks uses a WAN. A Bluetooth link from an engineer’s phone to a headset is a PAN.

LANs commonly permit direct local administration and high capacity. MAN and WAN designs additionally need routing, service-level agreements, path redundancy and strong protection when traffic crosses shared or untrusted facilities.

Physical and Logical Topologies

A network topology is either the physical arrangement of nodes, ports and links or the logical path followed by frames or signals. These views can differ: switched Ethernet is normally a physical star whose links are logically independent point-to-point connections.

Physical link arrangements and traffic direction for terminated bus, switched star, directional ring, partial mesh, full mesh and hierarchical tree topologies.

Topology	Arrangement	Advantages	Failure domain and limitations
Bus	Every station taps one backbone terminated at both ends	Low cable cost and simple small installation	A backbone break or bad terminator can stop all nodes; the shared medium has collisions and poor fault isolation
Star	Every station has a separate link to a central hub or switch	Easy addition, management and spoke-fault isolation	One spoke failure affects one node, but central-device or power failure affects the whole star
Ring	Each node has two neighbors in a closed loop; traffic follows the ring protocol	Orderly, predictable access without a central hub	One link or node can break a single ring; dual counter-rotating rings improve recovery at extra cost
Mesh	Nodes have several links; a full mesh joins every pair	Multiple paths, high availability and load sharing	Port/cable cost and routing complexity; full-mesh growth is quadratic
Tree	Access nodes feed distribution nodes under a core/root	Scalable hierarchy and policy aggregation	An upper-level failure can disconnect an entire branch unless the hierarchy is redundantly connected

Characteristics of common network topologies.

For $n$ nodes in a full mesh, each node could connect to $n-1$ others, but counting both ends would count every link twice. Therefore

Caution

Key Formula — Number of full-mesh links

$$\boxed{L=\frac{n(n-1)}{2}}.$$

[!NOTE]

Example — Choosing a topology

A modern office normally uses a switched star because one failed access cable isolates only one host. A WAN backbone commonly uses a partial mesh to obtain alternate routes without full-mesh cost. Five fully meshed routers require

$$L=\frac{5(5-1)}{2}=\boxed{10\text{ links}},$$

whereas ten routers would already require $45$ links.

IP Addressing and Internet Protocols

An IP address is a logical network-layer identifier assigned to an interface. Routers compare its network prefix with routing-table entries to choose a next hop. It is not the same as a data-link MAC address, a DNS name, or the physical location of a device.

An IPv4 address contains 32 bits and is normally written as four decimal octets, for example 192.168.10.25. Early IPv4 allocation used the following classful interpretation.

Class	First octet	Default prefix	Historical purpose
A	–126	/8	Very large unicast networks
B	–191	/16	Medium unicast networks
C	–223	/24	Small unicast networks
D	–239	Not host addressing	Multicast
E	–255	Not ordinary host addressing	Reserved or experimental

Historical classful IPv4 ranges.

Note

Key Point — CIDR replaced classful routing

Classes A, B and C explain historical defaults but do not control modern allocation or route selection. Classless Inter-Domain Routing (CIDR) uses address/prefix-length; a /n prefix fixes the first $n$ network bits, permitting aggregation and variable-size subnets.

The entire 127.0.0.0/8 block is reserved for loopback: traffic sent there remains within the host. Zero-valued addresses also have special meanings; for example, 0.0.0.0 can denote an unspecified local address and 0.0.0.0/0 is the default route, so neither should be treated as an ordinary assigned host address without context.

Prefix	Meaning
10.0.0.0/8	Large private range
172.16.0.0/12	Private range from 172.16.0.0 through 172.31.255.255
192.168.0.0/16	Common home and organizational private range

Private IPv4 address space.

Private addresses can be routed internally but are not globally routed on the public Internet. An edge router often translates them with NAT, although NAT is not encryption and does not replace a firewall.

Subnet Mask and Prefix Length

A subnet mask separates the network and host portions of an IPv4 address. It has $n$ leading one bits followed by $32-n$ zero bits for prefix /n; bitwise AND of the address and mask gives the subnet address.

Caution

Key Formula — IPv4 subnet size

$$N_{\text{addresses}}=2^{32-n}, \qquad N_{\text{traditional usable hosts}}=2^{32-n}-2.$$

The subtraction reserves the all-zero host part as the network address and the all-one host part as the directed broadcast address. It is the traditional exam rule, not a universal law: /31 point-to-point links can use both addresses and /32 identifies one host route.

Note

Example — Finding an IPv4 subnet

For 192.168.10.25/24, the mask is 255.255.255.0. The subnet address is 192.168.10.0, the traditional directed broadcast is 192.168.10.255, and the usual host range is 192.168.10.1--192.168.10.254; hence $2^8-2=254$ usable host addresses.

IPv6 and Address Comparison

IPv6 expands the address to 128 bits and writes eight hexadecimal groups separated by colons, for example 2001:0db8:0000:0000:0000:ff00:0042:8329. Leading zeros in a group may be omitted and one consecutive run of zero groups may be compressed as ::.

Feature	IPv4	IPv6
Address length	bits	bits
Notation	Dotted decimal	Hexadecimal colon notation
Address space	$2^{32}\approx4.3\times10^9$	$2^{128}\approx3.4\times10^{38}$
Base header	Variable length, normally 20–60 bytes	Simplified fixed 40-byte base header; optional information uses extension headers
Broadcast	Supported	No broadcast; multicast and anycast serve the relevant delivery cases
NAT	Common because public addresses are scarce	Usually less necessary for address conservation, though translation may still be deployed

IPv4 and IPv6 compared.

Layered Protocol Architecture

Layering divides communication into manageable services. At the sender, application data is carried by TCP, UDP or a transport built over UDP such as QUIC; IP supplies logical addressing and routing; Ethernet, Wi-Fi or another link technology creates a local frame; and the physical layer sends bits as electrical, optical or radio signals.

OSI and TCP/IP layer mapping with functions, protocol examples, PDUs and the encapsulation sequence from application data to transmitted bits.

The seven-layer OSI model is a conceptual ISO reference framework, whereas the four-layer TCP/IP model describes the practical Internet architecture. Their functional mapping is:

• OSI Application $+$ Presentation $+$ Session $\rightarrow$ TCP/IP Application;

• OSI Transport $\rightarrow$ TCP/IP Transport;

• OSI Network $\rightarrow$ TCP/IP Internet; and

• OSI Data Link $+$ Physical $\rightarrow$ TCP/IP Network Access.

The protocol data units (PDUs) are data at the upper layers, a TCP segment or UDP datagram at transport, an IP packet at network, a frame at data link and bits at physical. During encapsulation, each lower layer adds control information around the complete higher-layer PDU; the data-link layer commonly adds both a header and an error-detecting trailer/FCS. The receiver decapsulates in reverse and delivers the payload upward.

Note

Key Point — Layering qualification

Layering gives standard interfaces, interoperability, modular replacement and systematic fault isolation. The mapping is functional rather than perfectly one-to-one: OSI is a reference model, TCP/IP merges layers, and mechanisms such as TLS, MPLS and tunnels can appear to span conventional boundaries.

Domain Name System

The Domain Name System (DNS) is a distributed hierarchical database that maps domain names to typed resource records, including IPv4/IPv6 addresses, mail exchangers and authoritative name servers. DNS resolution does not itself carry the later web or email content.

The host’s stub resolver first checks an unexpired browser or OS cache and then sends one recursive query to its configured recursive resolver. If that resolver has no valid cached answer, it performs iterative queries: a root server refers it to the relevant top-level domain (TLD) server, the TLD server refers it to the domain’s authoritative server, and the authoritative server returns the requested record.

DNS resolution as one recursive client query followed, on a cache miss, by iterative root, TLD and authoritative queries and TTL-controlled caching.

The resolver returns the answer to the stub and caches the result for its time-to-live (TTL). Referrals, address records and negative answers may all be cached. A recursive resolver performs work on the client’s behalf; an iterative server normally returns the best answer or referral it has. The client stub does not normally walk the root/TLD hierarchy itself.

Common DNS Records

Record	Purpose
A	Maps a name to an IPv4 address
AAAA	Maps a name to an IPv6 address
CNAME	Makes one domain name an alias of another canonical name
MX	Identifies the mail exchanger for a domain
NS	Identifies an authoritative name server for a DNS zone
TXT	Carries text information, including data used for SPF and DKIM verification

Common DNS resource records.

Note

Example — Resolving a web name

For www.example.com A, a cache miss leads to root referral $\rightarrow$ .com referral $\rightarrow$ authoritative A answer. The recursive resolver returns the IPv4 address and reuses it only while its TTL remains valid. DNS supplies an address; the browser must still establish transport and, for HTTPS, validate the TLS peer before requesting the resource.

URL Structure

A Uniform Resource Locator (URL) identifies how and where a client can access a resource. Its complete general anatomy is

scheme://host:port/path?query#fragment

https://www.example.com:443/docs/page?unit=ict#dns

Component	Example	Function
Scheme	https	Selects the access protocol and its default port
Host	www.example.com	Names the destination host; DNS normally resolves it to one or more IP addresses
Port	443	Optionally selects the destination server process; the scheme supplies a default when omitted
Path	/docs/page	Identifies a resource within the server’s namespace
Query	unit=ict	Supplies parameters to the server-side resource
Fragment	dns	Identifies a client-side section of the returned representation and is normally not sent in the HTTP request

Meaning of each URL component.

HTTP and HTTPS

HTTP is a stateless application-layer request/response protocol. A client sends a method, target, headers and possibly a body; a server returns a status code, headers and possibly a representation. Stateless means each HTTP request is independently interpretable, not that a web application cannot maintain state through cookies, tokens or server-side sessions.

HTTP request and response between a browser and a web server, with the transport protection added by HTTPS/TLS.

HTTPS carries HTTP through TLS. Certificate-based server authentication helps the client verify the named peer, encryption provides confidentiality in transit, and integrity protection detects modification. HTTPS does not prove that the site’s content is honest, remove endpoint malware or replace application authorization.

HTTP Methods and Status Codes

Method	Intended use
GET	Retrieve a resource representation
POST	Submit data for processing or create a subordinate resource
PUT	Replace, or create at, the addressed resource
PATCH	Partially update a resource
DELETE	Request deletion of the addressed resource

Common HTTP methods.

Code	Meaning	Interpretation
200	OK	Request succeeded
301 / 302	Redirect	Resource is permanently/temporarily available at another location
400	Bad Request	Server cannot process malformed or invalid request syntax
401	Unauthorized	Authentication is required or failed; despite the name, this is primarily an authentication response
403	Forbidden	Server understood the request but refuses authorization
404	Not Found	Requested resource was not found
500	Internal Server Error	Server encountered an unexpected failure

Frequently tested HTTP status codes.

FTP, FTPS and SFTP

Protocol	Foundation	Connections and common port	Security meaning
FTP	File Transfer Protocol	TCP control connection on port 21 plus a separate active or passive data connection	Basic FTP does not encrypt credentials or payload
FTPS	FTP with TLS	FTP control/data model with explicit or implicit TLS	Adds TLS protection while remaining FTP
SFTP	SSH File Transfer Protocol	One SSH connection, commonly TCP port 22	A different protocol carried by SSH; it is not an FTP “secure mode”

File-transfer protocol distinction.

Hub, Switch and Router

A hub is a physical-layer multiport repeater. A switch is a data-link bridge that learns source MAC addresses and forwards frames toward a learned destination port. A router is a network-layer device that matches destination IP addresses against a routing table and forwards packets between IP networks.

Hub, one-VLAN switch and router behavior, showing collision domains and the Layer-2 broadcast boundaries separated by routing.

Feature	Hub	Layer-2 switch	Router
OSI layer / unit	Layer 1 / bits	Layer 2 / frames	Layer 3 / packets
Forwarding basis	None; repeats incoming bits	Learned MAC address table	Longest matching IP prefix and next hop
Collision domains	One shared domain	One per switched port	One per interface or point-to-point link
Broadcast domains	One	One per VLAN	Each routed interface is a separate IP broadcast domain
Typical use	Obsolete small or shared test segment	Connect hosts inside a LAN/VLAN	Connect LANs/VLANs to other networks or a WAN

Hub, Layer-2 switch and router compared.

A hub repeats unicast and broadcast traffic to every port and permits collisions on a shared half-duplex segment. A switch gives each port a separate full-duplex collision domain, but floods broadcasts and unknown unicasts within the VLAN. A router does not normally forward Layer-2 broadcasts, so its interfaces bound separate IP broadcast domains.

Client-Server Model and Server Roles

A server is a software process that listens for requests and provides controlled resources to client programs. A client initiates a request. One physical or virtual computer may run several server processes, distinguished by transport addresses and ports.

Client-server operation: the client initiates a request, the listening server processes it and returns a controlled response.

Web, Email and Print Servers

The destination IP address selects the server host and the TCP or UDP port selects the listening process. The server authenticates or validates the request as required, accesses a resource, returns a response and should log security-relevant activity.

Server	Main function	Protocols and common ports	Example
Web	Stores or generates pages, files and API responses	HTTP/TCP 80; HTTPS with TLS commonly TCP 443	Browser requests a self-care page from Nginx, Apache or an application server
Email	Accepts, relays, filters, stores and retrieves messages	SMTP relay TCP 25; submission 587; IMAP 143/993; POP3 110/995	Sender server looks up the recipient MX record and transfers mail by SMTP
Print	Shares printers, accepts jobs, queues/spools them and reports status	IPP commonly TCP 631, plus vendor and legacy print protocols	Several clients submit jobs that are serialized for one office printer

Web, email and print server roles and common ports.

Port numbers identify conventional service endpoints, but configuration can change them. A listed port therefore aids connection and filtering; it does not by itself authenticate the service.

Transmission Media

Transmission media are guided when a physical conductor or waveguide confines the signal and unguided when electromagnetic waves propagate through free space.

Guided Media

Medium	Main characteristics	Typical use
Twisted pair	Cheap and easy to install; copper attenuation and interference limit reach and rate	Ethernet LAN and telephone line
Coaxial cable	Better shielding than twisted pair; robust shared radio-frequency path	Cable television and older Ethernet
Optical fiber	Very high bandwidth, low loss, electrical isolation and immunity to EMI	Backbone, FTTH and long-distance links

Guided transmission media.

Unguided Media

Medium	Main characteristics	Typical use
Radio wave	Often omnidirectional and able to penetrate buildings; shared spectrum suffers interference	Wi-Fi and mobile communication
Microwave	Directional and normally line-of-sight; high antenna gain and frequency reuse	Terrestrial point-to-point link and satellite uplink
Infrared	Short range and usually line-of-sight; does not readily penetrate walls	Remote controls and short-range device links

Unguided transmission media.

Media selection balances required rate and reach against attenuation, noise, EMI, terrain, spectrum/licensing, security exposure, installation effort, availability and total cost.

Transport Protocols: TCP and UDP

TCP and UDP both use port numbers for process-to-process delivery, but provide different service contracts over IP.

Feature	TCP	UDP
Connection	Connection-oriented byte stream	Connectionless datagrams
Reliability	Sequence numbers, acknowledgements and retransmission	Best effort; an application must add recovery if needed
Ordering	Delivers bytes in order and suppresses duplicates	No delivery, ordering or duplicate-suppression guarantee
Flow/congestion control	Built in	Not supplied by UDP itself
Overhead and latency	More setup/state/header overhead	Small header and no connection setup; low protocol overhead
Typical use	Web over HTTP/1.1 or HTTP/2, email, FTP and SSH	Many DNS queries, streaming, VoIP, gaming and QUIC/HTTP/3

TCP and UDP compared.

“UDP is faster” is only a shorthand for lower built-in overhead. Application design, loss recovery, congestion and network conditions determine observed performance. DNS normally uses UDP for compact queries but can use TCP for large responses, retries and zone transfer; modern HTTP/3 uses QUIC over UDP while implementing reliability and security above UDP.

Electronic Mail Protocols

Email separates message transfer from mailbox access. SMTP pushes a message from a client to a submission server and between mail servers. The recipient then uses IMAP to synchronize a server-resident mailbox or POP3 primarily to download messages.

Protocol	Direction and function	Important distinction
SMTP	Sends/submits and relays mail toward the recipient server	Transfer protocol; MX DNS records identify destination mail exchangers
POP3	Downloads mail from server to client	Simple retrieval, often oriented toward local storage
IMAP	Synchronizes mailbox state between server and one or more clients	Folders, flags and messages normally remain coordinated on the server

Core email protocols.

Email delivery from the sender client through SMTP submission and server-to-server relay to recipient mailbox access by IMAP or POP3.

TLS-protected variants or STARTTLS protect links in transit, but mail can cross several administrative systems; endpoint access control, anti-spam checks and appropriate end-to-end content protection remain separate concerns.

Gateway, Intranet and Extranet

A gateway connects dissimilar systems or performs protocol/format translation. In ordinary IP host configuration, the default gateway is more specifically the router to which a host sends packets for destinations outside its local subnet.

Term	Meaning
Internet	Public global interconnection of IP networks
Intranet	Private organizational network that uses Internet technologies such as IP, DNS, web applications and browsers
Extranet	Controlled extension of private-network services to selected external partners, suppliers or customers

Internet, intranet and extranet.

An extranet is not simply a public website: identities, authorization and network/application controls restrict which external parties reach which private resources.

Basic Network Security Controls

Security is layered because no single control supplies identity, confidentiality, authorization, availability and detection simultaneously.

Control	Purpose and qualification
Firewall	Permits or denies traffic according to address, port, protocol, state or application policy; rules must be least-privilege and reviewed
Authentication	Verifies a user, service or device identity; strong methods include MFA and certificate/key-based credentials
Encryption	Protects confidentiality and usually integrity in transit; it does not decide whether an authenticated party is authorized
VPN	Creates an authenticated encrypted tunnel across an untrusted network; the protected endpoints and access policy still require security
Access control	Limits resources and operations to authorized identities, roles, devices or network segments
Segmentation	Uses VLANs, subnets, routing/firewall policy or separate zones to limit reachability and fault/attack propagation
Logging and monitoring	Records activity and detects suspicious behavior; logs need time synchronization, protection, review and response procedures
Patching and secure configuration	Removes known weaknesses, disables unused services and changes insecure defaults
Backup and recovery	Restores essential configurations and data after failure or attack; backups must be protected and restoration tested

Basic network controls and their purposes.

Tip

Exam Focus — Answering a network-security question

Name the protected asset and traffic path, then place controls at the endpoint, access link, segment boundary and application. State what each control does and one limitation; for example, a VPN encrypts transit but does not make an infected remote endpoint trustworthy.

Quick Review

• Scope: PAN is personal, LAN covers a site, MAN spans a city and WAN spans regions to the globe.

• Topology: modern LANs commonly use switched star; WANs often use partial mesh, while full mesh needs $n(n-1)/2$ links.

• Addressing: IPv4 is 32-bit and IPv6 is 128-bit; CIDR prefixes, not obsolete address classes, define modern networks.

• Layering: OSI has seven layers and TCP/IP is commonly shown with four; each lower layer encapsulates the higher-layer PDU with its own control information.

• Transport: TCP is reliable, ordered and connection-oriented; UDP is connectionless with low built-in overhead.

• DNS: the stub makes a recursive request to a resolver; on a cache miss the resolver makes iterative root, TLD and authoritative queries.

• URL: remember the exact order scheme://host:port/path?query#fragment.

• Web: HTTP is stateless request/response; HTTPS adds TLS peer authentication, confidentiality and integrity in transit.

• File transfer: FTP uses separate control and data connections, FTPS adds TLS to FTP, and SFTP is a distinct SSH protocol.

• Devices: a hub repeats bits, a switch forwards frames by MAC address and a router forwards packets by IP prefix.

• Domains: a hub shares one collision and broadcast domain; a switch separates collision domains per port and VLANs separate broadcast domains; router interfaces bound IP broadcast domains.

• Media and services: guided media use copper/fiber, unguided media use free-space radio; server ports identify web, mail and print processes.

• Email: SMTP transfers mail, POP3 downloads it and IMAP synchronizes a server-resident mailbox.

• Private access: an intranet is internal; an extranet grants controlled access to selected external parties; the default gateway routes off-subnet traffic.

Tip

Exam Focus — Ten-mark Internet/networking answer

Begin with a precise definition, draw the relevant layered or topology figure, compare alternatives in a table, and finish with one operational caveat. Keep identifiers distinct: DNS name $\rightarrow$ IP address $\rightarrow$ route, then transport port $\rightarrow$ server process.