Network architecture refers to the layout of the network, consisting of the hardware, software,

connectivity, communication protocols and mode of transmission, such as wired or wireless.

There are several ways in which a computer network can be designed. Network architecture
refers to how computers are organized in a system and how tasks are allocated between these
computers. Two of the most widely used types of network architecture are peer-topeer and client/server. Client/server architecture is also called 'tiered' because it uses multiple
levels.
Peer to Peer Network:
In a peer-to-peer or P2P network, the tasks are allocated among all the members of the network.
There is no real hierarchy among the computers, and all of them are considered equal. This is
also referred to as a distributed architecture or workgroup without hierarchy. A peer-to-peer
network does not use a central computer server that controls network activity. Instead, every
computer on the network has a special software running that allows for communications between
all the computers.
Peer-to-peer is mostly used for file sharing. Here is how file sharing works: One computer user
makes some of the files on the hard disk drive available for sharing. Information on these files is
made available to the rest of the users so they can decide if they want to download one or more
of these files. Once a second user has downloaded a file, this can also be made available to the
rest of the users. So now there are two possible sources from which to download the same file.
This is how files can be spread over thousands of users, one download at time.
One of the earliest peer-to-peer file sharing networks was Napster. One of the more recent
protocols for sharing files is BitTorrent. This protocol is used by the many users who provide
links to their files on websites, such as The Pirate Bay.
A peer-to-peer network is robust in the sense that if one or several of the individual computers
stop working for some reason, the network continues to function. On the other hand, the quality
of the network depends completely on the contribution of individual participants. For example,
in the case of file sharing, if very few people make their files available, there is very little for
users to download.
While there are a number of different applications of peer-to-peer network architecture, file
sharing is by far the most popular. This also includes file sharing networks that distribute
copyrighted material, such as movies, music and books, without permission. This is against the
law in most jurisdictions. Since peer-to-peer networks lack a central control system, such file
sharing systems present a serious challenge to agencies trying to prevent this type of sharing.
Client/Server
In a client/server network, a number of network clients or workstations request resources or
services from the network. One or more network servers manage and provide these resources or
services. The clients are computers that depend on the server for data and software. Network
servers are also referred to as computer servers or simply servers. Sometimes a server is
described in terms of the specific service it provides, such as e-mail server, print server or storage
server. Some servers, however, can provide all these services.

Servers are typically computers with more processing speed, memory and hard disk space than a
regular desktop computer. The network servers run their own operating system that manages the
various network tasks as well as services that run on the network. Depending on the need for
network storage and services, a single network may only use one or a large number of servers.
Clients are hardware devices which provide end users with access to data and services on the
server. You can use these devices more or less independently. For example, you can open up
software applications, create and edit documents and save files on the local storage medium,
such as a hard disk. However, in a typical client/server network, a number of essential tasks are
not performed by the client alone. Some typical examples are:

A printer is connected to the network - to print a document, the client sends a print
request to the server.

All the e-mail messages are stored on the server - the client is used to view and manage
the messages but does not store the actual files.

Very large databases are stored on the server - the client accesses these without copying
them onto the local hard drive.

Types of Networks
There are several different types of computer networks. Computer networks can be characterized
by their size as well as their purpose.
The size of a network can be expressed by the geographic area they occupy and the number of
computers that are part of the network. Networks can cover anything from a handful of devices
within a single room to millions of devices spread across the entire globe.
Some of the different networks based on size are:

Personal area network, or PAN

Local area network, or LAN

Metropolitan area network, or MAN

Wide area network, or WAN

In terms of purpose, many networks can be considered general purpose, which means they are
used for everything from sending files to a printer to accessing the Internet. Some types of
networks, however, serve a very particular purpose. Some of the different networks based on
their main purpose are:

Storage area network, or SAN

Enterprise private network, or EPN

Virtual private network, or VPN

Let's look at each of these in a bit more detail.

Personal Area Network

A personal area network, or PAN, is a computer network organized around an individual person
within a single building. This could be inside a small office or residence. A typical PAN would
include one or more computers, telephones, peripheral devices, video game consoles and other
personal entertainment devices.
If multiple individuals use the same network within a residence, the network is sometimes
referred to as a home area network, or HAN. In a very typical setup, a residence will have a
single wired Internet connection connected to a modem. This modem then provides both wired
and wireless connections for multiple devices. The network is typically managed from a single
computer but can be accessed from any device.
This type of network provides great flexibility. For example, it allows you to:

Send a document to the printer in the office upstairs while you are sitting on the couch
with your laptop.

Upload the photo from your cell phone to your desktop computer.

Watch movies from an online streaming service to your TV.

If this sounds familiar to you, you likely have a PAN in your house without having called it by its
name.
Local Area Network
A local area network, or LAN, consists of a computer network at a single site, typically an
individual office building. A LAN is very useful for sharing resources, such as data storage and
printers. LANs can be built with relatively inexpensive hardware, such as hubs, network adapters
and Ethernet cables.
The smallest LAN may only use two computers, while larger LANs can accommodate thousands
of computers. A LAN typically relies mostly on wired connections for increased speed and
security, but wireless connections can also be part of a LAN. High speed and relatively low cost
are the defining characteristics of LANs.
LANs are typically used for single sites where people need to share resources among themselves
but not with the rest of the outside world. Think of an office building where everybody should be
able to access files on a central server or be able to print a document to one or more central
printers. Those tasks should be easy for everybody working in the same office, but you would
not want somebody just walking outside to be able to send a document to the printer from their

cell phone! If a local area network, or LAN, is entirely wireless, it is referred to as a wireless
local area network, or WLAN.
Metropolitan Area Network
A metropolitan area network, or MAN, consists of a computer network across an entire city,
college campus or small region. A MAN is larger than a LAN, which is typically limited to a
single building or site. Depending on the configuration, this type of network can cover an area
from several miles to tens of miles. A MAN is often used to connect several LANs together to
form a bigger network. When this type of network is specifically designed for a college campus,
it is sometimes referred to as a campus area network, or CAN.
Wide Area Network
A wide area network, or WAN, occupies a very large area, such as an entire country or the entire
world. A WAN can contain multiple smaller networks, such as LANs or MANs. The Internet is
the best-known example of a public WAN.
Private Networks
One of the benefits of networks like PAN and LAN is that they can be kept entirely private by
restricting some communications to the connections within the network. This means that those
communications never go over the Internet.
For example, using a LAN, an employee is able to establish a fast and secure connection to a
company database without encryption since none of the communications between the employee's
computer and the database on the server leave the LAN. But what happens if the same employee
wants to use the database from a remote location? What you need is a private network.
One approach to a private network is to build an enterprise private network, or EPN. An EPN is a
computer network that is entirely controlled by one organization, and it is used to connect
multiple locations. Historically, telecommunications companies, like AT&T, operated their own
network, separate from the public Internet. EPNs are still fairly common in certain sectors where
security is of the highest concern. For example, a number of health facilities may establish their
own network between multiple sites to have full control over the confidentiality of patient
records.
Computers in a network are connected in some logical manner, referred to as network topology
or network architecture. Learn about the different types of network architecture.
Network Topology
Computers in a network have to be connected in some logical manner. The layout pattern of the
interconnections between computers in a network is called network topology. You can think of
topology as the virtual shape or structure of the network. Network topology is also referred to as
'network architecture.'
Devices on the network are referred to as 'nodes.' The most common nodes are computers and
peripheral devices. Network topology is illustrated by showing these nodes and their connections

using cables. There are a number of different types of network topologies, including point-topoint, bus, star, ring, mesh, tree and hybrid. Let's review these main types.
Point-to-Point
Point-to-point topology is the simplest of all the network topologies. The network consists of a
direct link between two computers. This is faster and more reliable than other types of
connections since there is a direct connection. The disadvantage is that it can only be used for
small areas where computers are in close proximity.
Bus
Bus topology uses one main cable to which all nodes are directly connected. The main cable acts
as a backbone for the network. One of the computers in the network typically acts as the
computer server. The first advantage of bus topology is that it is easy to connect a computer or
peripheral device. The second advantage is that the cable requirements are relatively small,
resulting in lower cost.
One of the disadvantages is that if the main cable breaks, the entire network goes down. This
type of network is also difficult to troubleshoot. For these reasons, this type of topology is not
used for large networks, such as those covering an entire building.
Star
In star topology, each computer is connected to a central hub using a point-to-point connection.
The central hub can be a computer server that manages the network, or it can be a much simpler
device that only makes the connections between computers over the network possible.
Star topology is very popular because the startup costs are low. It is also easy to add new nodes
to the network. The network is robust in the sense that if one connection between a computer and
the hub fails, the other connections remain intact. If the central hub fails, however, the entire
network goes down. It also requires more cable than bus topology and is, therefore, more
expensive.
Ring
In ring topology, the computers in the network are connected in a circular fashion, and the data
travels in one direction. Each computer is directly connected to the next computer, forming a
single pathway for signals through the network. This type of network is easy to install and
manage.
If there's a problem in the network, it is easy to pinpoint which connection is defective. It is also
good for handling high-volume traffic over long distances since every computer can act as a
booster of the signal. On the downside, adding computers to this type of network is more
cumbersome, and if one single computer fails, the entire network goes down.
Mesh
In mesh topology, every node has a direct point-to-point connection to every other node. Because
all connections are direct, the network can handle very high-volume traffic. It is also robust

because if one connection fails, the others remain intact. Security is also high since data travels
along a dedicated connection.
This type of topology requires a lot of cables and is, therefore, expensive. Many of the
connections are also redundant since there are several different paths for data to travel from one
node to another.