4809-3 ch01.

F 5/15/01 9:45 AM Page 4

PRE-TEST
1. What factors should be taken into consideration when choosing net-
work cabling?
2. Explain the importance of examining the components that have been
delivered with a new server.
3. How should servers be organized in a rack to prevent it from tipping?
4. What is the purpose of a firewall server?
5. What hardware is needed to turn a server into a router?
6. What is the purpose of electrical grounding?
7. What is the difference between a RJ-11 and RJ-45 connector?
8. What is the purpose of a server configured as a router?
9. Why should you consult a hardware compatibility list for your OS?
10. Explain the importance of a line conditioner?
4809-3 ch01.F 5/15/01 9:45 AM Page 5

P re-installation planning is a critical stage for server installations. Before pur-

chasing server equipment, you must decide on what type of servers will be
required, and how much CPU power, memory, and hard disk drive space they will
need. Your servers need a home, and a suitable location must be scouted for proper
power sources, network availability, and physical and environmental protection.
After the equipment has been delivered, it must be checked to ensure that all the
proper components have arrived, and that they are of the proper type. This chapter
explains how to handle all these elements.

Installation Strategy
Before any server or network installation, you must have a plan in place to ensure
that the new servers will meet your specific hardware and service requirements.
Some services have dependencies, and knowing how different types of servers
interact with each other is integral in pre-installation planning. An assessment of
your current and future needs will aid you in this respect. Whether you are adding a
server to an existing environment, or creating a brand new server network, you will
need a knowledge of the different types of servers and services available.

Installation planning
Objective 1.1 Conduct pre-installation planning activities
• Plan the installation
• Verify the Installation plan

Before deciding on your choice of server equipment, you should examine your cur-
rent needs. They may be as simple as replacing an aging e-mail server, or as com-
plex as installing a large enterprise database server to host a new human resources
and accounting package. You may realize that one of your servers is currently run-
ning under capacity, and can adequately host the new program or service without
impairing its primary function. You may find that with some minor upgrades, such
as extra memory, or a new CPU, you can update an existing server to bring it to a
level needed to operate the new program or service. If your current server installa-
tions do not meet the requirements of present and future needs, you need to plan a
new server installation.

Depending on the type of service or function your server will perform, you must
carefully check for any prerequisite services that may need to be in place. In order
to function properly, a certain server may need access to an available local DNS
server to resolve host names. If you do not currently have a DNS server, you must
factor that into your planning activities, and know that you will either need to
install it on an existing machine if available, or purchase another server to perform
that function.
4809-3 ch01.F 5/15/01 9:45 AM Page 6

Although you may be only thinking about present concerns, you must consider the
possibility of future growth and capacity planning. Depending on the type of appli-
cation or service you are installing, you must plan for that system to grow, and
careful planning must be considered for the amount of CPU, memory, and disk
space you will need. A year from now, you may find that the server you purchased
is already overloaded, and if upgrading is not an option, you may have to purchase
yet another server.

All of these planning concepts must be carefully balanced with the budget available
to management. You will have to work within the constraints of certain spending
limits, while still maintaining your server plan with respect to current server avail-
ability, prerequisite services, current and future needs, and capacity planning.

Before executing your installation plan, you should spend some time to verify and
test it. Ideally, this can be done in a lab situation, where you have a small network
set up internally that does not connect to your main network. This way, you can
test how the certain applications and operating systems will react with different
services, without affecting your normal network operations. If you can’t do this, the
new server can be set up on your own network, but you should perform the tests
outside of work hours, to minimize any impact to daily operations. Test how each
of your client and server operating systems will communicate with the new server
services. Although you will not be using an exact replica of your new equipment,
you will get a good idea of how the services that will run on that new server, will
communicate and interact with the services and applications on your network.

Exam Tip For the exam, make sure that you know all of the various server and network
device functions, and how they interact with each other.

Types of servers
Before you can plan your installation, you need to know what kind of server will fit
your needs. Some of these types of servers can also be performed by dedicated net-
work devices, such as routers or gateways. Many of these types of server services
can be combined in one machine. For example, a Web server and an FTP server can
reside on the same machine if needed. Combining several services onto one server
can greatly reduce your costs, because you do not need a separate server for each
service. The following are some of the most common kinds of servers:

✦ File-and-Print server: A file-and-print server provides file storage and printing

services to clients. When client files are stored on the server, they are central-
ized in a common place that can be more easily backed up. Print servers
enable clients to send a request to a printer. The print requests are queued by
the print server for delivery to the final destination printer.
✦ Mail server: A server that stores e-mail. Clients connect to the mail server by
supplying their mail account information, with which they can send and
receive e-mail. The e-mail system may be strictly internal, but it may also send
and receive e-mail over the Internet.
4809-3 ch01.F 5/15/01 9:45 AM Page 7

✦ Application server: An application server is basically a server that acts as a

middle tier in a multi-tiered application. An application server is often used in
a Client/Server architecture, where the client and server computers work
together to perform tasks. The server typically handles storage and process-
ing of data, while the client requests these resources from the server for its
own use.
✦ Database server: A database server stores structured data in a filing system
that can be retrieved by multiple users simultaneously. These types of severs
should have large amounts of RAM and disk space to be able to store and pro-
cess the data.
✦ Backup server: A backup server is used specifically for backing up system
data. The backup server is usually attached to tape unit, and runs the backup
applications to store data from your network servers to the tape.
✦ Fax server: Combined with special faxing software, and a modem/telephone
line connection, a fax server enables clients to fax documents directly from
their computer.
✦ Web server: A Web server delivers Internet Web pages to client computers.
The client runs a Web browser, which makes a request for an HTML web page
from the server. The server receives that request, and sends the desired page
to the client’s Web browser, where it is displayed.
✦ FTP server: The File Transfer Protocol (FTP) is a protocol for transferring
files. It facilitates these transfers with proper integrity, efficiency, and security.
A client computer connects to an FTP server, where they can supply logon
credentials, and then be granted access to retrieve files on the server.
✦ Proxy server: A proxy server forwards network requests on behalf of another
client or server. A proxy server is typically configured to facilitate Internet
Web page requests between a client and a Web server. The proxy server can
also cache Web data, so that a request for an external Web page can often be
taken from the local cache, greatly increasing performance. If the data does
not exist locally, the proxy server will go out on the Internet and fulfill the
request. Proxy servers are also useful for filtering requests, so that adminis-
trators can restrict users from viewing unauthorized Web sites.
✦ Firewall server: A firewall acts as a security filter to separate and protect a
private network from users of a public network. For example, a company will
restrict external users from the Internet from accessing the company’s
intranet. All network traffic flows through the firewall, and access is granted
or denied depending on the filters and rules that are set up by the administra-
tor. Although a firewall can be a dedicated network device, a server can be
made into a firewall by installing two or three network interface cards (NICs)
and special firewall software. Firewall security is usually set up in three zones:
an internal network, an external network, and a Demilitarized Zone (DMZ),
which allows users from the Internet to access certain services on the internal
network without risking unauthorized access to sensitive internal services. If
you are going to use a DMZ network, you will need a third network interface
card (NIC) installed on the server. Figure 1-1 shows an example of typical
4809-3 ch01.F 5/15/01 9:45 AM Page 8

firewall server setup. The mail server and the Web server are in the DMZ,
which means they can be accessed from the Internet, but the rest of the work-
stations are safe behind the firewall.

Workstation

Internet

Firewall Server

Hub Workstation

Local
LAN

Workstation
Hub

Web Server Web Server

DMZ
Figure 1-1: A Firewall server separating the Internet from the local LAN
and DMZ Internet services

✦ DHCP server: A DHCP server runs the Dynamic Host Configuration Protocol,
which assigns network IP addresses to clients on the network at start-up. With
this configuration, each client workstation does not need to be set up with a
static IP address. When the client computer starts, it sends a request to the
DHCP server, which assigns the client an IP address.
✦ DNS server: A domain name service (DNS) server keeps a database of tables
that translate fully qualified Internet domain names to their respective IP
address. This enables you to refer to Internet servers by name, such as
www.hungryminds.com, rather than by an IP address, such as 38.170.216.15.
4809-3 ch01.F 5/15/01 9:45 AM Page 9

✦ WINS server: A WINS server is a special Windows NT server that runs the
Windows Internet Naming Service (WINS). It enables clients to resolve
Windows NetBIOS names to standard Internet domain naming conventions.
For example, a computer with the NetBIOS name of Joe, would be visible to
other clients as joe.hungryminds.com.

Exam Tip For the exam, be careful not to confuse a WINS server and DNS server. The easiest
way to remember the difference is that a WINS server is for Windows clients only.

✦ RAS server: A Remote Access Server (RAS server) is a Windows NT service

that allows users to dial in to the server using a modem or WAN link. Clients
connect to the RAS server using a built-in RAS client, or point-to-point proto-
col (PPP).
✦ SNA server: System Network Architecture (SNA) is a set of protocols devel-
oped by IBM for mainframe computers. An SNA server acts as a gateway
between the client computer and the mainframe computer.
✦ Gateway server: Using a combination of hardware and software, a gateway
acts as a link between different types of networks. A gateway server must be
multihomed, which means it has more than one network card. Each card will
be connected to a different network, which are then linked internally. The
functions of a gateway server can also be performed by a dedicated network
device. See Figure 1-2 for an example of a SNA gateway providing SNA connec-
tivity to a mainframe computer. The AS/400 mainframe uses the Systems
Network Architecture (SNA) network model. With the gateway server in place,
the workstations can communicate with the mainframe no matter what oper-
ating system they’re running.

Workstation

Workstation Hub
Mainframe AS/400

SNA Gateway

Workstation
Figure 1-2: An SNA gateway server enables the network clients access to
the SNA-based AS/400.
4809-3 ch01.F 5/15/01 9:45 AM Page 10

✦ Router server: A router can connect any number of Local Area Networks
(LANs) by routing network packets to and from these different networks.
Routing can also be performed by a dedicated network device. A router con-
tains a routing table, which defines the best route for forwarding network
information to its respective destinations. A router server must be multi-
homed, meaning it must have at least two network cards, one on each net-
work. Figure 1-3 shows a server acting as a router. All packets from LAN 1 on
the left pass through the router, which forwards them to the appropriate host
on LAN 2 on the right. The router also forwards packets from LAN 2 to LAN 1.

Workstation Workstation

Workstation Hub Hub Workstation

Router Server

Workstation Workstation

LAN 1 LAN 2
Figure 1-3: A router server transfers packets between the two networks

✦ Bridge server: Much like a router, a bridge forwards packets between two dis-
similar networks. For example, a bridge can connect an Ethernet network with
a Token Ring network. Bridges are protocol-independent, and do not perform
any type of special routing functions that would be performed by a router.
Figure 1-4 shows an example of a bridge server connecting to different types
of networks.
4809-3 ch01.F 5/15/01 9:45 AM Page 11

Workstation

Workstation Workstation

Ethernet Token-ring
Workstation

Workstation Workstation Router Server

Workstation
Figure 1-4: A bridge server connects an Ethernet network with a Token
Ring Network

Verifying OS Hardware Compatibility

Objective ✦ Verify hardware compatibility with operating system

Before ordering any server equipment, you must examine the documentation for
your operating system to establish what types of hardware will and will not work
with the operating system. Each type of hardware, whether it is the motherboard,
memory, disk drives, network cards, or other peripheral cards, must be cross-
referenced with the OS hardware compatibility list. If the hardware is not supported
by the operating system, it will not work properly, and may even cause the system
to fail.

Exam Tip The most recent information on hardware compatibility will always be found on
the OS vendor’s Web site.

The hardware compatibility list for each vendor should be included with the OS
CD-ROMs, although they could be somewhat out-of-date. The OS CD will only con-
tain information at the time of the initial release of the software. To get the latest
versions of the list, go to the OS vendor’s Web site.
4809-3 ch01.F 5/15/01 9:45 AM Page 12

Verifying Power Sources and UPS Installation

Objective ✦ Verify power sources, space, UPS, and network availability

Verifying a power source is often the most overlooked aspect of a server installa-
tion. You can spend a lot of time and money on planning out a server, such as defin-
ing its role, carefully comparing and choosing proper components, hardware
compatibility with the OS, the rack size and type, air conditioning, and ventilation.
When the server is finally installed, you realize that there is nowhere to plug the
server in. Verifying the location of your power source, measuring its quality and
consistency, and protecting your equipment with an uninterruptible power supply
(UPS) are all vital and important aspects of a server installation.

Location
Before choosing a location for your server and rack, scout out the area and verify
that there are existing power outlets in close proximity to the location. Make sure
there are enough outlets to power your equipment, including monitors and other
peripheral devices such as external tape drives. You should avoid having all your
devices on the same electrical circuit. Too many devices on one circuit could affect
the consistency of the power flowing to the devices. Some power outlets are color-
coded to show which circuit an outlet is on.

Caution Avoid plugging power strips into the UPS to create more sockets. This will increase
the load on the UPS.

Quality
If your server is located in an older building, with outdated electrical subsystems, it
is possible that the power you are getting is of very low quality. High deviations in
the voltage and amperage could cause problems with modern, highly sensitive elec-
trical equipment. This can cause permanent damage to your components.

Before the servers are installed, you should measure the outlets with a power
meter that can accurately assess the quality of your power. If updating the electri-
cal system is not an option, consider the use of a line conditioner, which is a special
device that is connected between the power source and the destination that cleans
up electrical power to correct any deviances.

Grounding
The primary purpose of electrical grounding is to protect equipment from short-
circuiting. A ground refers to a path for the overloaded electrical current to dissi-
pate. Ensure that your electrical outlets in your server room are properly grounded.
You can check this by examining the outlet to see if it contains a third grounding
4809-3 ch01.F 5/15/01 9:45 AM Page 13

wire, made for a three-pronged plug. You should also have an electrician test the
power outlet. Non-existent or improper grounding can cause great damage to your
electrical equipment.

Uninterruptible power supply

Another electrical consideration for a server room is the effect of a power outage.
With no type of power backup, your servers will shut down immediately, without a
chance to gracefully shut down and save any current information to the storage
system. AUPS is a must for any type of server installation. A UPS will supply battery
power to servers in the event of a power failure. It will usually keep them up for
10–20 minutes before the battery is depleted, depending on the number of devices
it is connected to. This gives you enough time to shut down the system properly.

Most UPS units come with special software that interacts with the operating system
of the server. The UPS uses a special signaling cable that is attached to the serial
port of the server. In the event of a power failure, the UPS powers the server from
battery, and then notifies the OS to shut down the server. This allows the server to
be shut down properly without any technician being on site.
In the Once a UPS is installed, very few people actually test it. Even fewer test the auto
Real World
shutdown of the server feature. In the case of a real power failure, your current
setup might not work properly, so it is a good idea to test these capabilities before
relying on them.

UPS systems come in a variety of sizes. They are typically measured using a VA rat-
ing, which combines the total voltage multiplied by the amperage. For example, a
UPS with a 600VA rating can connect only as many devices as add up to that rating.
For example, a server that is rated at 400 VA, and its 14-inch monitor rated at 150 VA
would be fine on this UPS. Ensure that your UPS is large enough to handle the com-
bined VA rating of all your equipment. If there is too much load on the UPS, it will
not be able to provide any battery power at all during a power failure.

Verifying Rack Space

Objective ✦ Verify power sources, space, UPS, and network availability

When installing servers, you should have a special server cabinet or rack that will
hold the equipment. There are many reasons for using a server rack to house your
equipment:

✦ Safety: The equipment will not be subject to physical abuse and damage by
being left out in the open.
✦ Stability: When mounted in a proper rack, there is little chance that your
server will fall or tip over, compared to having it sit on a table or on the floor.
4809-3 ch01.F 5/15/01 9:45 AM Page 14

✦ Cooling and Ventilation: Most racks have proper ventilation slots on the
doors and sides to allow air to circulate. Some racks also have their own fans,
which help circulate the air.
✦ Security: Racks are usually lockable, so that no one can tamper with them.
✦ Manageability: By using a special keyboard/video/mouse switch, you can con-
trol all of the servers in the rack from one console, which can be installed as a
pull-out tray.

Exam Tip On the exam, questions on server racks and cabinets usually rely more on com-
mon sense for the answer than anything else. Always imagine how you would
perform the action in the real world using common sense.

Rack space is measured in Units. 1U is equal to 1.75 inches (4.4 centimeters). Each
server or device being installed into the rack will have its own Unit measurement.
For example, a medium size server has a rack height of 7U, meaning it is 12.25
inches (30.8 centimeters) tall. Some small servers are only 1U in height. A standard
rack height is approximately 42U. You must plan carefully how much rack space
you have available for your servers. Remember to include any other devices, such
as power strips, keyboard/video/monitor trays, UPS, and also leave some room for
expansion. When loading the rack, always use the manufacturer’s specifications. If
for some reason the specifications are not present, load the devices starting at the
bottom from heaviest to the lightest.

Network Cabling and Connectors

Objective ✦ Verify power sources, space, UPS, and network availability

There are many different ways in which your server can be connected to the net-
work. Several types of network topologies and cabling exist, each with their own
advantages and disadvantages.

Network topologies
Network topology refers to the actual shape of your local area network. The choice
of a topology will depend on your business needs, and the environment in which
your network will be operating.

Bus
A bus network topology is one in which all of the devices are connected together in
linear order by a central cable. A bus topology is most often used by coaxial
Ethernet networks. It is perfect for small, office networks, because it is easy and
inexpensive to set up. A disadvantage of a bus network is that if any device is dis-
connected from the network, it will affect the entire bus, and the entire network
chain will be disrupted. See Figure 1-5 for an example of bus network topology.
4809-3 ch01.F 5/15/01 9:45 AM Page 15

Workstation Workstation Workstation

Workstation Server

Workstation Workstation
Figure 1-5: A bus topology network

Star
A star network topology is one in which the devices all radiate from a central hub. If
one of the network devices fails, it will not bring down the rest of the network. The
real point of failure is with the central hub or computer. If it fails, the entire network
will be down. A star topology also uses much more network cabling than a bus or
ring topology. Star topologies are most often used by twisted-pair Ethernet net-
works, and support a wide variety of small to large networks. See Figure 1-6 for an
example of a star topology network.

Workstation

Workstation Workstation

Workstation Hub Workstation

Workstation Workstation

Workstation
Figure 1-6: A star topology network
4809-3 ch01.F 5/15/01 9:45 AM Page 16

Ring
A ring topology network is designed so that all of the network devices are con-
nected together in a ring or loop. Each device is connected to the next device in the
ring. Network information is passed along the ring until it reaches its destination
computer. A ring network is more expensive than a bus or start topology, but they
have the advantage of being able to span large distances. Physical breaks in the ring
cabling or a malfunctioning workstation can take down the entire network, and can
often be time consuming to track down. Token-ring networks use this type of topol-
ogy. See Figure 1-7 for an example of a ring topology network.

Workstation Workstation
Server

Workstation Workstation

Workstation
Figure 1-7: A ring topology network

Types of cable
Cabling your network depends on many factors, including network speed, length,
resistance to electrical interference, and cost. Each of the cable types discussed in
the following sections have their own advantages and disadvantages depending on
the type of your server and network installation.

Coaxial
Coaxial cable is one of the earliest types of network cabling. Coaxial cable consists
of a stiff, copper wire conductor, surrounded by a metallic mesh fiber, and then a
thick plastic sheath. The purpose of the plastic is to prevent the cable from being
bent so far that the copper conductor is broken. The metal fiber is used to shield
the wire from electromagnetic interference (EMI), which can interfere with network
communications. Coaxial cable is much more expensive than twisted-pair solutions,
but is cheaper than fiber optic cable. Because of its protective plastic, and EMI
shielding, it is most useful for industrial environments where the cable can be eas-
ily susceptible to damage.
4809-3 ch01.F 5/15/01 9:45 AM Page 17

There are two major types of coaxial network cabling, Thinnet, and Thicknet:

✦ Thinnet (10Base2): Thinnet coaxial cabling is a cabling standard for Ethernet

LANs. It uses RG-58 cable, with a maximum distance of 185 meters. The maxi-
mum network speed for 10Base2 is 10 MB/s. Due to its cost, Thinnet cabling is
usually used in place of Thicknet.
✦ Thicknet (10Base5): Thicknet is a larger, thicker type of coaxial cable used in
the earliest networks. It uses RG-8 or RG-11 cabling, and can run at 10 MB/s,
with a maximum distance of 500 meters.

Twisted-pair
Twisted pair wiring consists of pairs of wires that are twisted together. One wire
carries the data, while the other wire acts as a ground and absorbs electrical inter-
ference. Twisted pair is used for phone networks and lower-end networks, due to its
low cost. There are two different types of twisted pair cabling: unshielded twisted-
pair (UTP), and shielded twisted-pair (STP). STP cable uses an extra casing around
the wires to further prevent EMI from interrupting network communications.
Twisted-pair wiring is the most common network cabling used today, because of its
low cost, and the fact that it already exists in most buildings for phone connections.
There are several standards for types of twisted pair cabling:

✦ Category 1: This cabling is typically used for phone systems, not for data net-
works.
✦ Category 2: Can only support up to 1 MB/s, and should not be used for net-
works.
✦ Category 3: Can support 10 MB/s network speeds.
✦ Category 4: Primarily used for 16 MB/s token-ring networks.
✦ Category 5: Can support speeds of 100 MB/s. This is the most common type
of cabling today.

For most modern network, you will need an absolute minimum of category 5
twisted pair cabling to support 100 MB/s networks. There are a number of different
standards for twisted pair cabling:

✦ 10BaseT: 10BaseT Ethernet networks run at 10 MB/s, and have a maximum

segment length of 100 meters.
✦ 100BaseT: Also referred to as Fast Ethernet, 100BaseT runs at 100 MB/s, and
the maximum segment length is 100 meters.
✦ 1000BaseT: Also referred to as Gigabit Ethernet, 1000BaseT runs at 1 GB/s,
and the maximum segment length is 100 meters.
4809-3 ch01.F 5/15/01 9:45 AM Page 18

Fiber
Fiber-optic cable uses light to transfer data through glass fibers. A plastic covering
protects the cable. Fiber-optic cable is extremely fast and reliable, which also
makes it the most expensive of any cabling type. Companies will often build the
backbone of their network, which runs between departments and floors, with fiber
cabling, and then use twisted-pair to connect the backbone to the workstations and
servers.

Fiber cable is split into two different wires, one for transmitting and one for receiv-
ing. On the connecting device there will be two connectors, where these wires are
connected. The wires must go in the proper connectors, or the transmit and receive
feeds will be reversed, and the network will not work.

There are two types of fiber cabling, single mode and multimode. Single mode fiber
has a very narrow internal path that only permits one mode of light through at a
time. This provides much greater speeds and distances, but it is expensive com-
pared to multimode fiber. Multimode fiber has a larger core width, so modes of light
are reflected more through the pathway, causing lower speeds and higher attenua-
tion resulting in shorter cable lengths.

Connectors
Each type of network cabling requires its own type of connectors and devices to
work properly.

✦ RJ-11: This connector is used for connecting four- to six-wire connections. It

is usually referred to as two-pair wire. It is used for phone cabling.
✦ RJ-45: This connector is used to connect eight-wire connections, or four-pair
wires. It is widely used for LAN Ethernet networks.
✦ Attachment Unit Interface (AUI ): An AUI is typically used on coaxial-based
network cards. The AUI interface includes a transceiver that connects a coax-
ial cable with a 15-pin connector.
✦ BNC connector: A BNC connector is the type used on the ends of a RG-58
coaxial cable. For bus networks, a BNC T-Connector is used to connect
devices to the network, as it has three connections: one for each side of the
cable, and one to connect to the device. A BNC barrel connector is used to
connect two cables together. BNC connectors are locked into place by a rotat-
ing ring.
✦ Vampire tap: A vampire tap is used to connect older Thicknet types of net-
works. The cable is actually pierced and direct contact is made to the core
conducting wire. It is usually used in conjunction with a transceiver and an
AUI connector.

Exam Tip Although networking is not the focus of the exam, you should have good general
knowledge of the various networking topology, cabling, and connector standards
to aid you in troubleshooting questions.
4809-3 ch01.F 5/15/01 9:45 AM Page 19

Cable location
A large amount of cabling will be needed to connect your servers to the main
switch or router. Running them across the floor is a simple, but bad, alternative.
Running the cables along the floor leaves them susceptible to damage from people
walking on them, or tripping over them. They should be located either under the
floor, above the ceiling, or run along special cable trays on the sides of walls.

Many professional server rooms have false floors, with floor panels that can be eas-
ily lifted. Underneath is where most of the cabling is run, usually both power and
networking cabling. This is the best solution, as the cabling can be easily accessed
if devices need to be disconnected or moved. The cables are also covered and pro-
tected from external damage.

Server racks typically come with holes in the top, to enable cable drops to be
brought down from the ceiling into the rack, either by themselves, or through a pro-
tected conduit. The conduit is usually made of a hard but bendable plastic, so it can
be contoured to the environment. The cables are protected from external damage,
but if they need to be accessed or rerouted for any reason, accessibility is not easy.

An alternative to running cables right into the ceiling is through the use of cable
trays. These are simply brackets that run around the upper perimeter of the room,
allowing cable to be protected from external damage, but still accessible for moves
and rerouting.

Verifying Components
Objective ✦ Verify that all correct components and cables have been delivered

When your new servers are delivered, there is always the possibility that there are
parts missing, or incorrect items, that will delay the installation. Be sure to check
every component carefully, to ensure that it is the right hardware for your installa-
tion, and that it is compatible with other hardware and your network OS.

The following are some things to consider when verifying your server components
before installation.

✦ IDE and SCSI Cables: Ensure that any IDE and SCSI cables are the proper type
for your installation. For IDE systems, make sure that the cable will be able to
support all of your devices, and that the peripheral devices have the correct
jumper settings for master/slave/cable select abilities as appropriate. Check
the vendor’s documentation to ensure proper settings.
For SCSI cables, ensure that the connectors on the cables are the right width,
and have the same number of pins for your devices. There are quite an array
of different SCSI devices and cables, and they can easily be mixed up. Also
check that you have received the correct type of SCSI terminators to properly
terminate the SCSI bus.