Computer Network and Communication

When you send a video to a friend in Dharan, pay a bill with eSewa or join an online class, your data crosses several networks in less than a second. This unit explains how that happens: how signals travel through cables and through the air, which devices pass them along, how computers are arranged and given addresses, and how the Internet differs from the private networks inside schools and companies. Almost everything here is asked as a definition, a difference or a diagram, so learn the tables as carefully as the text.

1.1 Concept of telecommunication

Every telecommunication link, from a landline call to a 4G video stream, has the same five parts.

When devices are joined so that they can exchange data in this way, they form a network. The rest of this section defines the words used to describe how fast and how well such a link works.

Bandwidth

Bandwidth is measured in bits per second. In networking the prefixes are decimal, the small letter b means bit, and the capital B means byte (1 byte = 8 bits).

Radio engineers also use "bandwidth" for the width of a frequency band in hertz: a Wi-Fi channel, for example, is 20 MHz wide. A wider band can carry more bits per second, which is why the two meanings share one word.

Throughput

Throughput can never be more than the bandwidth, and it is usually less, because of

• many users sharing the same link at the same time (congestion);
• distance, walls and interference that weaken a wireless signal;
• damaged data that has to be sent again;
• the extra address information that travels with every packet of data.

A third measure, latency (delay), is the time a piece of data takes to reach the other end, in milliseconds (ms). For video calls and online games a low latency matters as much as a high throughput.

Broadband

Before broadband, homes used dial-up: a modem dialled a telephone number before every session, reached at most 56 kbps and kept the telephone line busy. Broadband is hundreds of times faster, stays connected, and leaves the telephone free.

Mobile generations: 3G, 4G and 5G

A generation ("G") is a major new version of mobile-phone technology, arriving roughly once every ten years. 1G (1980s) carried analogue voice and 2G (1990s) added digital voice and SMS. The three that matter now are:

4G gave Nepal fast mobile Internet in every province. 5G is faster still, has almost no delay and can serve a very large number of devices in a small area, which is what smart cities and the Internet of Things (Unit 5) need.

Data packets

Every packet has three parts:

Sending data this way is called packet switching. The packets may take different routes and arrive out of order; the receiver uses the packet numbers to rebuild the message and asks again for any packet that is missing or damaged. Because the link is shared packet by packet, many users can use it at once, a broken route affects only a few packets, and long-distance transfer becomes fast and reliable.

Frequency

Frequency decides how a signal behaves. A higher frequency has a shorter wavelength, can carry more data, but has a shorter range and is stopped more easily by walls and rain. A lower frequency travels farther and passes through walls better but carries less data. This is why 5 GHz Wi-Fi is faster than 2.4 GHz Wi-Fi in the same room but weaker one floor up.

Communication modes

Data communication is the exchange of data between two devices over a medium. According to the direction in which data can flow, there are three modes of communication.

1.2 Communication media: guided and unguided

Guided media

1. Twisted-pair cable. Pairs of insulated copper wires are twisted around each other. The twisting cancels much of the electrical interference that one pair picks up from its neighbours and from outside. UTP (unshielded twisted pair) is used in almost every LAN; STP (shielded twisted pair) adds a foil screen for electrically noisy places such as factories. Telephone lines also use twisted pairs. Twisted-pair cables are sold in categories:

A single run of twisted-pair Ethernet cable may be at most 100 m long.

2. Coaxial cable. A single copper core is surrounded by an insulating layer, a braided copper shield and an outer plastic jacket. Because the shield surrounds the core on every side, coaxial cable resists interference well. It is used for cable TV, for CCTV cameras, to connect dish antennas, and in older Ethernet networks.

3. Optical fibre cable. A hair-thin core of very pure glass (or plastic) is surrounded by glass cladding that bends light less strongly. Light pulses sent into the core strike the boundary at a shallow angle and are reflected back every time (total internal reflection), so the light stays in the core even when the cable curves. A buffer coating, strength fibres and a jacket protect the glass. Light travels in glass at about 200,000 km/s, two-thirds of its speed in a vacuum.

Single-mode fibre has a very thin core (about 9 µm) and a laser source and carries data for tens of kilometres; ISPs use it for their main lines. Multimode fibre has a thicker core (50 or 62.5 µm), uses cheaper light sources and serves shorter links inside buildings and campuses. Most of Nepal's Internet traffic enters the country through fibre links from India, and since 12 January 2018 also from China through the Rasuwagadhi–Kerung border.

Unguided media

Unguided media send electromagnetic waves through air or space, so no cable is needed. They are what make mobile phones, Wi-Fi and satellite TV possible. Because every wireless user shares the same airwaves, the government regulates who may use which frequency; Nepal's mobile operators, for example, use frequencies licensed through the Nepal Telecommunications Authority (NTA).

1. Radio waves (roughly 3 kHz to 1 GHz) spread out in all directions, pass through walls and travel long distances. AM and FM radio, mobile phones and walkie-talkies use them.

2. Microwaves (roughly 1 to 300 GHz) travel in straight lines, so the sending and receiving dish antennas must be in line of sight. Towers on hilltops relay microwave signals over hops of a few tens of kilometres, which made them useful in Nepal's hills where laying cable is hard. Heavy rain can weaken them.

3. Infrared waves carry data only a few metres, need a clear line of sight and cannot pass through walls. A TV remote control is the everyday example; because infrared stays inside one room, it is also hard to eavesdrop on.

4. Satellite communication uses an artificial satellite as a microwave relay station in space. A ground station sends the signal up — the uplink — and the satellite sends it back down to a wide area — the downlink — on a different frequency. A geostationary satellite, 35,786 km above the equator, turns with the Earth and appears fixed in the sky, so dish antennas never need to move. The first commercial communications satellite, Early Bird (Intelsat I), went into service in 1965. Satellites carry direct-to-home (DTH) television, weather images, GPS navigation and Internet to remote areas, but because the signal travels about 72,000 km up and down, each one-way trip takes about a quarter of a second. Nepal's own first satellite, the small research satellite NepaliSat-1, was launched in 2019.

5. Wi-Fi (Wireless Fidelity, the IEEE 802.11 standards) connects devices to a LAN by radio in the 2.4 GHz and 5 GHz bands, usually over 30–50 m indoors. A router or access point sends and receives the signals. Modern Wi-Fi uses channel bonding (joining two neighbouring channels into one wider, faster channel) and beamforming (aiming the signal towards each device instead of spreading it equally everywhere). A WPA2 or WPA3 password encrypts the data in the air.

6. Bluetooth connects devices over about 10 m using very little power, in the 2.4 GHz band. It uses frequency hopping spread spectrum (FHSS): the connection jumps between 79 narrow channels about 1,600 times a second, so interference on one channel spoils only a tiny part of the data. Earbuds, keyboards, smartwatches and car audio systems use it.

7. RFID (Radio Frequency Identification) uses radio waves to read data stored in a small tag fixed to an object. A passive tag has no battery and is powered by the reader's own signal, so it works only close to the reader; an active tag has a battery and can be read from much farther away. Library books, ID cards, contactless payment cards, toll gates and warehouse stock all use RFID.

1.3 Connectors: RJ-45, RJ-11, BNC, fibre connectors and the media converter

RJ-45 (Registered Jack 45). The clear plastic plug on every Ethernet cable. It has eight positions and eight contacts (8P8C), one for each of the eight wires of a twisted-pair cable, and a springy clip that locks it into the port. The order of the coloured wires follows the standard T568A or T568B.

RJ-11. The smaller plug of a telephone cable. It has six positions but usually only two or four contacts. It connects telephones and DSL modems to the phone line. An RJ-11 plug fits loosely into an RJ-45 port, but it is the wrong connector for Ethernet.

BNC (Bayonet Neill–Concelman). A round metal connector for coaxial cable that is pushed on and turned a quarter turn to lock (a bayonet joint). It is used on CCTV cameras, laboratory instruments and old coaxial Ethernet.

Fibre-optic connectors hold the glass fibre in a precise ceramic tube (the ferrule) so that the two fibre ends meet exactly:

• ST (Straight Tip) — round metal body with a bayonet twist-lock;
• SC (Subscriber Connector, also called Square or Standard Connector) — square plastic body that clicks in with a push and comes out with a pull; it is the usual plug on home fibre (FTTH) equipment;
• LC (Lucent Connector) — a small push-pull connector about half the size of SC, used on the fibre modules of switches;
• SMA (SubMiniature version A) — a connector with a threaded nut that is screwed on; mainly older and industrial equipment.

Media converter. A media converter joins two different kinds of cable, usually copper twisted pair and optical fibre. It turns the electrical signals arriving through its RJ-45 port into light pulses for the fibre port, and turns incoming light back into electrical signals. It is needed when

• a link is longer than the 100 m allowed for copper Ethernet, for example between two school buildings;
• a fibre line from an ISP or another building must join an ordinary copper switch;
• a cable must pass through a place full of electrical noise, such as near motors or high-voltage lines.

1.4 Networking devices

A network is built from three kinds of component: hardware (computers, cables and the devices below), software (a network operating system such as Windows Server or Linux, and application software), and the communication medium. The devices decide where each piece of data goes.

NIC (Network Interface Card). The circuit, built into every computer and phone or added as a card, that connects the device to the network medium — an RJ-45 port for a cable or a radio for Wi-Fi. Every NIC carries a permanent 48-bit MAC (Media Access Control) address, written in hexadecimal such as 3C-52-82-9A-7F-10, which identifies that card on the local network.

Repeater. A signal grows weaker the farther it travels. A repeater receives the weak signal, regenerates it to full strength and sends it on, so the network can reach farther. It does not look at addresses.

Hub. A box with many ports into which computers are plugged in a star. A hub is a "multi-port repeater": whatever arrives at one port is copied to every other port. All the computers share one bandwidth and their signals can collide. Hubs are cheap and simple but have been replaced by switches.

Switch. Also a multi-port box at the centre of a star, but intelligent. It reads the MAC address in each frame, learns which computer is on which port (its MAC address table), and sends the frame only to the port of the destination. Each port gets the full speed and can send and receive at the same time, so a switched LAN is faster and more private than one built on a hub.

Bridge. Joins two LAN segments that use the same protocol, inspects each frame and forwards it to the other segment only if the destination is there. In effect it is a two-port switch.

Router. Joins different networks — for example a school LAN and the ISP's network — and forwards packets between them using IP addresses. It keeps a routing table and chooses the best path for each packet. The "Wi-Fi router" in a home is really several devices in one box: a router, a small switch, a wireless access point, a DHCP server that hands out IP addresses, and a firewall.

Modem (modulator–demodulator). Converts the computer's digital signals into the form the line carries (modulation) and back again (demodulation). DSL modems work on telephone lines, cable modems on coaxial cable and 4G dongles or MiFi devices on the mobile network. On a fibre line the same job is done by an ONT (Optical Network Terminal), which converts light to electrical signals.

Wireless access point (WAP). Lets Wi-Fi devices join a wired LAN. It is connected to a switch by cable and broadcasts a network name (SSID); schools, hotels and offices mount several on ceilings so that the signal reaches every room.

Gateway. A device or program that connects two networks that use different protocols, translating between them — for example a VoIP gateway between telephone lines and the Internet, or an IoT gateway between sensors and a cloud server. In everyday use, the default gateway of a computer is the router address to which it sends every packet meant for another network.

1.5 Network topology

Bus topology. All nodes are attached to one main cable, the bus or backbone. A terminator at each end absorbs the signal so that it does not bounce back. A signal sent by one node travels along the whole bus and every node checks whether it is the addressee. A bus needs the least cable and is cheap for a few computers, but a single break in the backbone stops the whole network, faults are hard to find, and more nodes mean more collisions.

Ring topology. Each node is connected to the next and the last to the first, forming a closed loop. Data travels round the ring in one direction, from node to node; each node checks the address, regenerates the signal like a repeater and passes it on. In a token ring only the node holding a special message, the token, may send, so there are no collisions. But one broken cable or switched-off node can stop the whole ring (unless a second ring is laid in the opposite direction), and adding a node means breaking the ring.

Star topology. Every node has its own cable to a central switch (or, in old networks, a hub). It is the topology of almost every LAN today. A faulty cable affects only its own computer, nodes can be added or removed without disturbing the others, and faults are easy to trace. It needs more cable than a bus, and if the central switch fails, the whole star stops.

Mesh topology. Nodes are joined by many links, so there are several paths between any two of them. In a full mesh every node is linked to every other: n nodes need \frac{n(n-1)}{2} links and each node needs n-1 ports. A mesh keeps working when a link fails and carries heavy traffic, but it is the most expensive. The routers of the Internet's backbone form a (partial) mesh, and "mesh Wi-Fi" systems use the same idea at home.

Tree topology. Star networks arranged in levels under a root: a main switch feeds the switches of each floor or building, which feed the computers. A school with several buildings is usually wired this way. It is easy to extend branch by branch, but if the root fails, the branches below it are cut off.

Hybrid topology. A combination of two or more topologies — stars joined by a bus or ring backbone, or a tree with a mesh core. Most real networks are hybrid: each department keeps its own star while the backbone uses whatever suits the distance and the budget. A hybrid network grows easily and one failure rarely stops everything, but it is more complex to design and manage.

1.6 Networks by coverage: PAN, LAN, MAN and WAN

A network lets users share hardware (one printer or scanner for a whole office), share data and software, communicate (email, chat, video calls, the Internet), keep centralised control of users and security, and make central backups. It offers services such as file, print, message, application and database services. Networks are classified by the area they cover.

PAN (Personal Area Network). Connects the devices of one person within about 10 m: a phone with earbuds and a smartwatch over Bluetooth, or a laptop using a phone's hotspot. It is easy to set up and needs little configuration.

LAN (Local Area Network). Connects computers in one room, building or campus, up to about a kilometre across. It is privately owned (by the school or office), fast (100 Mbps to 10 Gbps), has very few errors, and today is almost always a star of switches. A LAN whose devices join by Wi-Fi is a WLAN (wireless LAN). A school's computer lab is a LAN; a LAN spread over a university's buildings is sometimes called a CAN (campus area network).

MAN (Metropolitan Area Network). Covers a city or a valley, from a few to about 50 km. It usually joins many LANs, is owned by one large organisation or a service provider, and links the LANs to the wider Internet. A cable-TV network across Kathmandu Valley, an ISP's fibre ring around a city, or a bank joining its branches within one city are MANs.

WAN (Wide Area Network). Covers a country, a continent or the whole world. It joins LANs and MANs through links owned by many organisations — telephone companies, ISPs and satellite operators — mainly optical fibre on land and under the sea, with satellite and microwave links where cables cannot go. Compared with a LAN its data rate per user is lower and its delay higher. The Internet is the largest WAN; a bank linking branches in every province, or Nepal Telecom's national network, are others.

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

Network architecture describes how the computers of a network share the work: whether some of them provide services for the rest, or all of them are equal.

Network protocols

The first host-to-host protocol of ARPANET, the network that grew into the Internet, was NCP (Network Control Program). ARPANET carried its first message on 29 October 1969 — "LO", because the system crashed before "LOGIN" was complete — and replaced NCP with TCP/IP on 1 January 1983.

1.8 IP addressing

IPv4

An IPv4 address is 32 bits long, written as four decimal numbers called octets, each from 0 to 255, separated by dots — dotted-decimal notation, for example 192.168.1.10. Each octet is 8 bits, so 2^{32} = 4,294,967,296 (about 4.3 billion) addresses are possible. That is fewer than the number of devices now online; the central pool of free IPv4 addresses ran out in 2011, and networks survive by giving private addresses inside and sharing one public address through NAT (Network Address Translation) on the router.

IPv6

IPv6 was designed to replace IPv4. Its addresses are 128 bits long, written as eight groups of four hexadecimal digits separated by colons, for example 2001:0db8:85a3:0000:0000:8a2e:0370:7334. That gives 2^{128} \approx 3.4\times10^{38} addresses — enough for every device that will ever be built. Two rules shorten an address: leading zeros in a group may be dropped, and one run of consecutive all-zero groups may be replaced by ::. So the address above is written 2001:db8:85a3::8a2e:370:7334.

Private and public, static and dynamic addresses

Private addresses are reserved for use inside LANs. They can be repeated in every home and school, and routers on the Internet do not forward them. Public addresses are globally unique and are assigned by ISPs (in our region, from the pool managed by APNIC); every device visible on the Internet needs one.

Two other special addresses: 127.0.0.1 (loopback) always means "this computer", and an address beginning 169.254 means a computer asked for an address by DHCP and got no reply.

Checking an IP address and the default gateway

A device's subnet mask (usually 255.255.255.0 in a small LAN) tells it which addresses are on its own network. For any other address it sends the packet to the default gateway, the address of the router — the door out of the LAN.

Wireless LAN adapter Wi-Fi:

   Connection-specific DNS Suffix  . :
   Link-local IPv6 Address . . . . . : fe80::1c2b:7d4e:9a10:3f21%12
   IPv4 Address. . . . . . . . . . . : 192.168.1.105
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Default Gateway . . . . . . . . . : 192.168.1.1

Pinging google.com [142.250.193.78] with 32 bytes of data:
Reply from 142.250.193.78: bytes=32 time=38ms TTL=117
Reply from 142.250.193.78: bytes=32 time=37ms TTL=117
Reply from 142.250.193.78: bytes=32 time=41ms TTL=117
Reply from 142.250.193.78: bytes=32 time=38ms TTL=117

Ping statistics for 142.250.193.78:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),

1.9 Internet, intranet and extranet

The Internet offers the World Wide Web, email, file transfer (FTP), chat (IRC and messaging apps), remote login (Telnet, SSH), video calls, online shopping, banking and cloud storage. A bank's internal portal where staff check circulars and apply for leave is an intranet; the same bank's portal where its merchants check their QR-payment settlements is an extranet. A school whose results and attendance pages can be seen only from inside the school is running an intranet; if parents may log in from home to see their own child's records, that part becomes an extranet.

Textbook exercises: model answers

1. Full forms. (a) DSL — Digital Subscriber Line. (b) bps — bits per second. (c) LAN — Local Area Network. (d) TCP/IP — Transmission Control Protocol / Internet Protocol. (e) IPv6 — Internet Protocol version 6. (f) ISP — Internet Service Provider. (g) RFID — Radio Frequency Identification. (h) CAT6 — Category 6 (twisted-pair cable). (i) NCP — Network Control Program (also called Network Control Protocol). (j) DNS — Domain Name System.

Practice questions

Group A — Multiple choice (1 mark each)

1. A student's plan is 50 Mbps but a speed test reads 38 Mbps. The 38 Mbps is the [1]abandwidthbthroughputcfrequencydlatency
2. A walkie-talkie works in [1]asimplex modebhalf-duplex modecfull-duplex modedbroadcast mode
3. Which medium carries data as pulses of light? [1]aCAT6bcoaxial cablecoptical fibredSTP cable
4. The connector on a telephone-line cable is [1]aRJ-45bRJ-11cBNCdSC
5. Which device sends a frame only to the port of the destination computer? [1]ahubbrepeatercswitchdmodem
6. In which topology does the failure of the central device stop the whole network? [1]abusbringcstardmesh
7. How many cables are needed to join 8 computers in a full mesh? [1]a8b16c28d56
8. Which is a private IPv4 address? [1]a8.8.8.8b172.32.0.5c192.168.10.25d1.1.1.1
9. Which command shows a Windows computer's IP address, subnet mask and default gateway? [1]apingbipconfigctracertdnslookup
10. An IPv6 address is written as [1]afour decimal numbers with dotsbeight hexadecimal groups with colonsc32 binary digitsdsix hexadecimal pairs with hyphens
11. A company lets its suppliers log in to check orders and stock. This network is an [1]aInternetbintranetcextranetdPAN
12. Which protocol gives devices their IP addresses automatically? [1]aDNSbDHCPcFTPdSMTP

Group B — Short answer (2 marks each)

1. Differentiate between bandwidth and throughput with two points. [2]
2. A 600 MB file is downloaded on a 20 Mbps connection. Find the shortest possible download time. [2]
3. Give two reasons why optical fibre, not CAT6, should join two school buildings 400 m apart. [2]
4. What is a gateway? How is it different from a router? [2]
5. Write two differences between a hub and a switch. [2]
6. Write the IPv6 address 2001:0db8:0000:0000:0000:0000:0000:0001 in its shortest form and state the length of an IPv6 address in bits. [2]
7. Differentiate between a static and a dynamic IP address, with one use of each. [2]
8. What is a data packet? Name two items carried in its header. [2]

Group C — Long answer (4 marks each)

1. What is network topology? Explain star and mesh topologies with diagrams, giving one advantage and one disadvantage of each. [1+3]
2. A school has two buildings 300 m apart. Building A has a lab of 30 computers; Building B has an office with 5 computers and a printer; teachers want Wi-Fi in both. Suggest the media, devices and topology, giving a reason for each choice. [4]
3. Define IP address. Differentiate between IPv4 and IPv6 in any three points. [1+3]
4. Explain PAN, LAN, MAN and WAN with one example of each. [4]