Data Communication and Networking

A computer on its own is a calculator; connected to others it becomes a communication system. This unit explains how a message gets from one machine to another — what the pieces of a communication system are, what carries the signal, what damages it on the way, how machines are addressed, and how the whole job is divided into layers. At 15 teaching hours this is the largest unit in the course and it supplies the largest share of the written marks.

2.1 Basic elements of a communication system

Data communication is the exchange of data between two devices through a transmission medium. For it to happen at all, five elements must be present.

Four characteristics decide whether a communication system is any good: delivery (to the correct destination), accuracy (no altered bits), timeliness (in time to be useful — critical for voice and video), and jitter (even spacing between arriving packets).

2.2 Concept of a communication system and its block diagram

A general communication system is described by five functional blocks. This is the diagram examiners ask for.

• The source produces the data.
• The transmitter converts it into a signal the medium can carry — a modem modulating digital bits onto an analogue carrier, for example.
• The transmission system carries the signal, and unavoidably adds noise.
• The receiver converts the received signal back into data — the modem at the far end demodulating.
• The destination consumes the data.

2.3 Elements of data transmission

A signal is the electrical, optical or electromagnetic representation of data. Signals are of two kinds:

	Analogue signal	Digital signal
Shape	Continuous wave	Discrete levels (0 and 1)
Described by	Amplitude, frequency, phase	Bit rate, bit interval
Example	Human voice, radio broadcast	Data inside a computer
Noise	Accumulates and cannot be removed	Can be regenerated exactly

Three more terms are examined constantly:

• Bandwidth — for an analogue channel, the range of frequencies it passes, in hertz; for a digital channel, the maximum bit rate, in bits per second. Larger bandwidth means more data per second.
• Bit rate — bits transmitted per second (bps, kbps, Mbps, Gbps).
• Baud rate — signal changes per second. If each signal change carries n bits then \text{bit rate} = \text{baud rate}\times n; they are equal only when one change carries one bit.
• Transmission mode — data may be sent serially (one bit after another on one wire; cheap, used over any distance) or in parallel (n bits at once on n wires; fast but only over short distances). Serial transmission is further divided into synchronous (a continuous stream timed by a shared clock) and asynchronous (one character at a time framed by start and stop bits).

2.4 Simplex, half duplex and full duplex

Mode	Direction	Channel capacity used	Example
Simplex	One way only	Full capacity one way	Keyboard to CPU, radio and TV broadcast
Half duplex	Both ways, alternately	Full capacity, one direction at a time	Walkie-talkie, old coaxial Ethernet with a hub
Full duplex	Both ways simultaneously	Capacity shared, or two separate channels	Telephone, switched Ethernet, mobile call

2.5 Concept of LAN and WAN

A computer network is two or more devices connected so that they can share data and resources. Networks are classified by the area they cover.

Point	LAN	MAN	WAN
Area	A room, building or campus (up to ~2 km)	A city (5–50 km)	Country to worldwide
Ownership	Private — one organisation	One organisation or an ISP	Usually leased from telecom operators
Speed	Highest: 100 Mbps to 10 Gbps	High	Lowest per user; long-haul links are shared
Error rate	Very low	Low	Higher
Media	Twisted pair, fibre, Wi-Fi	Fibre, microwave	Fibre, satellite, leased lines
Cost of setup	Low	Medium	High
Example	A school computer lab	Cable TV or ISP network inside Kathmandu	The Internet, a bank's branch network

A PAN (personal area network, a few metres — Bluetooth earphones) and a CAN (campus area network) are sometimes added at the small end.

2.6 Transmission medium: guided and unguided

Medium	Structure	Typical speed / bandwidth	Range before repeater	Notes
Twisted pair (UTP)	4 pairs of insulated copper twisted together; RJ-45 plug	Cat 5e 1 Gbps (100 MHz); Cat 6 up to 10 Gbps over 55 m (250 MHz); Cat 6a 10 Gbps over 100 m (500 MHz)	100 m	Cheapest, easiest to install; twisting cancels crosstalk; most LAN cabling
STP	UTP plus a metal shield	Same as UTP	100 m	Better noise immunity, costlier, must be earthed
Coaxial	Copper core, insulator, braided shield, jacket	Up to ~1 Gbps; hundreds of MHz	~500 m	Used by cable TV and cable internet; obsolete for LANs
Optical fibre	Glass core + cladding; light is carried by total internal reflection	Tbps possible; 10–100 Gbps routine	Many km (single-mode > 40 km)	Immune to EMI, very secure, very high bandwidth; costly, needs skilled splicing
Radio wave	Omnidirectional, 3 kHz–1 GHz	Up to a few hundred Mbps	Tens of km	Penetrates walls; FM radio, older wireless
Microwave	Line-of-sight dish, 1–300 GHz	Hundreds of Mbps to Gbps	~50 km per hop	Needs clear line of sight; Wi-Fi uses 2.4, 5 and 6 GHz
Infrared	Line-of-sight light, < 1 mm wavelength	Low to moderate	A few metres	Cannot pass through walls, so it is secure inside a room; TV remotes
Satellite	Microwave relayed by a satellite	High	Global	Large propagation delay for geostationary satellites (~250 ms one way)

Point	Guided	Unguided
Path	Along a physical conductor	Through free space
Direction	Point-to-point	Usually broadcast
Interference	Low (shielded)	High (weather, obstacles, other transmitters)
Security	Higher — tapping needs physical access	Lower — anyone in range can receive
Installation	Cabling work, hard in difficult terrain	Quick, reaches remote hills and islands
Mobility	None	Full

Current practice (2026). Wi-Fi 6/6E (IEEE 802.11ax) is the mainstream wireless LAN standard; Wi-Fi 7 (802.11be) was published in July 2025 and adds 320 MHz channels and 4096-QAM, and Wi-Fi 8 (802.11bn) is still under development with an expected release around 2028. Nepal's fixed broadband is now mostly FTTH (fibre to the home); mobile data runs on 4G LTE, which Nepal Telecom launched on 1 January 2017 — 5G in Nepal has so far only reached the trial stage.

2.7 Transmission impairments

A signal that arrives is never exactly the signal that was sent. The differences are impairments.

Impairment	What it is	Cause	Cure
Attenuation	Loss of signal strength with distance	Resistance of the medium	Amplifier (analogue), repeater (digital)
Distortion	Change in the shape of the signal	Different frequency components travel at different speeds	Equalisers; limit the bandwidth used
Noise	Unwanted energy added to the signal	See the four types below	Shielding, twisting, error detection
Crosstalk	Signal from one wire leaks into a neighbouring wire — you faintly hear another conversation	Electromagnetic coupling between adjacent conductors	Twisting the pairs, shielding (STP), fibre
Echo	The signal is reflected back to the sender, who hears a delayed copy of themselves	Impedance mismatch at a junction or hybrid	Echo suppressors and echo cancellers
Singing	A continuous howling or whistling tone — an echo that has become self-sustaining feedback	Echo plus excessive amplifier gain going round a loop	Reduce gain, use echo cancellers
Jitter	Variation in the arrival time (delay) of successive packets	Varying queueing delay in routers	Jitter buffer at the receiver, QoS
Bandwidth	A limit, not a fault: the channel simply cannot carry frequencies (or bit rates) above its bandwidth, so fast signals are rounded off	Physical properties of the medium	Use a wider medium — fibre instead of copper
Number of receivers	Each extra receiver tapped onto a line draws power and adds a reflection, so the signal weakens and distorts	Loading of the line	Use point-to-point links and switches instead of a shared bus

The four standard kinds of noise: thermal (random electron motion, present in every conductor, also called white noise), intermodulation (two frequencies mixing to produce a third), crosstalk, and impulse (sudden spikes from lightning or switching — the worst for digital data).

2.8 Network architecture: client–server and peer-to-peer

Point	Client–server	Peer-to-peer
Roles	Dedicated server provides, clients request	Every node is both client and server
Best size	Medium to very large networks	Up to about 10 computers
Cost	High — server hardware, server OS, admin	Low — ordinary PCs, no extra software
Administration	Centralised: one place for accounts, backup, security	Each user administers their own machine
Security	Strong, centrally enforced	Weak, depends on every user
Performance	Scales well; server may become a bottleneck	Degrades as the number of peers grows
Failure	Server down = network unusable	One peer down affects only its own shares
Example	A school's ERP, any website	File and printer sharing in a small office; BitTorrent

2.9 Basic terms

Term	Meaning
IP address	A logical address identifying a host on a network. IPv4 is 32 bits written as four decimal octets (202.79.32.33); IPv6 is 128 bits written in hexadecimal. Assigned by software; changes when the host moves network.
Subnet mask	A 32-bit number of consecutive 1s then 0s (255.255.255.0) that tells a host which part of an IP address is the network and which is the host.
Default gateway	The IP address of the router interface on your own network. Any packet whose destination is outside the local network is handed to the gateway.
MAC address	The 48-bit physical address burned into a NIC by its manufacturer, written as 12 hexadecimal digits (00:1A:2B:3C:4D:5E). The first 24 bits are the manufacturer's OUI. Permanent and globally unique; works only within one LAN.
Internet	The global public network of networks, using TCP/IP; open to everyone.
Intranet	A private network inside one organisation using the same Internet technologies (web, email), accessible only to its employees.
Extranet	A controlled extension of an intranet to selected outsiders — suppliers, dealers, partners — usually over a VPN with a login.

2.10 Network tools

• Cisco Packet Tracer — a free network simulator. You drag routers, switches, PCs and cables onto a canvas, configure them, and watch animated packets travel hop by hop. It lets a student build and debug a whole network without owning any hardware, so it is the standard lab tool for this course.
• Remote login — working on a distant computer as though you were sitting at it. Telnet (port 23) was the original tool but sends everything, including the password, as plain text; it has been replaced by SSH (Secure Shell, port 22), which encrypts the whole session. For a graphical desktop, Windows uses RDP (port 3389), and VNC and AnyDesk are common alternatives.
• Small diagnostic commands worth knowing: ping (is the host reachable, and how long does a round trip take), tracert/traceroute (list the routers along the path), ipconfig/ifconfig (show my IP address, mask and gateway), and nslookup (find the IP address behind a domain name).

2.11 Network connecting devices

Device	OSI layer	Function
NIC (network interface card)	1 and 2	Connects a computer to the medium; holds the MAC address; converts parallel data to serial signals
Modem (modulator–demodulator)	1	Modulates digital data onto an analogue carrier to send, demodulates on receipt — needed on telephone, cable and DSL lines
Repeater	1	Regenerates a weakened signal to extend distance
Hub	1	Multiport repeater: copies an incoming frame to every other port. Obsolete — wastes bandwidth and causes collisions
Bridge	2	Joins two LAN segments and filters traffic by MAC address
Switch	2	Multiport bridge: learns which MAC address is on which port and forwards each frame only to that port. Full duplex, no collisions
Router	3	Joins different networks, reads the destination IP address and chooses the best path from its routing table
Gateway	All 7	Joins networks that use different protocols, translating between them

2.12 Network topologies

Topology	Advantages	Disadvantages
Bus	Least cable, cheapest, easy to extend for a few nodes	One break in the backbone kills the whole network; collisions rise with traffic; hard to fault-find; needs terminators
Star	One cable fault affects one node only; easy to add or remove nodes; easy to manage; no collisions with a switch	Needs the most cable; the central switch is a single point of failure
Ring	Equal access for every node; no collisions (token passing); signal is regenerated at each node so it travels far	A single node or link failure breaks the ring (unless it is a dual ring); adding a node interrupts service; slow with many nodes
Mesh	Most reliable — many alternative paths; no traffic sharing; privacy	Very expensive: n(n-1)/2 links and n-1 ports per node; hard to install
Tree	Scales well; hierarchical, so easy to manage and to isolate groups	Failure of the root or a branch hub disconnects everything below it; much cabling
Hybrid	Combines the strengths of its parts; very flexible; scales to any size	Complex to design and manage; costly

2.13 Basic concept of the OSI reference model

The OSI (Open Systems Interconnection) model was published by ISO in 1984. It is a reference model: it divides the work of communication into seven independent layers, so that any layer may be changed without disturbing the rest. Data travels down the seven layers at the sender, across the medium, and up the seven layers at the receiver. Each layer adds its own header to the data it receives from above — this is called encapsulation.

#	Layer	Main function	PDU	Protocol / device
7	Application	Provides network services to the user's program	Data	HTTP, FTP, SMTP, DNS, Telnet
6	Presentation	Translation of formats, encryption/decryption, compression	Data	TLS/SSL, JPEG, MPEG, ASCII
5	Session	Establishes, manages, synchronises and terminates sessions	Data	NetBIOS, RPC, PPTP
4	Transport	End-to-end delivery, segmentation and reassembly, flow control, error control; port numbers	Segment	TCP (reliable), UDP (fast)
3	Network	Logical (IP) addressing and routing between networks	Packet	IP, ICMP — router
2	Data link	Framing, physical (MAC) addressing, error detection by CRC, access control	Frame	Ethernet, PPP — switch, bridge
1	Physical	Transmits raw bits: voltages, connectors, cables, data rate	Bit	RJ-45, fibre — hub, repeater

Point	OSI	TCP/IP
Layers	7	4 (Application, Transport, Internet, Network access)
Developed by	ISO, 1984	US DoD/ARPANET, 1970s
Status	A theoretical reference model	The model actually used on the Internet
Layer independence	Strictly separated	Application layer merges OSI 5, 6 and 7
Approach	Protocol-independent standard	Built around existing protocols (TCP, IP)

2.14 Internet protocol addressing

An IPv4 address is 32 bits, written as four decimal octets separated by dots — dotted decimal notation. Each octet is 0–255, so the whole space holds 2^{32} \approx 4.3 billion addresses. Every address has a network part (which network) and a host part (which machine on it); the subnet mask marks the boundary.

Class	First bits	First octet	Default mask	Networks	Hosts per network	Use
A	0	1 – 126	255.0.0.0 (/8)	126	16,777,214	Very large organisations
B	10	128 – 191	255.255.0.0 (/16)	16,384	65,534	Medium organisations
C	110	192 – 223	255.255.255.0 (/24)	2,097,152	254	Small networks
D	1110	224 – 239	—	—	—	Multicast
E	1111	240 – 255	—	—	—	Reserved / experimental

Two subtractions to remember: the host part of all 0s is the network address and all 1s is the broadcast address, so a network with n host bits has 2^{n}-2 usable hosts. 127.0.0.0/8 is reserved for loopback (127.0.0.1 is always "this computer"), which is why class A stops at 126.

Private addresses are never routed on the Internet and may be reused inside any organisation: 10.0.0.0/8, 172.16.0.0/12 and 192.168.0.0/16. A router performing NAT (Network Address Translation) lets a whole school share one public address.

IPv6. Because IPv4's 4.3 billion addresses ran out — IANA's free pool was exhausted in February 2011 — the Internet is migrating to IPv6, which uses 128 bits written as eight groups of four hexadecimal digits, for example 2001:0db8:85a3:0000:0000:8a2e:0370:7334. It gives about 3.4\times10^{38} addresses, has no broadcast (it uses multicast instead), and needs no NAT. Adoption has been slow but steady: traffic to Google from IPv6 users passed 50% for the first time in March 2026, up from roughly 43% at the start of 2025.

Chapter summary

• The five elements of data communication are message, sender, receiver, transmission medium and protocol; the communication model is source → transmitter → transmission system → receiver → destination, with noise added in the middle.
• Simplex = one direction; half duplex = both directions alternately; full duplex = both directions at once.
• LAN (a building, private, fastest) ⊂ MAN (a city) ⊂ WAN (country or world; the Internet is the largest WAN).
• Guided media (twisted pair, coaxial, fibre) confine the signal in a conductor; unguided media (radio, microwave, infrared, satellite) radiate it into space. Fibre has the highest bandwidth and immunity to interference.
• Impairments: attenuation, distortion, noise (thermal, intermodulation, crosstalk, impulse), echo, singing, jitter; bandwidth and the number of receivers limit what a channel can do.
• Client–server centralises resources and security; peer-to-peer is cheap and suits fewer than ten machines.
• MAC address = 48-bit, physical, permanent, layer 2. IP address = 32-bit (IPv4), logical, changeable, layer 3. Subnet mask splits network from host; the gateway is the router's address.
• Topologies: bus (one backbone, terminators), star (central switch), ring (one-way loop), mesh (n(n-1)/2 links), tree (hierarchy of stars), hybrid.
• OSI has 7 layers — Physical, Data link, Network, Transport, Session, Presentation, Application — against TCP/IP's 4. Routers work at layer 3, switches at layer 2, hubs and repeaters at layer 1.
• IPv4 classes: A (1–126, /8), B (128–191, /16), C (192–223, /24), D multicast, E reserved. A network with n host bits has 2^{n}-2 usable hosts. IPv6 is 128 bits.

Practice questions

Group A — Multiple choice (1 mark each)

1. Which is not one of the five elements of data communication? [1]amessagebprotocolcbandwidthdtransmission medium
2. A walkie-talkie is an example of which communication mode? [1]asimplexbhalf duplexcfull duplexdbroadcast
3. Which device works at the network layer of the OSI model? [1]ahubbswitchcrouterdrepeater
4. The default subnet mask of a class B address is [1]a255.0.0.0b255.255.0.0c255.255.255.0d255.255.255.255
5. In which topology does the failure of the central device stop the whole network? [1]abusbringcstardmesh
6. A MAC address is how many bits long? [1]a32b48c64d128
7. Unwanted coupling of a signal from one wire into a neighbouring wire is called [1]aechobjitterccrosstalkdattenuation

Answers to Group A

1. (c) — bandwidth is a property of the channel, not one of the five elements.

2. (b) — both can talk, but only one at a time, over one shared channel.

3. (c) — routers read IP addresses, which live at layer 3.

4. (b) — class B uses the first two octets for the network.

5. (c) — every node in a star hangs off the central switch.

6. (b) — 48 bits, written as 12 hexadecimal digits.

7. (c) — that is the definition of crosstalk.

Group B — Short answer (5 marks each)

1. Draw the block diagram of a communication system and explain the function of each block. [5]
2. Differentiate between guided and unguided transmission media on any five points, giving two examples of each. [5]
3. Define topology. Draw bus, star and ring topologies and state one advantage and one disadvantage of each. [5]
4. Explain any five transmission impairments and state how each can be reduced. [5]
5. A host has IP address 172.16.45.9 with mask 255.255.0.0. State its class, network address, broadcast address and the number of usable hosts on that network. Also distinguish an IP address from a MAC address. [5]
6. What is meant by client–server and peer-to-peer architecture? Compare them on any four points. [5]

Answers to Group B

5. The first octet 172 lies in 128–191, so it is class B (and a private address, since it is inside 172.16.0.0/12). With mask 255.255.0.0 the first two octets are the network, so the network address is 172.16.0.0 and the broadcast address is 172.16.255.255. There are 16 host bits, so 2^{16}-2 = 65{,}534 usable hosts. An IP address is a 32-bit logical address assigned by software, used by routers between networks, and it changes when the host joins another network; a MAC address is a 48-bit physical address burned into the NIC, used by switches inside one LAN, and it never changes.

Group C — Long answer (8 marks each)

1. (a) Explain the OSI reference model with a neat diagram, naming each layer and stating its main function and one protocol. [5](b) Compare the OSI model with the TCP/IP model on any three points. [3]

2. (a) What is a network topology? Explain bus, star, ring and mesh topologies with diagrams, listing the advantages and disadvantages of each. [6](b) A full mesh network is to be built for 10 computers. Calculate the number of cables and the number of ports required per computer, and say why a star is preferred in a school laboratory. [2]

Answer outline to Group C question 2(b)

Number of cables = n(n-1)/2 = (10\times 9)/2 = \textbf{45}, and each computer needs n-1 = \textbf{9} ports. A star needs only 10 cables and one switch, is far cheaper, lets a faulty cable be replaced without touching other machines, and can be extended simply by adding a port — so it is the practical choice for a lab, while mesh is reserved for backbone links where redundancy justifies the cost.