Week 1 · Lesson 8

Main Components of Network — Part 1

A network is not just boxes and cables. It is hosts that need addressing, media that carry the signals, and services that make the whole thing useful. This lesson covers the pieces you can see and touch — and the numbers used to describe how fast they move data.

CCNA 200-301 Domain 1.0 Network Fundamentals Verified 2026-10-11
Part 1

End devices and their roles

Every message on a network has a beginning and an end. The device at either end is called an end device or host. Hosts form the interface between the person or the software and the communication network: they originate traffic, receive traffic, or both.

Definition — end device (host). A device that is the source or destination of a message on the network. It has an address and runs the applications that people use. A host may act as a client, a server, or both.

Hosts are not limited to computers. The modern network carries a huge variety of end devices:

End deviceRoleExample
PC / laptop / tabletGeneral client computing; requests services and runs user applications.A workstation browsing the intranet
SmartphoneMobile client; voice, messaging, and data apps.A phone on the office Wi-Fi
ServerProvides a service to many clients at once (web, email, file, DNS).A file server in the data centre
Network printerReceives print jobs from clients over the network.A shared office printer
IoT deviceInternet of Things endpoint: a small embedded sensor or actuator that sends or receives data automatically.IP camera, smart thermostat, badge reader

The server is simply a host whose role is to serve. The hardware can be a powerful rack-mounted machine or an ordinary PC — the word describes the job, not the box. An IoT device is often a tiny computer with a network interface and very little else; it may send one reading every minute or stream video. IoT endpoints multiply the number of hosts a network must address, which is one reason IPv6 became necessary.

End devices · pick one to see its role

Part 2

End-device addressing

For two hosts to exchange data, each must be identifiable. Every end device therefore needs at least one address, and in practice two different kinds working at two different layers.

KindAddressLayerWhat it does
LogicalIP address (IPv4 or IPv6)Layer 3Identifies the host across the whole internetwork; routers use it to deliver data between networks.
PhysicalMAC addressLayer 2Burned into the network interface card (NIC); identifies the device on the local link and moves data across one link.
Watch the direction. IP addresses get data between networks; MAC addresses get it across one local link. Encapsulation (a later lesson) shows how both are used on a single journey.

Addresses can be assigned by hand (static) or handed out automatically by a DHCP server (dynamic). Most client devices use DHCP; servers, routers, and printers usually get static addresses so that clients always find them at the same place. Whatever the method, if a host has no valid address it cannot communicate beyond its own segment.

Part 3

Network infrastructure recap

Strip away the detail and every network is built from a small set of building blocks. Learn the four now; Part 2 of this topic expands on the second one.

Building blockWhat it isExamples
End devicesSources and destinations of traffic.PCs, servers, printers, IoT
Intermediary devicesForward traffic between hosts.Switch, router, access point, firewall
Network mediaThe path the signals travel.Copper, fibre, wireless
ServicesThe useful things the network delivers.File sharing, printing, DNS, collaboration

Intermediary devices are covered in Part 2 of this topic. Everything else in this lesson — media, bandwidth, and services — is what makes the hosts worth connecting in the first place.

Part 4

The three network media

Network media is the physical path that carries a signal from one device to another. There are exactly three families you must know: copper, fibre-optic, and wireless.

Three media families · select one
Copper cable Electrical pulses ~100 m per run Fibre-optic Light pulses kilometres, immune to EMI Wireless Radio waves (802.11)

Copper cable

Copper carries data as electrical pulses. The two forms in common use are twisted-pair (the familiar Ethernet patch lead with an RJ-45 plug) and coaxial cable (used for cable-TV and cable-internet, and in older Ethernet). Twisted pairs are used in unshielded (UTP) and shielded (STP) versions; shielding resists interference but costs more and is bulkier.

Copper is inexpensive, easy to install, and ideal for connecting hosts to the local switch. Its weaknesses are attenuation (the signal weakens with distance), interference from electromagnetic noise, and a hard distance limit. Standard Ethernet over twisted-pair is limited to about 100 m per cable run.

Fibre-optic cable

Fibre carries data as pulses of light through a glass or plastic core. There are two kinds: multimode fibre, with a larger core, used over shorter distances (inside buildings and data centres), and single-mode fibre, with a very small core, used over long distances (city to city, ISP backbones).

Fibre offers enormous bandwidth, very low loss, and complete immunity to electromagnetic interference, and it can span kilometres in one run. The trade-offs are higher cost and the need for skilled installation. Fibre is the medium of choice for network backbones and for any link that must cross a long distance or a noisy electrical environment.

Wireless

Wireless media carry data as radio waves through the air (the IEEE 802.11 family is the standard for Wi-Fi). No cable is required, so devices stay connected while they move — the key advantage of mobility, and a necessity where cabling is impractical.

Wireless has real drawbacks: the air is a shared medium, so all clients on one channel compete for capacity; radio signals suffer interference, are blocked by walls, and weaken with distance; and because anyone nearby can potentially listen, security must be designed in. Wireless usually supplements a wired network rather than replacing it.

Part 5

Bandwidth vs throughput

Two words that sound alike but mean different things, and one unit trap that the exam tests every single time.

Definitions. Bandwidth is the capacity of a medium — the maximum number of bits it can carry per second under ideal conditions. Throughput is the actual rate of successful data transfer measured in practice. Throughput is normally lower than bandwidth.

The gap between the two comes from real-world effects: the time data spends waiting (latency), the number of other users competing for the link (congestion), the protocol overhead of headers and acknowledgements, and retransmissions caused by errors. A link advertised at one speed almost never delivers that speed end to end.

Both quantities are measured in bits per second (b/s). The common prefixes are:

UnitSymbolBits per second
Bit per secondb/s1
Kilobit per secondkb/s1 000
Megabit per secondMb/s1 000 000
Gigabit per secondGb/s1 000 000 000
b vs B — the classic trap. A lowercase b means bit; an uppercase B means byte. One byte is 8 bits. So Mbps (megabits per second) is not the same as MBps (megabytes per second): 1 MBps = 8 Mbps. Link speeds are quoted in bits; file sizes are usually shown in bytes. Ignoring the case is the single most common unit error on the exam.
K / M / G: 1000 or 1024? In networking, rates use the decimal (SI) prefixes: 1 kb/s = 1 000 b/s, because the prefix’s real meaning is kilo = 1000. In computer storage and memory the binary values 1024 (K), 1 048 576 (M), and 1 073 741 824 (G) are often used instead. The IEC standard fixes the ambiguity with distinct names: kibi (Ki) = 1024, mebi (Mi) = 10242, gibi (Gi) = 10243. Rule of thumb: bits per second use 1000; bytes of storage are where you meet 1024.

Worked example: a 100 MB file on a 100 Mbps link. Convert megabytes to megabits: 100 × 8 = 800 Mb. At 100 Mb/s the raw time is 800 ÷ 100 = 8 s — and the real time will be longer once overhead and contention are added. Notice how the case of the letters changes the answer completely.

Bandwidth calculator · bits vs bytes

Part 6

Network services

The reason to build all this infrastructure is the services it delivers. Four are called out explicitly in the fundamentals material — remember them as F-P-C-D.

Network services · F-P-C-D

Exam tip. Remember the four services as F-P-C-D: File, Print, Collaboration, DNS. If a question asks which service maps a name to an address, the answer is DNS — and the port is 53.
Part 7 · Your turn

Review

Check what you have learned with these review questions.

Review · 6 questions

Sources

Sources & references

  1. Cisco. CCNA (200-301) Exam Topics, Network Fundamentals (network components). cisco.com.
  2. IEEE. IEEE 802.3 (Ethernet, including twisted-pair and fibre specifications and the 100 m twisted-pair limit). standards.ieee.org.
  3. IEEE. IEEE 802.11 (Wireless LAN — Wi-Fi). standards.ieee.org.
  4. IEC 80000-13:2008, Quantities and units — Part 13: Information science and technology, and IEC 60027-2:2005 (names and symbols for decimal vs binary prefixes: k/kibi, M/mebi, G/gibi). iso.org/standard/31898.
  5. Mockapetris, P. RFC 1034: Domain Names — Concepts and Facilities (1987) and RFC 1035: Domain Names — Implementation and Specification (1987). rfc-editor.org/rfc/rfc1034.
  6. Braden, R. RFC 1122: Requirements for Internet Hosts — Communication Layers (1989). rfc-editor.org/rfc/rfc1122.

Target exams: CCNA 200-301 v1.1 · CyberOps Associate 200-201. Last content verification: 2026-10-11.