NETWORK ELEMENTS

In the computer world, the term network means two or more connected computers that can share
resources such as data and applications, office machines, an Internet connection, or some
combination of these, as shown in Figure

Figure above shows a really basic network made up of only two host computers connected; they
share resources such as files and even a printer hooked up to one of the hosts. These two hosts
“talk” to each other using a computer language called binary code, which consists of lots of 1s
and 0s in a specific order that describes exactly what they want to “say.”

LOCAL AREA NETWORK

Just as the name implies, a local area network (LAN) is usually restricted to spanning a particular
geographic location such as an office building, a single department within a corporate office, or
even a home office. It’s still best to split a big LAN into smaller logical zones known as
workgroups to make administration easier In a typical business environment, it’s a good idea to
arrange your LAN’s workgroups along department divisions; for instance, you would create a
workgroup for Accounting, another one for Sales, and maybe another for Marketing. Figure
below shows two separate LANs, each as its own workgroup.

NB: The devices labeled hub and switch are just connectivity devices that allow hosts to
physically connect to resources on a LAN.
Notice that there’s a Marketing workgroup and a Sales workgroup. These are LANs in their most
basic form. Any device that connects to the Marketing LAN can access the resources of the
Marketing LAN—in this case, the servers and printer.

There are two problems with this:

■You must be physically connected to a workgroup’s LAN to get the resources from it.
■You can’t get from one LAN to the other LAN and use its server data and printing resources
remotely.

This is a typical network issue that’s easily resolved by using a device called a router to connect
the two LANs, as shown in Figure below

You can use routers for more than just connecting LANs, the router shown in Figure above is a
great solution because the host computers from the Sales LAN can get to the resources (server
data and printers) of the Marketing LAN, and vice versa.
Connecting the two workgroups with a type of cable instead of the route would result in having
only one big, cumbersome workgroup instead of separate workgroups for Marketing and Sales,
and that kind of arrangement isn’t practical for today’s networks.
This is because with smaller, individual-yet-connected groups, the users on each LAN enjoy
much faster response times when accessing resources, and administrative tasks are easier, too.
Larger workgroups run more slowly because there are many hosts within them that are all trying
to get to the same resources simultaneously. So the router shown in Figure which separates the
workgroups while still allowing access between them.
Common Network Components

There are a lot of different machines, devices, and media that make up our networks. The three
of the most common are:
■ Workstations
■ Servers
■ Hosts

Workstations

Workstations are powerful computers that run more than one central processing unit (CPU) and
whose resources are available to other users on the network to access when needed. Workstations
are often employed as systems that end users use on a daily basis.

Clients

Don’t confuse workstations with client machines, which can be workstations but not always.
People often use the terms workstation and client interchangeably. In colloquial terms, this isn’t
a big deal; we all do it. But technically speaking, they are different. A client machine is any
device on the network that can request for resources on a network. Workstations can be clients,
but not all clients are workstations. For example a printer can request resources from the
network, but it is a client, not a workstation.

Servers

Servers are also powerful computers. They get their name because they truly are “at the service”
of the network and run specialized software known as the network operating system to maintain
and control the network

In a good design that optimizes the network’s performance, servers are highly specialized and
are there to handle one important labor-intensive job. This is not to say that a single server can’t
do many jobs, but more often than not, you’ll get better performance if you dedicate a server to a
single task. Here’s a list of common dedicated servers:

File Server Stores and dispenses files

Mail Server The network’s post office; handles email functions
Print Server Manages printers on the network
Web Server Manages web-based activities by running Hypertext Transfer Protocol
(HTTP) for storing web content and accessing web pages
Fax Server The “memo maker” that sends and receives paperless faxes over the network
Application Server Manages network applications
Telephony Server Handles the call center and call routing and can be thought of as a
sophisticated network answering machine
Proxy Server Handles tasks in the place of other machines on the network, particularly
an Internet connection.
Servers are usually dedicated to doing one specific important thing within the Network, but not
always, sometimes they have more than one job. But whether servers are designated for one job
or are network multitaskers, they can maintain the network’s data integrity by backing up the
network’s software and providing redundant hardware (for fault tolerance). And no matter what,
they all serve a number of client machines.
Servers must have considerably superior CPUs, harddrive space, and memory—a lot more than a
simple client’s capacity—because they serve many client machines and provide any resources
they require. Because they’re so important, you should always put your servers in a very secure
area. My company’s servers are in a locked server room because not only are they really pricey
workhorses, they also store huge amounts of important and sensitive company data, so they need
to be kept safe from any unauthorized access.
In the following figure you can see a network populated with both workstations and servers. Also
notice that the hosts can access the servers across the network, which is much the general idea of
having a network in the first place!
Note that there are more workstations here than servers because one server can provide resources
to what can sometimes be a huge number of individual users at the same time but workstations
don’t.

Hosts

This can be kind of confusing because when people refer to hosts, they really can be referring
to almost any type of networking devices—including workstations and servers. But if you dig a
bit deeper, you’ll find that usually this term comes up when people are talking about resources
and jobs that have to do with Transmission Control Protocol/Internet Protocol (TCP/IP). The
scope of possible machines and devices is so broad because, in TCP/IP-speak, host means any
network device with an IP address. Yes, as an IT professional stick to the definition being
network devices, including workstations and servers, with IP addresses.
Wide Area Network

There are many people who, if asked to define a wide area network (WAN), just couldn’t do it.
Yet most of them use the the Internet—every day! With that in mind, you can imagine that WAN
networks are what we use to span large geographic areas and truly go the distance. Like the
Internet, WANs usually employ both routers and public links, so that’s generally the criteria used
to define them.

Here’s a list of some of the important ways that WANs are different from LANs:

■ WANs usually need a router port or ports.

■ WANs span larger geographic areas and/or can link disparate locations.
■ WANs are usually slower.
■ We can choose when and how long we connect to a WAN. A LAN is all or nothing, our
workstation is connected to it either permanently or not at all, although most of us
have dedicated WAN links now.
■ WANs can utilize either private or public data transport media such as phone lines.

We get the word Internet from the term internetwork . An internetwork is a type of LAN
and/or WAN that connects a bunch of networks, or intranets . In an internetwork, hosts still use
hardware addresses to communicate with other hosts on the LAN. However, they use logical
addresses (IP addresses) to communicate with hosts on a different LAN (other side of the router).
And routers are the devices that make this possible. Each connection into a router is a different
logical network. Figure below demonstrates how routers are employed to create an internetwork
and how they enable our LANs to access WAN resources.
The Internet is a prime example of what’s known as a distributed WAN—an internetwork that’s
made up of a lot of interconnected computers located in a lot of different places. There’s another
kind of WAN, referred to as centralized, that’s composed of a main, centrally located computer
or location that remote computers and devices can connect to. A good example is remote offices
that connect to a main corporate office, as shown in Figure above.

MPLS (CLOUD)

Multiprotocol Label Switching (MPLS) is one of the most popular WAN protocols in use today.
MPLS has become one of the most innovative and flexible networking technologies on the
market, and it has some key advantages over other WAN technologies:

■■ Physical layout flexibility

■■ Prioritizing of data
■■ Redundancy in case of link failure
■■ One-to-many connection

MPLS is a switching mechanism that imposes labels (numbers) to data and then uses those labels
to forward data when it arrives at the MPLS network, as shown in Figure below

The labels are assigned on the edge of the MPLS network, and forwarding inside the MPLS
network (cloud) is done solely based on labels through virtual links instead of physical links.
Prioritizing data is a huge advantage; for example, voice data could have priority over basic data
based on the labels. And since there are multiple paths for the data to be forwarded through the
MPLS cloud, there’s even some redundancy provided as well.
Network Architecture: Peer-to-Peer or Client-Server?

We’ve developed networking as a way to share resources and information, and how that’s
achieved directly maps to the particular architecture of the network operating system software.
There are two main network types you need to know about: peer-to-peer and client-server. And
by the way, it’s really tough to tell the difference just by looking at a diagram or even by
checking out live video of the network in operation. But the differences between peer-to-peer
and client-server architectures are major. They’re not just physical; they’re logical differences.

Peer-to-Peer Networks

Computers connected together in peer-to-peer networks do not have any central, or special,
authority—they’re all peers, meaning that when it comes to authority, they’re all equals. The
authority to perform a security check for proper access rights lies with the computer that has the
desired resource being requested from it. It also means that the computers coexisting in a peer-to-
peer network can be client machines that access resources and server machines and provide those
resources to other computers. This actually works pretty well as long as there isn’t a huge
number of users on the network, if each user backs things up locally, and if your network doesn’t
require much security. If your network is running Windows, MacOS, or Linux in a local LAN
workgroup, you have a peer-to-peer network. Figure below gives you a snapshot of a typical
peer-to-peer network. Keep in mind that peer-to-peer networks definitely present security-
oriented challenges; for instance, just backing up company data can get tedious.

Since it should be clear by now that peer-to-peer networks aren’t all sunshine, backing
up all your critical data may be tough, but it’s vital because security is not centrally governed,
each and every user has to remember and maintain a list of users and passwords on each and
every machine. Worse, some of those all-important passwords for the same users change on
different machines—even for accessing different resources.

Client-Server Networks

Client-server networks are the polar opposite of peer-to-peer networks because in them, a single
server uses a network operating system for managing the whole network. Here’s how it works: A
client machine’s request for a resource goes to the main server, which responds by handling
security and directing the client to the desired resource.
This happens instead of the request going directly to the machine with the desired resource, and
it has some serious advantages. First, because the network is much better organized and doesn’t
depend on users remembering where needed resources are, it’s a whole lot easier to find the files
you need because everything is stored in one spot—on that special server. Your security also gets
a lot tighter because all usernames and passwords are on that specific server, which is never ever
used as a workstation. You even gain scalability—client server networks can have many
workstations on them. And surprisingly, with all those demands, the network’s performance is
actually optimized. The figure shows a client-server network with a server that has a database
of access rights, user accounts, and passwords.

Many of today’s networks are hopefully a healthy blend of peer-to-peer and client server
architectures, with carefully specified servers that permit the simultaneous sharing of resources
from devices running workstation operating systems. Even though the supporting machines can’t
handle as many inbound connections at a time, they still run the server service reasonably well.
And if this type of mixed environment is designed well, most networks benefit greatly by having
the capacity to take advantage of the positive aspects of both worlds.