Lesson 1: Choosing a Network Design
This lesson focuses on the first step in creating a computer network: laying
the foundation upon which your network will be built. The decisions that
you make now can make life easier for you or come back to haunt you. In
this lesson, we explore the decisions and steps that you need to take to
design a functional network.
Peer-to-Peer or Server-Based?
A company that manufactures custom-made bicycles has asked you to install
an economical computer network that will bring it up-to-date in communication
technology and be flexible enough to allow for future expansion.
The company's network goals are to:
Table 8.1 provides background information about the bicycle company.
Network the existing computers so that they can share information and printers.
Add two additional computers to the network: one for the Product Design
Group and one for the Manufacturing Department.
Allow for the possible addition of three computers at a later date.
Provide an Internet connection for the Product Design Group.
Table 8.1 Background Information on Bicycle Company
|Number of employees
||Single-story building: 245 square meters (2625 square feet)
|Current number of computers
||Five personal computers, distributed as follows:
Managing Director: Pentium III 400 MHz
Accounting Dept.: 486/200 MHz
Sales Dept.: 486/200
Shipping Dept.: 286/25
Product Design Group: Pentium II 300
||Managing Director and Product Design Group: Windows 98
Accounting and Sales Depts.: Windows 95
Shipping Dept.: MS-DOS 5.0
||The Managing Director has a modem, Internet connection,
and a color ink jet printer.
The Design Dept. has an old laser printer. The Accounting and Sales
Depts. share a second old laser printer on a switch box.
The Shipping Dept. has a dot matrix printer.
Figure 8.1 illustrates the layout of the bicycle company, including
public areas, department offices, and manufacturing facilities.
Figure 8.1 Facility layout of bicycle company
The first decision we need to make for this new network is whether it should
be a peer-to-peer or a server-based network. The factors we need to examine
Throughout the lessons in this chapter, we will build, maintain,
and expand this hypothetical computer network. As we work through the lessons,
we will discuss various applications and choose those best suited for the
needs of our fictional bicycle company. Keep in mind that the choices we
make are based on the requirements of this hypothetical company and will
not address every company's networking situation and needs.
In a peer-to-peer network, all users on the network are equal. Each will
have equal access to all other computers on the network, provided the owner
of the computer has shared that resource with the network. In a small network
or business, this "one for all and all for one" system often works well.
Size of the network.
Level of security.
Type of business.
Level of administrative support available.
Amount of network traffic.
Needs of the network users.
Often, in a small business, no individual is able to devote full-time
attention to administering the network. This brings another advantage of
peer-to-peer networks to light. Here, responsibility for running the network
is distributed to everyone, and users determine which information or resources
on their computers will be shared.
While the peer-to-peer network option looks like a good choice for our
bicycle company, it has some drawbacks that must be considered. For example,
imagine a user on the network who has the laser printer attached to her
computer. If she turns off her computer and leaves the office, no one else
will be able to use the printer. If Computer A freezes or is rebooted while
Computer B is trying to use a resource, Computer B will be disconnected.
To summarize, usually, in a peer-to-peer scenario, no system administrator
is designated, security precautions are few, and each user is responsible
for his or her own data.
Another down side of the peer-to-peer network is its limited performance.
If another user is accessing the resources on your computer, that user
will also be using processor time on your computer. Therefore, regardless
of how fast your computer's processor is or how much memory you have, the
performance of your computer will slow down when someone else is drawing
on its resources.
Even with these drawbacks, a peer-to-peer network might still appear
to be a good choice for our network. However, we should also consider the
advantages of using a server-based network before we make our decision.
On a server-based network, resources are usually centralized. For example,
one server manages all the printers, and another server manages all the
files. Because servers are rarely turned off, resources will always be
available. Server-based networks are also scalable. This means that
their size can be easily adjusted to respond to changes in the load on
Server-based networks are also more secure than peer-to-peer networks.
With a peer-to-peer network, all resources are shared equally across the
network. If the Accounting Department shares the directory that contains
the salary files so that the Managing Director can access them, everyone
else on the network can also access these files. On the other hand, server-based
networks allow for the creation of accounts and permissions that provide
for further security. For example, a server-based network can share individual
files within a directory without making the directory itself available
to everyone on the network.
As it grows, a server-based network can be segregated according to organizational
needs. For example, one server might be designated for the Accounting Department
and another server designated for the Sales Department. Should our bicycle
company's network requirements reach this level, we will need to consider
using a network that supports file-level sharing and user groups with shared
rights to network resources.
At present, the better choice for our company is to use a peer-to-peer
network. But in order to provide more flexibility and to prepare it for
further expansion, another option exists: create a hybrid network. Thus,
while our basic network will be peer-to-peer, we will install one computer
as a file server. With this approach, access to the file server requires
an account and permissions, while access to other computers on the network
is shared equally.
So, after weighing these factors, we arrive at our network-design selection
for this bicycle company: a hybrid peer-to-peer network, with one new computer
to be installed and configured as a file server and used to centralize
After deciding on the overall network design, our next step in creating
a network is to take inventory to determine what hardware and software
is already available and what needs to be acquired. As an illustration,
we turn again to our bicycle company. It has a mixture of computers, ranging
from a legacy 286 to a new Pentium III, as well as some older printers.
Thus, some obvious updating will be required to get this network up and
running. Taking inventory is an important step, because it sets the stage
for future network expansion. For example, if all your computers run Microsoft
Windows 95 or Windows 98, you will be limited to using a peer-to-peer network.
To upgrade to a server-based network in the future, you will have to upgrade
one of the computers to run NetWare or Windows NT or add a new server with
one of those network operating systems installed.
To take inventory, you'll need to survey four categories:
This is actually a simple process, but one that should not be taken lightly.
Begin by recording the specifications of each computer; the details you
gather at this stage can save time in the long run. As we will see later,
in order to function effectively, networks often require that hardware
and software meet certain minimum standards. If you know the specification
details of the available equipment in advance, you can prevent many problems
For each computer, you will need to gather information, including:
Make a list of the manufacturer and model number for any peripheral devices,
such as printers, plotters, and scanners, whether they are installed or
simply sitting on a shelf. For each of these, note whether you have the
original disk with drivers.
Make and model.
Processor manufacturer and speed.
Amount of memory (RAM) installed.
The size and manufacturer of each hard drive.
Details of any other installed drives, such as compact-disc and removable
Monitor—make, model, and size.
Video card—make, model, and amount of memory.
Any installed peripherals.
Type of bus—EISA, Micro Channel, ISA, or PCI—the computer uses and whether
there are any free slots; you will need free slots to install network interface
cards. (For more information on bus architecture, refer to Lesson
2: The Network Interface Card, in Chapter 2.)
Be aware of all the software currently in use throughout the potential
network. For example, if you were to convert all the computers to Windows
NT while you were installing the new network, you might find that some
of the old standby programs, once used on a daily basis, now no longer
run. Be especially careful when evaluating custom-designed and proprietary
programs, such as accounting databases, that have been written especially
for the company. You might need to contact the manufacturer for information
about running proprietary programs on the network. Not all of these will
run in a network environment; the product-licensing arrangement might not
allow network operations.
For each software program, gather the following information:
As you carry out your survey of our bicycle company, also note any potential
software incompatibilities within and among company departments. For example,
the Accounting Department might be using WordPerfect, whereas the Sales
Department is using Microsoft Office. If you are planning to upgrade some
day, now is the time to make any changes needed to ensure that the same
system is used company wide.
Program version number
Availability of the original installation floppy disks or compact discs
Any licensing information
Telecommunications Equipment Survey
It might seem strange to review the existing telecommunications equipment
when you are installing a LAN, but this is actually a very important element
of your survey, especially if you intend to use Internet connections or
some form of remote access server. (Known as RAS, this is a host on a LAN
that includes modems and enables users to connect to the network over telephone
lines; RAS is discussed in Chapter 7, Lesson 2:
Connection Services.) Overlooking something as simple as the number
of phone lines wired into each office can have a major impact later if
you need modem and telephone connections at the same time. For example,
if the company has an automated telephone system, while telephone outlets
might be located in every office, they might not be capable of a modem
connection. In that case, a separate telephone outlet might be required
for voice and data communication. Also, if the company is using a high-speed
digital telephone service, you might not be able to connect with standard
modems. Don't assume a standard RJ-11 telephone jack is going to be sufficient
for you to connect a modem and start surfing the Web.
Requirements of the Network
After you have examined the existing facility and equipment, you need to
define the requirements of your network. You'll then match these requirements
to the existing hardware, software, and telecommunications features available
and determine what steps need to be taken to develop the network. At a
minimum, you should consider the following:
The size of the facility (located on a single floor vs. multiple floors)
The number of users
Whether the LAN will be extended to several buildings
The environment (office, manufacturing, out-of-doors)
The current network media, if any
The technical competence of users
The amount of network traffic (initially, and anticipated for the future)
The level of security
Building a Map
Now it's time to lay out the network. But before you begin to recommend
a network plan for our bicycle company, you will first need to make a map
of all the elements involved. During this step, you should consider two
aspects of the network: the physical layout, including the location of
each piece of hardware and how it relates to the others, and the physical
and logical topology of the proposed network.
Use a drawing or map of the facility—or make one if it doesn't exist
already—and mark the location of the existing equipment. Figure 8.2 shows
the facility drawing for the bicycle company, with the location of the
Figure 8.2 Existing equipment for the bicycle
The second step is to create a layout of the network topology. Don't
forget to include printers and other peripherals, such as scanners and
modems. Figure 8.3 shows the company network as a physical bus.
Figure 8.3 Bicycle company network as a physical
Figure 8.4 Shows the network as a physical star.
Figure 8.4 Bicycle company network as a physical
Adequately documenting a network is the key to successful troubleshooting
in the future. Start with the facility map and equipment survey.
Choosing Network Media
The choice of which media to select should not be taken lightly. The cost
of installation can be quite high, especially if you have to do it twice.
The media you choose will usually be related to the geographic requirements
of the site. For example, if several of the workstations are located in
a manufacturing environment in which a large amount of electrical noise
is generated, fiber-optic cable might be required because it is unaffected
by electrical signals. On the other hand, in a small office, simple twisted-pair
cable will usually be appropriate. The most important thing to keep in
mind is not the cost today, but the cost in the future. Being overly cost-conscious
now can limit the scalability, and thus the life span, of the network.
At our bicycle company, we might decide to install our network using
CAT 3 UTP cable. This would give us a functional network with our seven
workstations, but limit our network speed to 10 Mbps. Five years from now,
when we might have as many as 30 to 50 workstations, a 10 Mbps network
would be slow. However, by installing CAT 5 UTP now, we can upgrade our
network to 100 Mbps at any time in the future without needing to rewire
the building. And CAT 5 UTP cable costs only a few cents more per foot
than CAT 3 UTP cable.
Installing Network Media
Installing network media requires special skills and is best left to a
professional cable installer if the topology is complex. With a simple
topology, however, the necessary skills are well within our reach. Still
using our bicycle company as an example, we will next tackle the basics
of installing a CAT 5 Ethernet cabling system. (Bear in mind that at this
point we are focusing on the cables and are not yet ready to connect any
of the devices.)
The simplest way to lay out a network in our small-office environment
is to use a physical star. The focal point of the star will be a patch
panel. (A patch panel is an array of RJ-45 female connectors that
have terminals for connecting the wires.) Figure 8.5 shows a typical 12-connection
CAT 5 568B patch panel.
Figure 8.5 12-connection CAT 5 568B Patch Panel
Notice that the front of the panel has 12 RJ-45 connectors. On the back
is a terminal for each of the 12 cables. Looking closely, you will notice
that each terminal has eight connection points, one for each of the eight
wires. The wires are color-coded for easy installation. By mounting the
patch panel near the hub (usually next to the server), you can easily make
connections to the network.
On the other end of the cable, you will need to install a CAT 5 outlet.
This is similar to a standard telephone outlet (see Figure 8.6), but uses
an RJ-45 connection, just as the patch panel does. Notice that this connector
also has color-coded terminals.
Figure 8.6 CAT 5 outlet
Remember that when installing the cable, as discussed in Chapter
2, "Basic NetworkMedia," any cables that are to be run above the ceiling
or below the floor must meet local construction and fire codes. Be sure
to use plenum-grade cable where it is required.
Some CAT 5 outlets have two sets of color coding, one for 568A
and one for 568B cable termination. Be sure to match the color code of
the outlet and the patch panel.
If you plan to install a lot of cable and connections (patch panels
or CAT 5 outlets), you might want to purchase a punch down tool. (This
is a small hand tool that will ensure that the wires are properly connected
to the terminal.) Figure 8.7 shows a typical punch down tool.
Figure 8.7 Punch down tool
Connecting Media to Computers
Once the media is installed, connecting the computers is straightforward.
All connections are made through CAT 5 patch cables (prewired cables and
connectors, in a variety of solid colors). One end of the cable is secured
to the computer's NIC and the other to the RJ-45 outlet. If you are using
a hub, you will find that short patch cables connect the patch panel and
an outlet on the hub. Don't worry if you find that there is no computer
connected to an outlet and the cabling is connected all the way to the
hub. The hub will sense that the segment is unoccupied and ignore it. Figure
8.8 shows a hub connected to a patch panel and a workstation.
Figure 8.8 Installed media
Exercise 8.1: Creating a Peer-to-Peer Network
In this exercise, your task is to prepare a design of the network for our
bicycle company. All the information you need to refer to is provided in
this lesson. This exercise is presented in two parts. First define the
topology to use, and second, make a list of materials required.
Part 1: Design a Topology
Because peer-to-peer Ethernet can be implemented in the form of either
a physical bus or a physical star, you will need to look at these two different
topologies in planning for this network. On separate paper, draw:
Check your drawings against the answer figures for this exercise in Appendix
A logical bus with a physical bus.
A logical bus with a physical star.
Part 2: Select Your Materials
Make a materials list for each of the topologies. Using the facility drawing
provided earlier in Figure 8.1, identify the location of each workstation
and estimate all the materials you will need to complete the job. Use the
following table to help determine which materials are required.
||Logical bus; physical bus
||Logical bus; physical star
|Type of cable
|Length of cable
(type and count)
Save your work to use later in Exercise 2, which follows Lesson 2: Establishing
a Client/Server Environment, in which we will build on the physical star
logical bus design you have begun to build in this exercise.
The following points summarize the main elements of this lesson:
The first step in network design is to decide whether the network will
be peer-to-peer or server-based.
Take a detailed inventory of all the available networking hardware, software,
and telecommunications equipment on hand, before recommending any equipment
purchases for a new LAN.
Create a map of the facility layout and the network (logical and physical)
as the beginning of your network-documentation package.
Installing network media is difficult, and complex configurations require
a professional installer. Basic installations can be accomplished without
By knowing the specifics of each piece of hardware that will be on the
network, you can map a network design and foresee potential problems.
Addressing potential problems before implementing the network can save
time and expense.