|
The following questions are to be answered while reviewing the following text...
1. What does the acronym LAN stand for?
2. What are the advantages of having computers and other hardware connected to a network?
3. What is a diskless workstation? Why are they more secure?
4. What ls a node?
5. Describe the three ways that computers can be organized In a network's architecture. What Is one advantage and one
disadvantage of each arrangement?
6. What Is the difference between a baseband and a broadband network? Can you think of one advantage and one disadvantage
of each?
7. What Is the purpose of a communlcatlons Interface unit?
8. Describe two ways networks keep signals from Interfering with each other when they are being sent over a network.
SAVE AND ** E-MAIL YOUR "LOCAL AREA NETWORKS" REVIEW QUESTIONS FOR GRADING .......... /16
Once you feel confident that you have completed this lesson, e-mail your document as an attachment to Mr. Dunning at the
following address:
dunning_r@fc.sd36.bc.ca
Your document should include the appropriate file extension in order for the formating to remain in transmission.
For sending documents from Apppleworks, use the extension ".cwk". A correct file name and extension would look
similar to:
L2localareanetworks.cwk
Notice that any spaces in the file name have been removed and an indicator of which lesson (L2) has been added.
The assigned value for this assignment will be returned once it has been graded.
Local Area Networks
OBJECTIVES
After completing this topic, students will be able to:
~ Describe why computers are connected into networks
~ Describe typical network architectures
~ Explain how signals are sent along a network
~ Describe how traffic is controlled on a network
There is a saying that "The whole is greater than the sum of its parts." This is certainly true of microcomputers.
Individually, they are useful but when linked into a Local Area Network (LAN), they become even more so. Local area networks
can connect computers within a single building or in several nearby buildings, on a college campus for instance. The number
of computers that are linked in this fashion is growing at a phenomenal rate. There are two primary reasons for connecting
computers into local area networks.
* The computers can share expensive peripherals like plotters, scanners, laser printers, and hard disk drives and CD burners.
Where there are only a few computers, this can also be accomplished by switches that switch a printer or another device between
the computers.
* The computers can communicate with each other, and users can exchange Information and electronic mail.
* Local area networks are created by connecting microcomputers, printers, and hard disks with wires or cables. Microcomputers
connected into a network operate together as equals. Each microcomputer has its own computational ability, so it can function
either alone or as part of the network. LAN's differ from multiuser systems where dumb terminals or microcomputers that act
like dumb terminals and are connected to a central computer. In a multiuser system the dumb terminal can draw only on the
resources in the main computer: it cannot run its own programs. In a LAN the computers can be regular microcomputers or diskless
workstations, which are microcomputers without disk drives. All the programs it runs and all the data files it saves are stored
on the network's hard disk drive. Diskless workstations are more secure than computers with disk drives. For example, users
cannot illegally copy program or data files that are stored on the network's hard disk.
Several networks are available that connect microcomputers and their peripherals. In this topic, we explore the basic
concepts of local area networks.
Network Architecture
Microcomputers can be arranged to form a network in several ways. This layout is called network architecture (or topology).
All network architectures include hardware, called nodes and wires or cables over which data is sent between the nodes. A
node is a single, addressable device, for example, a microcomputer or printer. A network's architecture refers to the way
the data flows between nodes, not to the way they are physically arranged in an office.
Bus Architecture
Bus architecture connects all nodes to a single bus much as the components are organized within the computer. In a bus
arrangement a signal addressed to another node is sent from a microcomputer to the bus.
All other nodes on the network, which are also connected to the same bus, examine the signal. If the signal is addressed
to them they accept it. If it is not they ignore it. Since each node is separately connected to the network any node can break
down without affecting other nodes. This feature makes bus architecture one of the most popular arrangements for networks.
Ring Architecture
Ring Architecture arranges the nodes on the network in a circle. When one of the microcomputers on the network sends
a signal it passes it to the next node on the network. If it is not addressed to that node it is retransmitted to the next
node and so on around the circle until it reaches the node it is addressed to. Because each node on the network retransmits
any signal that is not addressed to it. If one node breaks down the entire network breaks down. Unlike the other two systems
where new equipment can be plugged into the bus or host computer a ring network requires rewiring when new equipment is added.
To insert a new node the connection between two existing nodes must be broken. The new node is then installed and wired to
the two adjacent nodes.
Star Architecture
Star Architecture has the nodes connected to a central, or host computer. When one of the microcomputers on the network
sends a signal, it is sent to the host that routes it to the node it's addressed to. There are no direct connections between
the nodes on the network except through the host computer. If the host breaks down so does the network, but if one of the
other microcomputers breaks down no harm is done. The others continue to function.
Network Servers
When printer's modems and hard disk drives are connected into a network additional hardware called network' servers are
required to allow these nodes to be shared by all the microcomputers on the network. There are four basic kinds of servers.
1. Print Servers allow computers on the network to use the same printer.
2. File Servers allow all computers on the network to use the same hard disk drive.
3. Routers connect two or more networks with the same architecture. Generally, the networks must be supplied by the same
vendor or be otherwise compatible. For example, a routing server can connect an IBM LAN to another IBM LAN if the IBM LANs
are compatible.
4. Gateway Servers connect two or more different types of networks. These gateways simplify the network and reduce network
expenses. For example if each computer in a network had direct access to a mainframe each would require a special add-on board.
But when a gateway server is used only the computer that acts as the gateway requires the add-on board. All the other computers
in the network access the mainframe through that board.
Network Connections and Signals
When signals are sent through a network they are sent between nodes along a common set of wires. Microcomputers and their
peripherals that are organized in a network must be connected by hardware that includes both the add-on cards that plug into
each node s expansion slots and the cables that connect them. Three types of cables are used to connect the computers in the
network: twisted pair, coaxial, and fibre optic
~ Twisted Pair wires are just like telephone wires. They are the least expensive way to connect computers because the
wire is cheap and easy to install. Their disadvantages are a relatively slow speed and a tendency to pick up electrical interference
that can cause high error rates.
~ Coaxial Cable is layered. An inner wire is surrounded by an insulating material that is in turn, surrounded by a braided
wire. This braided wire shields the inner wire from any noise or other signals in the environment that can affect the quality
of the transmission. Coaxial cables are reliable and have a wide bandwidth so they can transmit data much faster than twisted
pair wires
~ Fibre Optic cables are made of plastic or glass fibers covered with an opaque sheath that keeps light from entering
or escaping. The digital data signals from a microcomputer, or other node are used to modulate a light beam to convert it
into pulses. These light pulses are sent along the optical fibers and are unmodulated at the receiving end. Fibre optics is
fast and reliable, as well as small and light. They also have an extremely wide bandwidth so a large number of signals can
be sent simultaneously.
Since several nodes may be sending data at the same time the signals must be kept separated on the wires. This can be
done by sending them at different times or different frequencies.
Baseband
When the signals are sent at different times the network is called baseband. The baseband network merges the signals
so that those from the two computers are sent along the network one at a time separated from each other by time. Baseband
networks operate at medium speed. Because they use unmodulated signals modems are unnecessary. Baseband networks can also
transmit voice messages in digitized form that can be stored and played back when desired.
Broadband
When the signals are sent at different frequencies the network is called broadband. Unlike baseband networks where only
one signal is sent at a time a broadband network can simultaneously carry many signals. To send data at different frequencies,
a number of carrier frequencies are used to electronically divide the cable into a number of channels. To do this modems are
required at all nodes to convert digital signals into analog signals and then back again. A major advantage of a broadband
net work is that it can carry audio, video and digital data at the same time. Each device on the network uses its own channel
to transmit and receive data.
Communication Interface Units
All the computers, printers, and other hardware are connected to the network though a Communication Interface Unit (CIU).
The type of CIU required depends on what signal is used in the network. In a broadband network where the sending node modulates
the signal and the receiving node demodulates it, the CIU is a modem. On a baseband network where unmodulated digital signals
are sent along the network the CIU is a transceiver (a combination transmitter and receiver). In either case, the CIU performs
several functions.
~ Monitors the traffic on the network to see when the hardware connected to it can send a message.
~ Checks traffic on the network to see if a message is addressed to its node and, if so accepts it.
~ Retransmits or repeats message if the message is not addressed to its node. This is done only in a ring network since
bus and star networks do not require signals to be retransmitted from node to node.
~ Converts signals from the network into a form acceptable to its node and vice versa. For example on a broadband network,
It converts digital signals into analog signals when sending data and converts analog signals into digital signals when receiving
data. If necessary, it converts parallel signals into serial signals and vice versa, for example, when its node is a serial
printer.
~ Detects errors in messages addressed to its node and if found, requests that the sending node retransmit the data.
Network Traffic Control
The nodes connected to a network are each assigned a unique address, known as an IP address. When a node wants to communicate
with another node-(either computers or peripherals), it adds its address to the message it is sending. Since several computers
may be sending messages two or more may try to send signals at the same time. If this happens the messages will not get through.
Two common methods of solving this problem are collision detection and token passing.
Collision Detection
In a network using Collision Detection none of the computers nodes are aware of the others. Each node transmits data whenever
it wishes. If no other computer is sending data the message is received by the device it was addressed to, and acknowledgment
is sent back to the sending node. But if two computers transmit at the same time their messages collide and neither gets through.
Since no acknowledgment is sent to the sending computers both computers have to retransmit. To prevent the messages from colliding
on the next try, the computers randomly generate a pause before another attempt is made. Since both computers randomly time
their pauses they most likely will make their next attempts at different times. Because the number of collisions increases
as the number of computers using the line increases a heavily used network that uses collision detection may operate slowly.
Token Passing
Token Passing is a more sophisticated way to time signals so that they do not collide with each other. A single encoded
signal called the token is sent around the network. Any node can let it pass or choose to grab it. The node holding the token
is the only one allowed to transmit data on the network. An analogy is a group of people passing a tape recorder around a
conference table. Any person can choose to speak into it or pass it to the next person. In the network, a node holding the
token sends data addressed to another node. All the other nodes look at the message as it goes around the network to see if
the message is addressed to it. If the message is not addressed to it, the node passes it along to the next node on the network.
When the node it is addressed to has received the message, the receiving node sends back an acknowledgment that the data was
received. The token is then put back into circulation. Since there is only one token in circulation, no collisions can occur.
Because no single computer can use it twice in a row, other computers have fair access to it.
Network Software
Networks are managed by network management programs that route messages and allow computers to be connected and disconnected.
These programs should be as transparent or invisible to the user as possible. When fully transparent the user does not even
know they are connected to a network. Each computer operates just as it does when disconnected from the network. Unfortunately
this goal has not yet be achieved, and network management software adds another level of knowledge required by the user. Eventually
this problem will be solved and computers on a network will operate as simply as telephones which though connected to an extremely
complicated and immense network operate as very simple devices.
Moreover operating systems and applications programs used on a network must be modified. For example most programs are
designed for standalone computers. A user on a standalone can use these programs to retrieve or save files as he or she wishes.
But when these programs are used on a network these operations must be regulated because two or more users may want to work
on the same file at the same time. To prevent this from happening, applications programs, or the operating system, must be
able to lock files that are currently being used so that other users cannot retrieve and modify them. When the first user
saves or quits the file. It is then automatically unlocked so that another user can retrieve it. Many of the security problems
discussed on multiuser database management programs also apply to networks
|
