Welcome to the OUTWARD course “Computer Network

Overview ” ! The estimated runtime of this course is 60

1
Here you see how to navigate within the course.

2
3
The learning objectives for this course are as shown here.
You will learn why networks are connected and how data is
transmitted. You will also learn how our enterprise networks
are used.
You will learn about the network architecture of companies.
You will learn about the main hardware components used to
configure networks.
And finally, you will learn about the latest conditions related
to constantly evolving networks.
This course helps to develop an understanding of basic
network environments.

4
Networks have become an essential part of our day-to-day
lives.
Now, anyone can easily connect to a network not only
through a wired connection at home or in the office, but also
through wireless communication outdoors.
In this course, you will learn about computer networks.
In particular, we will provide an overview of software aspects
and hardware aspects to provide an understanding of the
mechanism of networks.
Then, you will learn what networks are and how networks
communicate.

5
This lesson describes what networks are and how data is
transmitted in a network.

6
A network is a number of computers and peripherals linked
together so that data can be passed between them.
The rise in the popularity of networks has led various
approaches to network construction.
To help distinguish between different approaches, engineers
categorize networks according to four elements.
The four elements are network size, the model or
architecture of how the network functions, the topology,
meaning how the networks are arranged, and the protocols
used, which are the set of rules they observe when
communicating.

7
The most common forms of data transmission in a network
use frames and packets of data.
A packet is a block of data at the network layer and is
destined from one point in the network to another point.
A frame is the unit of transmission of data at the data link
layer of the network.
Packets of data are encoded on top of frames, or in other
words frames are used to transfer packets of data across the
network.
The packets are transmitted via cabling or a wireless
communication medium used by the network.
In order to identify the destination of the data packet, an
address is attached to it.
The address is a unique identifier of the destination point,
whether it is a workstation, printer, or other network
hardware component.
In this example, a work station whose address is .002 sends
packet data to a workstation whose address is .003.
The destination address attached to the packet ensures that
it is directed to the right part of the network.

8
Only the addressee will be able to receive it.
The structure and addressing of packets is determined by the
network protocols the operating system implements.

8
Test your knowledge in a quiz!

9
A network is a number of computers and peripherals linked
together so that data can be transferred.
Because there are various ways to create networks, they are
classified based on four items: size, network architecture,
topology and protocols.
Data transmission is generally carried out using frames and
packets, and packets contain address information for
identifying the destination.

10
This lesson explains the technology used for network
connections.

11
Networks are generally grouped into three categories based
on their size: LAN, MAN and WAN.
A LAN is a group of two or more workstations that share a
common communications line in a small, defined area such
as a home, office or building.
A LAN has the flexibility to connect as few as two
workstations in a home or over a thousand workstations in a
company’s office.
A WAN connects computers that are geographically remote
each other to form a broader network, and interactively
connect LANs and other WANs.
This enables the creation of remote offices by connecting the
head office with the LANs of branch offices.
A 2-wire connection or satellite connection are used to
connect services, usually provided by telecommunications
carriers.
A MAN is a network positioned somewhere between these
networks, and is a network system for interconnecting LANs
on a city level.

12
LANs have become extremely popular because of their
various applications for business and personal use due to
their flexibility.
A LAN can be used to share files, hardware resources and
other network resources such as printers or mail servers.
There are many different approaches to setting up a LAN,
but the principal types you are likely to encounter are
explained in 2.2 Topology and Lesson 3: Protocols.

13
A WAN joins together LANs and other WANs to form a
broader network covering large and often remote areas.
WANs often use a 2 wire connection or satellite connection
to link together smaller, more localized networks.
The network that mobile phone companies provide
nationwide can be said to be a WAN.

14
MANs refer the networks that are larger than LANs and
smaller than WANs.
However, there is no definite boundary that can be used for
classification. This network is also used as a campus
network nowadays.
Applications include use in the entire area of a single
university, the entire area of a single company and the entire
area of a single city.
Optical fiber cable is used for the backbone and is
constructed with L2 or L3 switches using RJ-45 Ethernet.
L2 switches work as a hub which has a bridge function. L3
switches specify the route of data.

15
Topology is the underlying structure of a network.
Physical topology refers to how a network is physically
connected, and logical topology refers to how data is
transmitted around the network.
A network’s physical topology may be of one kind, and its
logical topology of another kind. For example, a network with
a physical star topology may have a logical bus topology. In
topology, the workstations, servers and other devices
forming the network are called nodes.

16
The physical topology refers to the way networks are
physically interconnected at the hardware level.
This affects how robust the network is and determines how it
reacts if, for example, a workstation fails.
Three typical topologies are the star, bus and ring types.
These are not always used alone and there are also hybrid
topologies and mixed topologies that combine several
network structures.
A WAN is a very common example of a hybrid network
because it is connected to a combination of different types of
networks.
Click the information button to display details on each
topology.

17
In a star topology, workstations are connected together via a
central distribution point.
In most LANs a specific device, called a network switch,
serves as the central distribution point.
Physical star topologies allow the network to be more robust
against cabling failures or workstations going offline.
Also, it is easier to extend a star network than other
topologies. Because of these factors, most networks run with
physical star topologies.
The drawback is that the entire network is affected when the
central switch or hub fails.

18
A physical bus topology is a system in which devices are
connected to a single cable called a bus or backbone.
The backbone is usually terminated by a special device
called a terminator.
Terminators are necessary to prevent signals bouncing back
into the bus after they have reached either end of the bus.
When a bus fails, all communication on the network is lost.
The physical bus topology is now hardly used in general
networks.

19
In a physical ring topology, each workstation is connected to
two other workstations.
A workstation can only receive data packets from one
direction and pass it on in the other direction.
A ring topology allows each workstation to boost the signal,
thus increasing normal cable distance restrictions.
The drawback is that when any of workstations fail, all
communication is lost. The physical ring topology is now
hardly used in general networks.

20
The logical topology determines how workstations
communicate with each other and is defined on a software or
functional level.
The two dominant configurations are bus and ring.
A logical bus topology can operate in either a bus or star
physical configuration, and messages are simultaneously
sent to all workstations.
Modern networks use network switches that only route
packets to the recipient workstations.
This improved behavior, however, does not affect the logical
bus topology.
In a logical ring topology, packets are transmitted
sequentially around the network in a given order.
Physical ring topologies suffer the disadvantage that every
node is essential to the network functioning well.
Failure of any node results in the failure of the network.

21
Ethernet is a standardized way of connecting computers to
create a network.
It specifies what kind of cables to use, how to connect the
cables together, how long cables can be, how computers
transmit data to one another using these cables and more.It
also specifies the maximum transfer speed.
Because there are rules for the naming of standards, a
certain amount of information can be obtained from the
name.

22
Ethernet works on the following principle.
If a workstation has a message to send, it first checks to see
if the line is free.
If the line is busy, the transmitter waits for a random period
of time and then tries sending the message again.
It then transmits a data packet containing the address and
message information onto the network.
Each workstation receives the message simultaneously.
The workstation whose network address matches that
attached to the message accepts the message, while all
other workstations simply disregard it. This method is called
the broadcasting.

23
Broadcasting presents a security risk, as every workstation
potentially has access to every message sent over the
network.
Data collision is also a risk with Ethernet systems, as two
workstations may start to transmit messages simultaneously.
When a collision occurs, the data is broken and cannot be
read.
This problem is dealt with in Ethernet systems by
reattempting the transmission of the data that was affected
by the collision.
In modern LANs where Fast Ethernet is used, it is typical to
deploy switches rather than hubs for interconnecting devices
to one another.
Switches are more intelligent than hubs and only transmit
data to the computer it is destined for, therefore alleviating
security issues, as not every workstation on the network is
able to receive all transmitted data.

24
CSMA/CD provides instructions on how to send data and how
to deal with collisions.
The following procedure is used when sending data.
1. Carrier Sense is a process for confirming that another node
is not transmitting before commencing transmission.
Data is sent if no other nodes are transmitting, and data is sent
after waiting for a random period of time if another node is
transmitting.
2. Multiple Access means that multiple nodes share the same
line and multiple pieces of data can be present on the same
line at once.
Each node compares its own address data with the destination
address, and only reads data addressed to itself, while
discarding all other data.
3. Collision Detection destroys data if a collision occurs within
a cable when multiple nodes are transmitting at the same time.
Because abnormal signal patterns and voltage levels occur
when data collides, each node detects collisions based on
these.
Transmission is promptly terminated when a collision is
detected, and transmission of data is resumed after waiting for
a random period of time.
25
There are a number of Ethernet network variants. Each
standard is defined by factors such as the cable type,
transfer speed and number of nodes.
10Base5 was standardized by IEEE in 1983, and became
the first Ethernet standard to become popular in the private
sector.
10Base5, 10Base2 and 10Base-T were standardized, in that
order. A bus topology using a co-axial cable was used until
10Base2.
A star topology using a Unshielded twisted pair cable, or
UTP cable for short, and a hub similar to the modern
configuration has been used since 10Base-T. Since then,
100BaseT, known as Fast Ethernet, has become standards,
enabling transfers at 100 Mbps, as has Gigabit Ethernet,
enabling transfers at 1Gbps.
Now, 1000BaseT is the most widely used standard in
general environments.
There are also standards for 10Gbps or more, but they are
currently only used in certain data centers and have not
become widespread.

26
Click the information button to display details on each
standard.

26
A wireless network uses radio waves to create a LAN without
cables.
There are several ways to connect, and infrastructure mode
is the wireless LAN format most widely used now.
It is a star network where communication is carried out
through an access point without direct connections between
clients.
Ad-hoc mode provides direct peer-to-peer communication
between clients.
It enables direct exchanges of data and printing on printers
without an access point, but the configuration is somewhat
complicated.
Wi-Fi Direct uses a software approach in which one device
has an Access Point function and behaves as the AP, and
the clients directly communicate with each other. Devices
negotiate which one will act as the AP on the first connection.
It is a direct connection, but configuration of the connection
and security can use infrastructure mode because it is seen

27
as a normal access point by clients.
A commonly used wireless network standard is IEEE802.11x,
which is defined by the Institute of Electrical and Electronics
Engineers, or IEEE.
The OSI reference model described in Lesson 3 and the
IEEE standard map to each other as shown in the table.

27
SSIDs are identifiers for access points in a wireless LAN.
They can be set for each access point using up to 32
alphanumeric characters.
Access points are identified based on this name when
connecting to a wireless LAN.
A wired LAN cannot be joined without connecting a LAN
cable, but wireless LAN can be accessed or intercepted from
outside within the range of the signal. This problem is
prevented by encryption. At present there are three types of
encryption: WEP, WPA and WPA2.
However, the use of WEP and WPA is not recommended
because the encryption has already been broken.
WPA2 is an improved version of WPA, and is currently the
most secure because it uses the more robust Advanced
Encryption Standard, or AES for short, for encryption. It can
also communicate with devices supporting WPA because it
has backward compatibility with WPA.

28
Ethernet protocols use CSMA/CD. However, wireless
networks use CSMA/CA because CSMA/CD cannot be
implemented.
Under the control of CSMA/CA, the station needs to listen to
the channel during a preset time in order to confirm whether
other stations are transmitting.
When there are other activities, the channel is called “busy”
and the station continues to listen until the channel becomes
“idle”, or free status.
When the channel becomes idle, the station is allowed to
transmit finally.
However, an inevitable collision sometimes occurs when the
station cannot detect the busy status correctly because of
obstructions that are difficult for radio waves to pass through.

29
The IEEE802.11x family of standards used now is shown in
the table.
MIMO is an acronym for Multiple Input Multiple Output, and
has been adopted in 802.11n and ac to significantly increase
transfer speeds.
MIMO is a technology that simultaneously transmits and
receives using multiple antennas to increase transfer speed.
4x4 refers to having four antennas for transmitting and
receiving respectively, and enables communication at
approximately four times the speed without MIMO.
802.11ac/ad is a 5th-generation standard enabling extremely
high-speed transfers, which is also referred to as 802.11vht
or Gigabit Wi-Fi.
802.11ac uses 5 GHz like 802.11a/n and is backwards
compatible.
802.11ad uses the extremely high frequency of 60GHz,
making it susceptible to obstacles and giving it an extremely
short range of around 10 meters. However, because it has a

30
large channel width, high-speed communication is possible
without using MIMO.

30
Token ring network technology, developed by IBM, is based
on a logical ring topology.
The data collisions experienced on Ethernet systems do not
occur in a token ring.
The token ring system works by circulating a token, which is
an empty packet, in one direction only.
When a workstation is ready to send a message, it has to
wait until it receives the token.
This prevents two senders from transmitting simultaneously.
The sender copies its message, including the data, address
and source, into the empty packet. When data packet is filled,
its status is set to “full”.
The full data packet circulates around the ring until it reaches
the recipient.
The recipient copies the message and sends a response
frame to the sender.
Once acknowledgment of receipt has been delivered, the
status of the token is set to “empty” and it continues to

31
circulate, ready to carry another message. As a result, most
token ring systems have been replaced by Ethernet.

31
With the early token release method, the node that sent data
releases the token after a certain period of time without
waiting for arrival confirmation from the receiving node. This
mechanism increases speed by circulating multiple tokens
and data on the network.
FDDI is a standard for a dual token-ring network using
fiberoptic cable to transmit data at up to 100Mbps.
The high speed of FDDI makes it suitable as a link between
LANs or in high-performance networks transferring large
amounts of data.
By using fiberoptic cable, an FDDI network can cover great
distances of up to 60 kilometers between nodes without loss
of signal.
The same type of network can run over shielded and
unshielded twisted-pair cabling for shorter distances, and is
known as CDDI.

32
Test your knowledge in a quiz!

33
Networks can be classified into LAN, MAN and WAN
depending on their size, and the format used for their
respective connections differ greatly.
To connect network hardware components, it is necessary to
understand both the hardware aspect such as cabling and
the logical aspect such as the path data actually follows.
Ethernet is currently the most widely used connection
standard. Although there are technical issues such as data
collisions and security, these have been resolved now,
resulting in the standard becoming widespread.
The use of hubs and the mechanism for dealing with
collisions are used in current wireless network access points
and CSMA/CA, forming the basis for the latest systems.

34
Cables, cable connectors, and hubs and traffic management
hardware for connecting numerous workstations are
essential for connecting to networks. This lesson introduces
the hardware required for creating networks.

35
A NIC is a device that allows an individual hardware
component to communicate with a network.
The card translates and controls information flow between
the device and the network.
Physically, NIC’s are usually electronic component cards that
are slotted into the computer’s motherboard.
The card is connected to the network either by a physical
connection, or by wireless systems.
However, most recent new computers have a built-in NIC on
the motherboard.

36
Each NIC has its own address, and other nodes are able to
communicate by identifying the address.
This address, called the MAC address or, less commonly,
the NIC address, is a 12-digit hexadecimal unique identifier.
The first six digits of the MAC address are assigned by IEEE
and identify the manufacturer. The remaining six digits are
unique numbers for identifying individual devices and are
assigned by the manufacturer. This identifier should not be
changed.
The IP address is a virtual address, identifying an entity in a
network, and can be allocated to any node within the
network, but the MAC address is assigned to a separate
physical device.
In a TCP/IP network, the component of the IP address that
identifies the host device needs to be mapped to the MAC
address of the individual machine. The NIC monitors network
traffic for packets addressed to its unique MAC address, only
decoding those packets that match.

37
Hubs perform the task of interconnecting network devices.
Ethernet often operates in a physical star topology, based on
a hub, with a logical bus topology.
Hubs are also widely used in networks of all topologies to
extend the network by acting as the central communication
point distributing signals to a new group.
Centralizing the network’s wiring is an advantage, making it
easier to manage a large and potentially complicated cabling
system.
When physically relocating workstations, a workstation can
be removed from one part of the system and simply plugged
into a different part of the network. Physically, a basic hub is
a collection of connectors and switches.
In current LANs, switches are generally used instead of hubs
as devices interconnecting different network nodes.

38
Passive hubs deliver transfer signals without doing anything
to each workstation.
In this configuration, the distance between the hub and
workstation is limited, and the danger of losing the signal is
eliminated.
However, if the connection distance is long the signal
strength becomes weaker.
An active hub, or repeater, boosts the signal strength before
re-transmitting it, allowing connection distances to be
extended.

39
In star networks connecting multiple devices through line
concentrators such as hubs, the network can be expanded
by connecting line concentrators. This method is called
cascading.
However, there are limitations to extending a network and
the limitations are best explained by the IEEE’s “5-4-3” rule.
There may be no more than five repeated segments
between any two Ethernet network devices.
No more than four repeaters can be used between any two
Ethernet network devices.
A maximum of three of the five repeated cable segments can
be populated.
If the network cascades through repeating hubs, this rule
has to be observed to ensure point to-point communication.
IEEE’s 5-4-3 rule does not apply to the ends on switches or
switch ports.

40
Dual speed hubs have the ability to negotiate network traffic
between segments operating at different speeds.
For example, different speeds such as 10BASE-T,
100BASE-T, Fast Ethernet and Gigabit Ethernet are able to
be recognized and operated on a single hub.
Dual speed hubs use speed-sensing to interpret Fast Link
Pulses interspersed between the packets, and set the port to
the appropriate speed for the incoming traffic. An internal
switch connects between the 10Mbps and 100Mbps domains.
LED indicators on the front panel of the hub show the speed
of each port.
A dual speed hub can be cascaded with other 100 Mbps and
10M bps hubs to provide more ports.
Intelligent hubs communicate network management
information, thus increasing overall LAN efficiency.
They have SNMP agent functionality and can also provide
bridging and routing functions.

41
There are a number of devices used to direct traffic from one
node to another within a network or across multiple networks.
The devices you are most likely to encounter are as follows.

42
Bridges are devices used to link separate LANs or LAN
segments that operate on the same data link layer of the OSI
model.
They do not convert network data between protocols as they
operate only at the data link level.
If two separate LANs are connected with a bridge they
behave as one network.
Bridges serve to reduce network traffic by filtering broadcast
messages. They contain a list of workstation addresses.
If the data packet address matches any of the addresses on
the list, the message is transmitted to that workstation.
If there are no matching addresses, the packet is forwarded
to the next bridge, where the process is repeated.
Filtering can be performed more efficiently because the
bridges learn which address belongs to which network and
can develop a database.
Bridge features are simpler and cheaper compared to
routers.

43
A switch is a device that can interconnect network devices
and direct information from one data path to another.
A switch can usually connect several different networks
together, including networks using different architectures.
In practical terms switches are drop-in replacements for
hubs.
Switches vary in their level of sophistication.
A layer 2 switch operates as a hub with a bridge function, but
it cannot be configured. It only connects to the network with
the power on.
A layer 3 switch incorporates decision-making algorithms
and provides routing functions.
They typically have a built-in operating system and can be
configured on the fly with security, access control and other
settings.

44
A router is an intelligent device that connects two or more
LANs, which send data packets from nodes on one network
to nodes on a different network. Routing is a function
associated with the network layer in the OSI model.
Routers can send data between networks using different
architectures and different low-level protocols, however, they
are restricted to operating between networks using the same
high level protocols.
Routers determine which path a data packet should be sent
along in order for it to most efficiently reach its final
destination.
They make the determination according to knowledge of the
state of each network path such as free, blocked and busy.
Routers work to reduce traffic on the network by only
allowing data operating with certain protocols to enter the
LAN.

45
A gateway is a connection between a LAN and another
system. For example, it may be used between a LAN and a
mainframe computer or a larger network such as the Internet.
Gateways perform protocol conversions to transmit
messages between these systems. Generally, they are
slower than bridges or routers.
In typical business networks that are homogeneous,
gateways are essentially routers that interconnect the
company’s LAN to the Internet, its WAN or other external
networks.

46
A peripheral is any computer device that can be connected
to and controlled by a computer.
However, is not part of the integral functioning of the
computer in the way that the CPU, RAM memory or the data
path are.
Some peripherals such as hard disk drives may be located
within a computer housing.
Others are external to the computer, for example, printers
and scanners.
One of the major advantages of a LAN is its ability to
efficiently share peripheral devices between networked
workstations.
Networked printers have increased in popularity because
multiple printers can be connected to a single network,
serving different needs.

47
Cabling is an essential part of any LAN. Cabling must not
only link workstations together but also link peripherals.
There are many different types of cabling available, such as
twisted pair cable and co-axial cable, each having particular
advantages and disadvantages.
Connectors are the physical link between network hardware
and cabling.
The type of connector used depends upon the components
to be coupled and the topology of the network.
A connector can serve a number of functions. For example,
connectors may connect equal or near equal segments of
co-axial cable, connect different types of cabling such as co-
axial and twisted pair, terminate a cable or terminate a cable
to ground.

48
Twisted pair cabling is by far the least expensive and most
commonly used type of LAN cabling.
Twisted pair consists of two separate insulated wires twisted
together, which ensures that each cable receives the same
amount of outside electrical interference.
Cables can be shielded or left unshielded from electrical
interference.
STP cables are more resilient to noise and offer better
performance.
However, the LAN port must support STP to provide such
performance, and UTP may be more stable in some cases.
The advantages of twisted pair cabling are ease of
installation, cost effectiveness and the ability to easily add
new users to the system.

49
Co-axial cable, or coax, consists of a central copper core
encased in a plastic sheath.
It is surrounded by copper or aluminum sheathing and finally
encased in plastic.
The signal is transmitted along the central copper core,
shielded from electrical interference by the sheath.
The main advantages are insensitivity to outside electrical
interference and performance over long distances.
Co-axial cable is almost as easy to install and maintain as
twisted pair, however, it is awkward to install in small, tight
places.
Co-axial cable is now less popular than it was, having been
overtaken by newer technologies such as fiberoptic cable.

50
Fiberoptic cabling works by transmitting light pulses instead
of electrical signals.
A device called a light emitting diode, or LED, converts
electrical signals into a light source.
At the reception node the light signal is translated back to
electrical signals for the workstation to process.
Cables are bundled in groups of up to 24 fibers but generally
groups contain only two to four pairs.
Fiberoptic cable has a number of advantages over twisted
pair and co-axial cabling. It can transmit data at high rates,
normally 1 Gbps or more.
It is also completely immune to electrical interference and its
low resistance means that cabling distances can be large.
However fiberoptic cabling is expensive to install and it is
difficult to add new workstations once the network has been
installed.
These are the main barriers to wider use of fiberoptic
technology in smaller networks.

51
RJ connectors are used for twisted pair cables such as
attaching telephone cables to modems.
The most popular are the RJ11 and RJ45, with RJ11 being
used for cables of two pairs of and RJ45 for four pairs.
BNC connectors connect cable of the same type such as
thin and thick co-axial cable.
There are various theories on the unabbreviated form of
BNC, including bayonet nut connectors and British Navy
connectors.
Fiberoptic connectors are used to connect segments of
fiberoptic cable together. Installation of fiberoptic connectors
requires special equipment and know-how because a
precise connection is required to reduce signal loss.
Installation is typically performed by trained technicians.

52
Connectors are sometimes assembled together into a patch
panel. A patch panel is a mounted hardware unit containing
an assembly of port locations. Patch panels contain ports
where the cabling connects to the hardware device such as
a hub.
Typically, a network uses a patch panel as a sort of
switchboard, using cables to interconnect all of the
computers within the area of a LAN.
The interconnecting cables are called patch cords.

53
Test your knowledge in a quiz!

54
A network interface card is a piece of hardware required for
a workstation to connect to a network.
Each card has a MAC address that is unique to the NIC for
identification on the network.
Traffic management hardware not only connects to different
LANs and the Internet, but also enables efficient
communication by reducing network traffic.
Peripherals work by connecting to a computer. Peripherals
can be shared between the work stations that are connected
to the network.
Each type of connection cable have strengths and
weaknesses, and need to be selectively used as appropriate.

55
Protocols are sets of rules that govern the overall network
communication process.
Just as it is impossible to have a conversation if you speak
different languages, communication is not possible if
different protocols are supported.
Protocols can be divided into hardware and communication
protocols.
Hardware protocols determine standards for network
hardware devices to maximize compatibility when modifying
networks with new pieces of hardware.
Communication protocols govern the rules about how a
communication session should be set up, carried out and
terminated.
This lesson introduces the OSI reference model used as a
standard for considering protocols and explains details on
communication protocols concerning IP, TCP/UDP, Bonjour
and NetBIOS/NetBEUI.

56
ISO, an organization that develops industry standards,
developed the OSI standard.
ISO provides a performance standard to allow the flexibility
to add and replace network devices independent of the
vendor.
OSI is a model only, and reflects a way of looking at things
rather than hard and fast rules, but it is accepted as a
standard way of understanding networks. The table shows
the OSI reference model, its content, and the related
software and standards.
OSI specifies 7 layers of protocols that communicating
systems should adopt.
They are the physical layer, data link layer, network layer,
transport layer, session layer, presentation layer, and
application layer.
Each layer’s protocol is written so that it works together with
the protocol above and directly below it.
When two different operating systems are communicating

57
with each other, each layer communicates with the
corresponding layer in the other system.

57
IP is an abbreviation for Internet Protocol, and plays an
extremely important role in today’s networks.
First, IP divides data into the packets that are the basic unit
of communication.
After that, IP transfers data by delivering these to the lower
data link layer after adding an IP header containing the
address information.
The recipient side reconstructs the packet and sends it to the
transport layer.
A packet is made up of the IP header containing information
on the data’s destination and source, and the payload storing
the content being communicated. The destination and
source information in the IP header is specified by the IP
address mentioned later.

58
An IP address is an identification number in the network
layer specified for identifying devices on the network under
the Internet Protocol.
Whereas the MAC address in the data link layer is called the
physical address, this is called the logical address.
In IPv4, which is currently the prevailing protocol, the IP
address is a 32-bit integer.
To make this easier to understand, it is divided into four sets
of 8 bits divided by dots expressed in dotted decimal notation.
This conversion can be calculated. The conversion from
binary to decimal is calculated using the sum of the weights
of each binary digit.
The conversion from decimal to binary is obtained by
continuing to divide the decimal number by two and
obtaining the quotient and remainder until the quotient is
zero. The remainders obtained are the result of conversion
into binary.

59
The subnet mask is a 32-bit number that defines how many
bits from the start of the IP address are used as the network
address.
If the subnet mask is defined as follows, the first 24 bits are
the network portion and the last 8 bits are the host portion.
There are two main ways to represent this, and the most
common is to state both the IP address and subnet mask.
This is the notation system that is based on Classless Inter
Domain Routing, so it is called CIDR notation.
However, simplified notation can also be used because it is
only necessary to specify how many bits of the subnet mask
to fill with ones.
If the network portion is different, this is treated as belonging
to a different network, and it is necessary to pass through a
router, and so on, to communicate. To use an analogy, the
network portion represents the country, and the host portion
represents the address within the country.

60
The IP addresses mentioned until now are called IPv4.
The IP address is represented using 32 bits and is able to
identify approximately 4.3 billion devices.
However, the rapid spread of the Internet has led to fears of
depletion of these IP addresses in future.
In response to this, it is scheduled to be replaced by IPv6 as
a new standard.
IPv6 is represented using 128 bits, and approximately
3.4×10^38, or 2128 addresses can be used.
In addition, improvements have been made such as stronger
security through IPSec and so on, and increased transfer
efficiency.
IPv6 notation is hexadecimal and divided into 16-bit sections
by semicolons.
The first 64 bits are the network portion and the past 64 bits
are the host portion, and in contrast to IPv4, this division
does not change.
Because IPv6 addresses are very long and tend to have

61
many zeros, they can be abbreviated according to certain
rules.
However, it must be noted that there is some variation in
notation even when these rules are followed.

61
TCP and UDP are both transport layer protocols, and bridge
data via IP in the network layer and protocols in the session
layer and higher.
TCP is a highly reliable protocol that transfers data while
exchanging information called control flags to confirm receipt,
or others.
When commencing communication, the connection is
established between the sending and receiving devices
using a procedure called a three-way handshake.
After the connection is established, the data is divided and
given TCP headers including sequence numbers, and
delivered to the network layer to be communicated.
The recipient restores the data by rearranging it based on
the sequence numbers, enabling highly reliable data transfer.
UDP does not perform confirmation as TCP does, and does
not guarantee the reliability of data.
However, it is suitable for situations that need speed and
real-time access more than reliability, such as streaming

62
playback.
In addition, TCP always uses 1:1 communication to establish
connections, but UDP is able to broadcast, making it the
perfect protocol for telephony and videoconferencing systems.

62
Bonjour is a network technology not requiring configuration
that was devised by Apple. It is mainly used in LANs to use
devices without the need for any configuration.
The main functions are automatic assignment of IP
addresses and host names, and automatically searching for
services.
Devices connected to a Bonjour network do not need a
DHCP server or a DNS server.
Bonjour is used for both Ethernet and Wireless connections,
however, our equipment mainly uses it for wireless purposes.
When devices are connected, they search for a free IP
address from the 169.254.0.0/16 range to use as their own
IP address.
Another characteristic is that all devices connected to the
network are queried for their host names, and connections
are made to the matching host name.
For example, if you wish to print something, a query about
whether there are any devices providing print services is

63
made over the LAN by multicast. Only the relevant devices
respond, enabling communication to take place.
Normally, to reduce load, host information is cached instead
of multicasting every time.

63
NetBIOS is a system enabling small-scale communication
based on the names given to each computer by users.
NetBIOS defines the fundamental configuration of how data
is passed in and out of workstations within a LAN, and it
operates at the Session layer of the OSI model.
NetBIOS’s API provides a common interface between
applications and the underlying network operating system for
the purpose of transmitting messages between nodes.
Furthermore, NetBIOS functions can be used through
TCP/IP. When doing so, the NetBIOS name is mapped to an
IP address by WINS.
NetBEUI is an extended version of NetBIOS which has
become the industry standard as part of a network’s general
communication protocol.
It was supplied with Microsoft’s LAN Manager, Windows for
Workgroups, and Windows NT.
It is generally used only in small networks because it has no
routing capability.

64
Test your knowledge in a quiz!

65
Protocols are rules for communicating with each other, and
communication is made possible by adding a variety of
information to data being transferred through multiple
protocols.
IP serves the role of identifying the nodes communicating on
a network with many nodes.
Now, there are concerns about a shortage of IP addresses
due to network advances, and it will be necessary to migrate
to the IPv6 system in the near future.
TCP and UDP are protocols on the same network layer, but
their applications differ greatly.
Whereas the objective of TCP is to correctly transfer data,
the objective of UDP is to transfer data at high speed in
multiple directions.
Now, IP is used in most networks, but there are also systems
such as NetBIOS that are not reliant on IP addresses.

66
There is a lot of network usage and architectures.
This lesson introduces the model of client/server networks
and peer-to-peer networks.
Plus, this lesson describes the roles of servers and clients
used for the effective utilization of networks.

67
The majority of substantial business network installations are
client/server systems.
In this type of network, services such as printing and file
storage and retrieval are centralized and managed by a
program called a server.
There are various types of servers. Print servers manage
print requests and file servers handle the storage and
retrieval of data.
The advantages of client/server networks are efficient use of
network resources.
This topology also helps keep the network environment
consistent and under control and offers ease of management.
However, by introducing a client/server network, the user
can expect an increase in network traffic and slower
response times.
In this case, this can be minimized by increasing network
bandwidth or by increasing the server processing speed.

68
Peer-to-peer networks provide access to files, software and
peripheral devices via communication between individual
workstations.
Each workstation has the same network status as any other,
and any workstation can initiate a communication session.
In some cases, peer-to-peer communication is implemented
by giving each workstation both server and client capabilities.
Any workstation can behave as a server to other software
and interconnect to other networks using different operating
systems.
Printing is usually accomplished by either sharing a device
physically connected to a workstation, or by a third party
utility to share a standalone device attached to the network.

69
In a client-server system, the client can request the server to
perform tasks such as file retrieval and printing.
A workstation can function as both a client and as a server,
but there are several ways to configure a client in networks
with dedicated servers.
The most popular approach places all program logic in the
client workstations and uses servers to store data and
provide network services.
Other approaches vary the division of program logic and
data stored on the server and workstations.
In a thin-client system, all storage, applications and handling
of peripheral devices are provided by the server, and the
clients only serve as terminals for accessing and remotely
controlling the server.
Therefore, clients need not have a lot of memory or
processing power, hence they are called thin.

70
A dedicated server is a permanent operating server that is
used for a certain task or service. This server is not used as
a workstation.
For non-dedicated servers, the server software is running on
a host. This server may also be used as a workstation.
A shared server is a server that may be used for different or
multiple tasks or services.

71
In the past, file management and the execution of
applications was performed by the client without heavy
reliance on the server.
However, recently the increase in wireless LAN speed and
the emergence of tablets has led to a migration to terminal-
based solutions.
For example, by virtualizing desktops and applications, it is
possible to perform work as if in the office regardless of the
place or device used.
In addition, by unifying computing resources, applications
and systems can be updated and upgraded economically.
The use of computers in business is increasingly moving
toward terminal-based thin-client systems.
Examples of terminal solutions include Microsoft Remote
Desktop Service and Citrix XenApp.
Using remote desktop services, end users can execute
applications as if they had direct access to a server.
The Citrix solution for remote access incorporates functions

72
exceeding terminal services in additional products.
For example, a complete client software application is not
required and access to applications can be easily requested
through a Web browser.

72
Test your knowledge in a quiz!

73
The client-server approach most commonly used by
enterprises now centrally manages services such as printing
and file storage and retrieval on a server. In contrast, clients
are equal in a peer-to-peer system, and act as servers as
required, while managing the network.
Servers not only manage files and printers, but are also able
to perform actions normally handled by the client side such
as execution of applications.
Now, there is a growing trend of moving toward thin-client
systems in which the client terminal only plays the role of
accessing data and applications.

74
This concludes the Computer Network Overview course. In
closing, let’s look back on what you have learned.
Networks enable communication by combining a variety of
technologies.
Each technology basically interacts with upper and lower
layers, and this is easier to understand by referring to the
OSI reference model.
Most of the use of networks in enterprises previously used
client-server systems.
However, this is moving toward thin-client systems as the
specs of servers and networks improve.
Early networks had a maximum transfer speed of around 10
Mbps despite using thick cables that were awkward to use.
However, there are now Ethernet standards with a
theoretical transfer speed in excess of 10 Gbps despite
using thin cables.
Wireless networking is becoming more advanced with the
latest standards enabling communication at Gigabit speeds,

75
and network technologies continue to evolve.

75
Congratulations! You have now completed the OUTWARD
course "Computer Network Overview”.

76