CS 55 - Local Area Networks

7/31/2019 CS 55 - Local Area Networks

1/181

1-1

Local Area Networks

CS 55 Computer Networks

Topic 5


2/181

1-2

Loc al Area Networks (LANs)(Study Map)

Introduction Review of Data Communications Concepts

Review of Basic LAN Concepts

Hardware LAN Media

Topologies and Media Access Control

LAN Hardware

Software LAN System Software

Client/Server Architecture


3/181

1-3

The EthernetLocal Area Network

I came to work one day at MIT and thecomputer had been stolen, so I called DEC

to break the news to them that this$30,000 computer that they'd lent me wasgone. They thought this was the greatestthing that ever happened, because it turns

out that I had in my possession the firstcomputer small enough to be stolen!-Robert Metcalfe on the trials and tribulations of inventing

the Ethernet.


4/181


5/181

1-5


Metcalfe had two challenges: thenetwork had to be fast enough todrive the very fast new laser printer;

and it had to connect hundreds ofcomputers within the same building.Never before had hundreds ofcomputers been in the same building

-- at that time no one had more thanone, two or maybe three computersin operation on any one premise.


6/181

1-6


The press has often stated that ethernetwas invented on May 22, 1973, whenMetcalfe wrote a memo to his bosses

stating the possibilities of ethernet'spotential, but Metcalfe claims ethernet wasactually invented very gradually over aperiod of several years. In 1976, Robert

Metcalfe and David Boggs (Metcalfe'sassistant) published a paper titled,"Ethernet: Distributed Packet-SwitchingFor Local Computer Networks."


7/181

1-7


Metcalfe left Xerox in 1979 to promote theuse of personal computers and local area

networks (LANs). He successfullyconvinced Digital Equipment, Intel, andXerox Corporations to work together topromote ethernet as a standard. Now an

international computer industry standard,ethernet is the most widely installed LANprotocol.


8/181

1-8

End of the Story

Not yet we have a

compressed lecture classtoday hmmm ...


9/1811-9

Review of Data

CommunicationsConcepts


10/1811-10

Section Preview

In this section we will study:

The requirements for communication

What constitutes a network The various types of networks in

common use today

The OSI Reference Model Some of the basic terminology of

data communications and networks


11/181


12/1811-12

Understanding the Message

In computer systems, data can berepresented by any of several

different codes, the two mostcommon being:

the American Standard Code for Information

Interchange (ASCII) the Extended Binary Coded Decimal Interchange

Code (EBCDIC)


13/1811-13

Security

Sensitive data like your credit cardnumber or other secret data should be

safeguarded during transmission. Themost common mechanism for protectingdata during transmission is encryption. Encryption transforms plain text into an (presumably)

undecipherable form called cipher-text.


14/181


15/1811-15

Three Network Subtypes

Local Area Network (LAN)

Metropolitan Area Network (MAN)

Wide Area Network (WAN)


16/1811-16

LAN/MAN/WAN Comparison

Limitedtypically upto 2,500 meters or 2miles

Hightypically inexcess of 10 Mbps10,100 and 1,000 arestandard

Locally ownedtwisted-pair wires, fiberoptic cable, wireless(not satellite)

Can be any, but mostare desktop computers

Limitedtypically upto 200 kilometers or100 miles

Hightypically100 Mbps

Locally owned andcommon carriertwisted-pair wires, fiberoptic cable

Can be any, but mostare desktop computersand minicomputers

Unlimited

Slowerusually 1.5Mbps

Locally owned andcommon carriertwisted-pair wires,coaxial cable, fiberoptic cable, wireless toinclude satellite

Can be any, but mostare desktop computers

Distance

Speed

Media

Nodes

LAN MAN WAN


17/1811-17

The OSI Reference Model

The problem of network interconnection isso important that the ISO created the OSIReference Model that describes the

functions a generic network needs toprovide.

The OSI Reference Model has become the

basis for many data communicationsstandards. Because these standards are placed in the public domain, they are

called open standards and lead to open systems.


18/181


19/181


20/1811-20

The OSI Model at Work

Application Layer The application on Node A builds a record with a transaction

identifier, the number of the account to be updated, the date and

time of the transaction, and the amount to be deducted.

Presentation Layer The presentation layer is responsible for translating from one format

to another.

Session Layer The session layers major functions are to set up and perhaps

monitor a set of dialogue rules by which the two applications can

communicate and to bring a session to an orderly conclusion.


21/1811-21

The OSI Model at Work (cont.)

Transport Layer The transport layer is the first of the OSI layers responsible for

actually transmitting the data.

Network Layer The network layer provides accounting and routing functions.

Data Link Layer The data link layer is responsible for data delineation, error

detection,and logical control of the link.

Physical Layer The physical layer does not append anything to the message. It

simply accepts the message from the data link layer and translates

the bits into signals on the medium.


22/1811-22

Receiving the Message

1. The message is passed over the link connecting Nodes A

and X.

2. The message is passed to the data link layer in Node X. Themessage is checked for transmission errors, the PDU

information applied by As data link layer is removed, andthe message is sent to Xs network layer.

3. Xs network layer records the accounting information for

the message and then strips off the network layer protocoldata and examines the destination address. The destination

is not Node X in this case, so the network layer consults itsnetwork routing table and determines the next link on thepath to Node M. Xs network layer affixes the proper

network layer protocol data and sends the message toNode Xs data link layer.


23/1811-23

Receiving the Message (cont.)

4. Node Xs data link layer creates its PDU and sends the

message to Node M.

5. Node Ms data link layer receives the message, strips off

Node Xs data link layer protocol data, checks for

transmission errors, and passes the data up to Node Ms

network layer.

6. Node Ms network layer gathers accounting data, strips off

the network layer protocol data, and fins that the messageis destined for an application in this node.

7. The message is passed up to Ms transport layer, where the

sequence number is checked to ensure that no messageshave been lost. The transport layer protocol data isremoved.


24/1811-24

Receiving the Message (cont.)

8. The message arrives at the session layer, whererelevant protocol data is examined and remove.

9. The message arrives at Node Ms presentation

layer, where appropriate action is taken.

10.The message arrives at the application, where

it is acted on.


25/1811-25

General Network Implementations:LANs

A LAN services a limited geographic area at high speedsusually 10 million bits per second or higher. Allcomponents of the LAN are commonly owned by theorganization that uses it.

The nodes in many of todays LANs are desktop systemslike personal computers. Henceforth, we will also use theterms workstations, clients, and serversin referring to LANnodes.

A workstationis used here to represent a LAN users

computer; other terms used in referring to a workstation areclient and node.

A server is a network node that is dedicated to providing

services to client nodes.


26/181


27/181


28/181

1-28

General Network Implementations:VANs

A VAN is a network owned by a communicationsutility that sells the services of the network toother companies.

A communications utility that owns a VANprovides connectivity to multiple locations.

The value added by the communications utility isthe maintenance and management of the

communications circuits.


29/181

1-29

General Network Implementations:Enterprise Networks

An enterprise network is an organizations

complete network.

With the advent of LANs, many companiesinstalled departmental LANs to improve theproductivity of work groups.

Soon, these companies realized that there was abenefit to having users on one LAN communicate

with users or applications on other LANs or onthe WAN, and the various networks wereconnected together to form one corporate-widenetwork, the enterprise network.


30/181

1-30

General Network Implementations:The Internet

An internet (with a lowercase i) is the

interconnection of two or more networks. Anenterprise networks just described is an example

of an internet.

The Internet (with an uppercase I) is a specific

instance of an internet.

The Internet is a global network of networks. The

Internet is made up of hundreds of networks,thousands of nodes, and millions of usersthroughout most countries of the world.


31/181

1-31

General Network Implementations:Intranets

An intranet is an organizations private Web.

Companies have found that WWW capabilitiescan improve the information flow and availabilityin a company.

Companies may use an intranet rather thanpublishing on the Internet because theinformation being provided is intended for

corporate use only and not for the public at large.


32/181

1-32

Review of Basic Local

Area Networks Concepts


33/181


34/181

1-34

Why Use LANs?

Resource Sharing The need to share resources is still one of the primary reasons for

using a network.

Dedicating resources (disk drives, tape drives, printers, data,

application programs, scanners, etc.) to each user is expensive. In anetwork, some or all of these resources may be shared.

Communication We use data communications networks for communications, but the

objects that are communicating are not always people. A person-to-person communication may be an electronic

conversation with the two parties exchanging messages in real time.

An application-to-application dialogue may be the transfer of a file

from one node to another.


35/181

1-35

Why Use LANs? (cont.)

Management Control A LAN can help standardize the microcomputer

environment.

Application standards can be set up more easily in anetwork because most application programsfor

example, word processors and spreadsheet programs

can be installed on one or more network nodes called

servers.

LANs can also help control computer viruses. A diskless

workstation, as the name implies, has no local disk

drives. This reduces the ways in which a virus can be

introduced.


36/181

1-36


Cost Effectiveness The ability to share resources has a direct impact on an

organizations expenses. If users can share hardware,

less hardware is needed.

The less obvious cost effectiveness comes from the

ability of users to communicate and thus improve their

productivity.

Downsizing In some companies, LANs have been used to downsize

the data processing hardware, software, and personnel

requirements.


37/181

1-37


New Application Software One of the newer application technologies associated

with networks is called groupwareapplications orientedtoward improving the productivity of a group of people

working together.

Groupware automates work functions that require the

communication and collaboration of two or more

individuals. These tools raise office productivity and

make LANs even more cost effective.


38/181


39/181

1-39

Application Software: PersonalProductivity Applications (cont.)

Database Management A database management system (DBMS) gives you the ability to

define, use, and manage data. With LA-oriented databases, users share the database. This reduces

the amount of redundant data storage that occurs when individuals

maintain their own databases on stand-alone systems.

Presentation Services Graphics hardware and software can be rather expensive. A LAN

allows these expensive resources to be shared, providing a cost-

effective way of distributing graphics capabilities.

Applications that allow users to make more effective use of theirmicrocomputers on a LAN include:


40/181

1-40

Application Software: WorkgroupApplications

Electronic Mail Todays e-mail systems allow correspondents to exchange

communications containing text, graphics, and voice images in

batch or real-time mode. For many companies, e-mail has become aprimary mode of communications.

Electronic Appointment Calendars One user can consult other users appointment calendars to find a

time at which each user is available for a meeting.

Electronic Filing Cabinets E-mail and other machine-readable documents can be stored in disk

folders that are equivalent to file folders in conventional filing

cabinets.

Applications that allow users to make more effective use of theirmicrocomputers on a LAN include:


41/181


42/181


43/181

1-43

LAN Characteristics

Transparent Use

Mixed Hardware and Software

Limited Geographical Area

High Speed

Resource Sharing

LAN Media Communication Protocols

Local Ownership


44/181

1-44

LAN Selection Criteria

Cost For most of us, cost is an overriding constraint, an you must choose

the best solution within your budget. Usually, cost is the most

inflexible constraint under which you must operate, and in the final

analysis the LAN must be a cost-effective solution to your problem.

Number of Workstations Each LAN is physically capable of supporting some maximum

number of workstations. If you exceed that maximum number, you

must make some provision for extending the maximum number.

Type of Workstations The LAN hardware and software must be compatible with the

workstation used and their operating systems.


45/181


46/181

1-46

LAN Selection Criteria (cont.)

Distance and Medium Attaining high speed over long distances can be very expensive.

Thus, each LAN has a maximum distance it can cover.

Speed It is important that you select a LAN capable of meeting your

performance goals. Common LAN speeds available for

microcomputers are 4, 10, 16, 100, and 1,000 Mbps, and the trend is

for increasing speeds.

Applications If an application uses an interface not supported by a particular LAN,

then the application probably will not work on that network.

Moreover some software is not LAN compatible.


47/181


48/181


49/181

1-49

LAN Selection Criteria (cont.)

Security

Data in a LAN is shared. Sharing should not, however,imply that all users should have unlimited access to all

data. The LAN software must have the ability to control

access to data.

I di t d R i t f


50/181

1-50

Immediate and Recurring costs ofa LAN

Equipment upgrades

Documentation

Installation of cabling

System software installation

Creating user environments

Space required for newequipment

LAN managementpersonnel

costsConsumable suppliestoner,paper, etc.

Immediate Costs

Recurring Costs

Training users, operators, administrators

Site preparation

Hardware installation

Installing applications

Testing

Supplies and spares

Hardware and software maintenance

Training new users, administrators


51/181


52/181

1-52

Hardware and LAN

Media

Section


53/181

1-53

SectionPreview

What makes up a LAN system

Several of the Leading LAN media

Characteristics of LAN media

Error sources, detection, andcorrection



54/181

1-54

Three Major LAN Components

LAN software

Topology

The media access control (MAC)protocol


55/181


56/181

1-56

Two Major Classes of LAN Media

Conducted Media uses a conductor like a wire or a fiber optic cable to

move the signal from sender to receiver

includes twisted-pair wires, coaxial cables, and fiber

optic cables

Wireless Media

uses radio waves of different frequencies or infrared lightbroadcast through space

does not need a wire or cable conductor to transmit

signals


57/181


58/181

1-58

Conducted Media (cont.)

UTP and STP Shielded twisted-pair (STP)

These wires have a metal foil or wire mesh wrapped around

individual wire pairs with a metal braided shield around the

twisted-pair wire bundle itself.

Twisting pairs of wires helps eliminate interference fromneighboring wires; the metal shielding helps prevent ambient

distortion like heavy-duty motors, electrical or magnetic fields,

and fluorescent lights.

Unshielded twisted-pair (UTP)

These wires have no protective metal covering. UTP wires are

more susceptible to environmental noise that can disrupt the

signal.

UTP is used because it is cheaper than STP, and it may safely

be used in environments where external disruptions are rare.

Twisted-Pair Wire Category


59/181

1-59

Twisted-Pair Wire CategorySummary

1

2

3

4

5

Maximum Data RateCategory

1 Mbps

4 Mbps

10 Mbps

16 Mbps

100, 155, and1,000 Mbps

Telephones

Token Ring LANs

Ethernet LANs

Token ring LANs

Ethernet, fast ethernet,and gigabit ethernetLANs, CDDI LANs andasynchronous transfermode (ATM)

Typical UseCost (Relative toCategory 1)

1

1.5

2

3

4


60/181

1-60

Coaxial Cable

Most early microcomputer-based LANimplementations used coaxial cable as themedium.

Coaxial cable comes packaged in a variety ofways, but essentially it consists of one or twocentral data transmission wires surrounded by aninsulating layer, a shielding layer, and an outer

jacket. Coaxial cable is most commonly used in two

types of LANs, ethernet and ARCNET.


61/181

1-61

A Single Conductor Coaxial Cable

OuterInsulation Mesh Shielding Insulation Conductor


62/181

1-62

Fiber Optic Cable

Fiber optic cables come in two varieties,multimode and singlemode, each with a differentway of guiding the light pulses from source to

destination. Fiber optic links for very short distances cost

more than wires, but as distance or the requiredtransmission rate increases, fiber optic cables

become cost effective. Fiber optic cables will not corrode, so they can be

used in environments unsuited for copper media.


63/181

1-63

Views of a Fiber Optic Cable

PlasticCovering

GlassCladding

GlassConductor


64/181


65/181

1-65

Wireless Media (cont.)

Infrared Transmission Infrared transmission is a line-of-sight technology. It can

be used to provide LAN connections between buildings

and also is the medium used in some wireless local area

networks.

The Frequencies of Various


66/181

1-66

The Frequencies of VariousWireless Media

1016

1015

1014

1013

1012

10111010

109

108

107

10

6

105

104

103

102

101

X rays, gamma raysUltraviolet lightVisible lightInfrared light

Millimeter waves

MicrowavesUHF televisionVHF televisionVHF TV (high band)FM radioVHF TV (low band)Short-wave radio

AM radio

Very low frequency

Frequency (Hz)


67/181

1-67

LAN Media Selection Criteria

Cost

Speed or

Capacity

Availability

Expandability Error Rates

Security

Distance

Environment

Application

Maintenance


68/181


69/181

1-69

Media Selection Criteria (cont.)

Error Rates The propensity for error influences not only the quality of

the transmission but also its speed.

Security Although most of the hacker incidents reported relate to

wide area networks, similar concerns occur on LANs.

Distance

Before deploying a medium, LAN designers need todetermine the distances that need to be covered andensure that the wiring configuration or wirelessconfiguration does not exceed the distance limitations ofthe technology being used.


70/181

1-70

Media Selection Criteria (cont.)

Environment The constraints of environment can eliminate certain

types of media.

Application In some applications, the characteristics of the required

equipment may dictate the type of medium and interfaces

to be used.

Characteristics of Common LAN


71/181

1-71

Characteristics of Common LANMedia

Unshielded twisted-pair

Shielded twisted-pair

Coaxial Cable

Fiber optic cable

Broadcast radio

Spread spectrum radio

Microwave radio

Infrared light

Common Speeds(Mbps)

MediumType

1, 4, 10, 16, 100, 1000

1, 4, 10, 16, 100, 1000

10, 16, 50

10, 16, 50, 100, 1000, 2000

2

2, 10, 16

5.7

4, 10, 16

Less capable than other conducted media

Better than unshielded; less capable than fiberoptic or coaxial cables

Good; less capable than fiber optic cable

Excellent

Subject to interference

Good

Subject to interference

Objects can block transmission

Error Characteristics


72/181

1-72

Error Sources

White Noise White noise, also referred to as thermal noise and Gaussian noise,

result from the normal movements of electrons and is present in al

transmission media at temperatures above absolute zero.

Impulse Noise In LANs, it can be caused by lightning striking the medium, by jarring

loose connections, or by transient electrical impulses such as those

occurring on a shop floor.

Crosstalk Crosstalk occurs when signals from one channel distort or interfere

with the signals of a different channel.


73/181

1-73

Error Sources (cont.)

Echo Echo is essentially the reflection or reversal of the signal

being transmitted.

Attenuation Attenuation is the weakening of a signal as a result of

distance and characteristics of the medium.


74/181

1-74

Error Detection

Parity Check A parity check (also known as vertical redundancy check [VRC])

involves adding a bitknown as the parity bitto each character

during transmission.

Longitudinal Redundancy Check (LRC) With LRC, an additional, redundant character called the block check

character (BCC) is appended to a block of transmitted characters,

typically at the end of the block.

Cyclic Redundancy Check (CRC) A CRC can detect bit errors better than either VRC or LRC or both.

The transmitting station generates the CRC and transmits it with the

data.


75/181


76/181


77/181

1-77

Topologies and MediaAccess Control

S i P i


78/181

1-78

Section Preview

LAN topologies

Media access control protocols

Common ways in which topologiesand media access control protocolsare combined

LAN standards Strengths and weaknesses of

different LAN configurations


LAN T l i


79/181

1-79

LAN Topologies

The term topologyderives from amathematics field that deals with

points and surfaces in spacethatis, with the layout of objects inspace. The LAN topology is thephysical layout of the network.

LANs have three basic topologies:ring, bus, and star.

Ri T l


80/181

1-80

Ring Topology

Data Flow

B T l


81/181

1-81

Bus Topology

St T l


82/181

1-82

Star Topology

Wiring Hub

D t Li k P t l


83/181

1-83

Data Link Protocols

In general, a data link protocol establishesthe rules for gaining access to the mediumand exchanging messages. To do this, theprotocol describes several aspects of themessage-exchange process. Six of themost important are: Media Access

Delineation of Data

Error Control

Addressing

Transparency

Code Independence

D t Li k P t l ( t )


84/181

1-84

Data Link Protocols (cont.)

Media Access Media access defines how a node gains the right to

transmit data on the medium.

Delineation of Data A data link protocol must define or delineate where the

data portion of the transmitted message begins and

ends. This can be accomplished in two basic ways: byframing the data with certain control characters or by

using a standard message format wherein data is

identified by its position within the message.

D t Li k P t l ( t )


85/181

1-85

Data Link Protocols (cont.)

Error Control Error control is used to determine whether data has been

corrupted during the transmission.

Addressing Communication between two network nodes is

accomplished through an addressing scheme. Networks

use a hierarchical addressing scheme, with the hierarchybeing application, network node, and network.


86/181

MAC P t l


87/181

1-87

MAC Protocols

LAN technology adheres to two primaryMAC protocols: token passing andcontention.

Contention In a pure contention MAC protocol, each network node has equal

access to the medium. Variations of this protocol exist, some of

which allow for node priorities.

Token Passing Token passing is used on both bus and ring topologies. Token

passing is a round-robin protocol in which each node gets an equal

opportunity to transmit. With token passing, the right to transmit is

granted by a token that is passed from one node to another.

LAN Standards


88/181

1-88

LAN Standards

Standards exist covering most aspects ofLAN technology

Implementing according to establishedstandards generally results in theavailability of components from multiplevendors, competition among vendors, andlower prices.

The organizations most active in settingstandards for LAN topologies and MACprotocols are the IEEE and ANSI.

IEEE Standards


89/181

1-89

IEEE Standards

The 802 Committee established by the IEEE is divided intosubcommittees, each of which addresses specific LANissues and architecture. High-Level Interface

Logical Link Control

CSMA/CD

Token Bus

Token Ring

MANs

Broadband Technical Advisory Group

Fiber Optic Technical Advisory Group

Integrated Data and Voice Networks

LAN Security Wireless LANs

Demand Priority Access Method

Data Transport over Cable TV

Short-Distance Wireless Networks

Broadband Wireless Access

LLC and MAC Sublayers of the


90/181

1-90

yOSI Data Link Layer

Presentation Layer

Application Layer

Session Layer

Transport Layer

Network Layer

Data Link Layer

Physical Layer

Logical Link Control

CommunicationsInterface Unit (CIU)

Bus Interface Unit

Media Access Control

Media Signaling

OSI Reference ModelLayers

ANSI Distributed Queue Dual Bus


91/181

1-91

LAN

Bus A, Unidirectional

Bus B, Unidirectional

Suggested Broadband Frequency


92/181

1-92

gg q yAllocations

280

240

200

160

120

80

40

0

Frequency

(Nh

z)

Video Transmission Channel

High-Speed Data Channel

Voice or Data Channel

Low-Speed Data Channel

Comparison of Token-Passing and


93/181

1-93

p gCSMA/CD Media Access Control Protocols

Equal access for all nodes.

Predictable access window.

Maximum wait time to transmit is tokencirculation time.

Average wait time to transmit ispredictable (half the maximumcirculation time).

Network congestion does notadversely affect network efficiency.

A node needs to wait for the tokenbefore being able to transmit.

One node cannot monopolize thenetwork.

Large rings can cause long delaysbefore a node obtains a token.

Consistent performance for large, busynetworks.

Equal access for all nodes.

Access window can be unpredictable.

Maximum wait time to transmit isunpredictable and depends on collisions

Average wit time to transmit isunpredictable.

Network congestion may cause collisionsand reduce network efficiency.

A node may be able to transmitimmediately.

One node may be able to monopolize thenetwork.

A node can transmit when the network isquiet.

Unpredictable performance for large, busynetworks due to possibility of collisions.

Token Passing CSMA/CD

Topology and Protocol Trade Offs


94/181

1-94

Topology and Protocol Trade-Offs

There are two primary LAN implementations inuse, CSMA/CD buses (ethernet) and token-passing rings.

Ethernet has the major market share between thetwo.

Ethernet provides fair access for all nodes, but heexact time a node waits before sending a

message may vary. Ethernet provides higher speeds and lower costs

than token rings.

Broadband and Baseband


95/181

1-95

Technologies

Broadband transmission andbaseband transmission are different

ways in which you can use amedium. Broadband transmission divides the medium into several

channels, thus allowing the medium to be used for

distinct transmission needs. Baseband transmission dedicates the entire data-

carrying capacity of the medium to the LAN.


96/181

1-96

LAN Hardware

Section Preview


97/181

1-97

Section Preview

Servers

Backup devices

Workstation hardware

LAN adapters

Printers

Miscellaneous hardware


Server Platforms


98/181

1-98

Server Platforms

File Services File Servers

A file server allows users to share files. Only one

copy of application software like word processing,

spreadsheet, or database must reside on a file server.Individual clients share these applications.

Database Servers

The database server was developed to solve

problems like passing an entire file over themedium. The most common example of a database

server is the SQL server. Structured query language

(SQL) is a standard database definition, access, and

update language for relational databases.

Server Platforms (cont )


99/181

1-99

Server Platforms (cont.)

Print Services When a client shares a printer on the LAN, it can submit

jobs to the printer at any time. The print job is collected

on disk, and printing is deferred until the printer isavailable for printing a new job and the print job is the

next one eligible to be printed.

Print Server Technology


100/181

1-100

Print Server Technology

Application

ClientSpooler

Router

Client

ServerSpooler

PrintMonitor

Print Server

Server Selection Considerations


101/181

1-101

Server Selection Considerations

Server Disk Drives When choosing a server, you should carefully

select the servers disk subsystem. There are

five factors that are critical when choosing aservers disk subsystem:

Storage Capacity

Average Access Time

Separation of Function

Fault Tolerance

Drive Interface

Server Selection Considerations( t )


102/181

1-102

(cont.)

Server Memory In Intel-based servers, there are two classes of

memoryLevel 1/2 cache and standard random access

memory (RAM)

Processors and Processor Speed The processing power of the server is also a critical

factor. It seldom makes sense to select a server that hasfast disks and plenty of memory but a slow CPU. In

general, the server ought to be one of the fastestif not

the fastestcomputers on the network.

Server Selection Considerations( t )


103/181

1-103

(cont.)

Processor Bus A variety of buses are available. The bus is the path over which data

is transferred between the CPU, memory, and peripherals.

Expansion and Power A server should have sufficient expansion capability and the power

to use the expansion slots effectively. Network server capacity can

be expanded by adding hardware to the existing server or by adding

additional servers.

Compatibility The server platform you choose must be compatible with the

topology, MAC protocol, applications, and system software on the

network.

IBM-Compatible System Buses


104/181

1-104

IBM-Compatible System Buses

Industry standardarchitecture (ISA)

Enhanced industrystandard architecture(EISA)

Micro channel

architecture (MCA)VESA local (VL-Bus)

Video electronicsstandards association(VESA)

Peripheral componentinterconnect (PCI)

PC-Card (formerly

PCMCIA)

Backup Devices


105/181

1-105

Backup Devices

Floppy Diskette Drives The major disadvantages of this backup method is the low capacity

and speed of the backup media.

Removable Disk Drives Manual intervention is necessary for changing disk cartridges,

whereas some tape backup system provide tapes with much higher

storage capacity and with automatic tape changing.

Hard-Disk Drives The arguments for and against this alternative are much the same as

those for diskettes. The major difference is that the capacity of hard-

disk drives is greater than that of diskettes.

Backup Devices (cont )


106/181

1-106

Backup Devices (cont.)

Optical Disk Drives Optical disk drives are gaining popularity as input,

output, and backup devices. The reasons for this are

their decreasing costs and large storage capacity.

Magnetic Tape Drives A magnetic tape drive is the usual choice for a backup

device. Magnetic tapes are less expensive than the otheroptions. They hold large volumes of data, are easy to use

and store, and generally provide good performance.

Primary Backup Technologies


107/181

1-107

Primary Backup Technologies

Diskette backup

Hard drive, fixed

Hard drive, removable cartridge

Tape backup, 4mm or 1/4 inch

Tape backup, 8mm or VCR

Tape backup, 9-track

Optical drives

Digital versatile disks (whenavailable)

1.44 MB

Multiple capacities

40 MB to over 1 G

To 15 GB

160 MB

2.2 GB

To 2.2 GB

To 100 MB

To 4 GB

10-14 GB

2.88 MB

60 MB

500 MB

15 GB

20 MB

150 MB

1.2 GB

70 GB(compressed)

Backup Functions


108/181

1-108

Backup Functions

Back up all files

Differential backup

Back up all files modified since aparticular date

Back up by directory

Back up automatically by time orcalendar

Back up all but a list of files to beexcluded

Start backup from workstation or server

Back up by interface to a database

Back up using wildcard characters in filenames

Incremental backup

Maintain index on tape and disk

Maintain cross-reference of tapeserial numbers and backup

Back up manually

Back up by list of files

Back up by index

Compress data

Back up multiple volumes

Generate reports

Workstations


109/181

1-109

Workstations

Diskless Workstations A diskless workstation does not have any local disk

drives. Instead, a diskless workstation has its boot logic

in a read-only memory (ROM) chip located on the LAN

adapter. Because diskless workstations have no disk drives, they

are inherently less expensive than those with disks.

Workstation Memory and Speed

The speed of the workstations processor must be

compatible with the type of work for which it is being

used. Basically, it is the application, not the LAN, that

determines the power of the workstations.

LAN Adapters


110/181

1-110

LAN Adapters

If you have chosen an architecture, medium, andmedia access control protocol, you havenarrowed the options for LAN adapters, also

known as a network interface card (NIC). LAN adapters provide the connection between

the medium and the bus of the workstation orserver. LAN adapters are designed to support a

specific protocol using a specific medium,although a few can support two or three differentmedium types.

Printers


111/181

1-111

Printers

One major factor that affects the success of aLAN is printer support. Some LANs haverestrictions regarding the distribution of printers

and the number of printers that can be supportedby one server.

A printer driver is a software module thatdetermines how to format data for proper printingon a specific type of printer. The printers you

intend to use must be supported by the softwaredrivers provided by the LAN operating system orprinter vendor.

Hubs and Switches


112/181

1-112

Hubs and Switches

Some LAN architectures use wiringhubs to provide deviceinterconnection.

Hubs vary in the number of portsavailable. A very commonconfiguration is 8 ports, with 12- and

16-port hubs as commonalternatives. Hubs may be stand-alone or stackable.

Making Connections


113/181

1-113

Making Connections

The objective of network connectionthatis, connecting a computer to the LANmediumis to provide a data path

between the medium and the computersmemory.

The interface or connection to the mediumis called the communications interface

unit (CIU), and the interface or connectionto the computers bus is called the bus

interface unit (BIU).


114/181

Transceiver Located on a LANAd t


115/181

1-115

Adapter

EthernetController

Bus

Interface

Encoder/Decoder

RAM

Transceiver

Cable


116/181

1-116

Software Introduction


117/181

Classes of Software


118/181

1-118

Classes of Software

Application Software

Operating System Software Memory management

The OS manages memory and allocates it among itself and the

requesting processes according to a memory management

scheme adopted by the OS designers

File management

The OS provides the ability to organize bits into files, files into

directories, and so on

User interface Creating user environments is an OS function

I/O interfaces

The OS and file management system translate the request to

access a file into an address the disk can act on

Classes of Software (cont.)


119/181

1-119


Resource allocation

The OS has an algorithm that determines which

ready process gets access to the CPU and the

conditions under which a running process maintains

control of the CPU

Accounting

CPU usage

Lines printed

Disk space used Transactions submitted (by type)

Data communications medium usage

bytes read from and written to disk


120/181



121/181

1-121

( )

Utility Software Utilities augment the function of the OS or other subsystem, such as the

database management system or the data communications subsystem.

Development Software Development software includes text editors, text formatters, compilers,

interpreters, link editors, cross-reference programs, testing utilities, andrelated programs and utilities

Network Management Software Network management includes a variety of activities such as backups,

problem resolution, capacity planning, and establishing user environments

Network Access Software Network access software for a LAN includes the components that allow us

to access LAN resources from directly attached workstations and remote

workstations

Generic Functions of LAN SystemSoftware


122/181

1-122

Software

LAN system software is essentially an extensionof the OS. It carries out hardware-oriented LANtasks, such as interfacing to the medium, and I/O-oriented tasks such as directing print jobs and

disk read/write requests to a server.

The purpose of system software is to insulateapplications from hardware details such as I/Oand memory management. System software

provides an interface through which theapplications can request hardware serviceswithout needing to know the details of how theservices are carried out.

LAN System Software in Serverand Workstation


123/181

1-123

and Workstation

ApplicationProgram

A1Network

Redirector

MediumInterface

Software

OperatingSystem

DiskDrive

ServerOperating

System

MediumInterface

Software

Server

SystemSoftware

DiskDrive

LAN Medium

LAN Redirector Functions


124/181

1-124

ApplicationProgram

A1

NetworkRedirector

MediumInterface

Software

OperatingSystem

LAN Medium

Disk

RedirectorinterceptsI/O requestand routes toappropriate softwaremodule.

NetworkRequest

LAN Server Request Queue


125/181

1-125

q

ServerOperating

System

LANSystem

Software

Database Read

Write to Printer

Download Application

Database Read

Write to Printer

Write to Printer

Queueof

WaitingRequests

ServerDisk


126/181

A Contention Problem


127/181

1-127

Record A

Karen readsRecord A with lock

Juan issues a readrequest for Record A

Record A is locked byKaren so Juan must wait

Record A

Karen's Application

Juan's Application

Read

Record A

PendingRequest

Database

Exclusive, Protected, and SharedOpen Combinations


128/181

1-128

Open Combinations

Exclusive

Protected

Shared, update

Shared, read-only

Denied

Denied

Denied

Denied

Denied

Denied

Denied

Granted

Denied

Denied

Granted

Granted

Denied

Granted

Granted

Granted

Open ModeRequested Exclusive Protected

SharedUpdate Shared Read-only

Currently Opened As

Waiting for Lock Release


129/181

1-129

g

Record A

Application 1 readsRecord A with lock

Application 2 issues a read

request to Record A

Record A is locked byApplication 1, soApplication 2 must wait

Record A

Application 1

Application 2

Read

Record A

PendingRequest

Database

Deadlock Situation


130/181

1-130

Record A

Record BLocked by A2

Application 1 has readand locked Record A

Application 2 has readand locked Record B

Application 1 is waitingfor Application 2to release Record B

Application 2 is waitingfor Application 1to release Record A

Application 1 and 2 aredeadlocked andwill wait forever

Record A

Application 1

Application 2

Read

Record A

PendingRequest

DatabaseRecord B

Software Standards


131/181

1-131

Most of the LANs in operation todayare called open networks, meaning

that they adhere to a set of standardsformulated by organizations such asISO, ANSI, IEEE, and EIA. Thesestandards are placed in the public

domain and are called openstandards.


132/181

Software License Agreements


133/181

1-133

The license agreement states the conditionsunder which you are allowed to use the product.

In general, license agreements take one of the

following forms: Single-user, single-workstation licenses

A single-user, single-workstation license agreement is the most

restrictive. It specifies that the software is to be used on one

workstation only and by only one person at a time.

Single-user, multiple-workstation licenses

The constraints of the single-user, single-workstation agreement are

relaxed by the single-user, multiple-workstation license agreement. It

usually relies on the honor system for enforcement. The purchase of a

single copy of the software allows the owner to install it on several

systems.

Software License Agreements(cont.)


134/181

1-134

(cont.) Restricted number of concurrent users licenses

Most LAN-compatible software is designed formultiple users; however, some software vendors

limit the number of concurrent users with a

restricted number of concurrent users license

agreement. The main idea behind this strategy is tocharge by the number of users.

Server licenses

A server license allows an application to be installed

on one server. All users attached to that server mayuse the application. If a company has several

servers, say three, and wants to use the application

on each of them, the company must purchase three

licenses or three copies of the software.,

Software License Agreements(cont.)


135/181

1-135

(cont.)

Site licenses A site license gives the user unlimited rights to use the

software at a given site. The site may be a single LAN or

multiple LANs at one location.

Corporate licenses A corporate license gives the corporation unlimited use

of the software at all locations. Some companies restrict

a corporate license to all locations within one country.

Sometimes, the right to reproduce documentation is also

granted.

License Agreement Owners

Rights


136/181

1-136

Rights

The license agreement is intendedprimarily to protect the rights of themanufacturer. However, the owner of a

license agreement also has certain rights: The owner can transfer or assign the license to another user.

The owner can get a refund if the product is defective or does not

work as stated.

The owner has legal rights granted by certain states or countries

regarding the exclusion of liability for losses or damage resultingfrom the use of the software.

The owner can terminate the license by destroying the software and

documentation.


137/181

1-137

LAN System Software

Section Preview


138/181

1-138

Generic LAN system softwarefunctions

Workstation system softwarefunctions

Server system software functions

Printer software

Backup software


Server/WorkstationCommunications


139/181

1-139

Communications

The first problem the workstation must address isconnecting to a server. Once a client has a serverconnection, the user at the client can log onto thenetwork. Depending on the LAN OS being used, a

user can log onto a single server, a collection ofresources called a domain, or the entirecorporate network.

Regardless of the environmentsingle server,

domain, or network directorythe users networksoftware must find a server resource that canauthenticate the logon request.

Server/WorkstationCommunications (cont.)


140/181

1-140

Communications (cont.)

The address of the server with which theworkstation must communicate is essentialbecause the address is part of the data linkmessage header.

The network layer protocols determine the formatof addresses. Leading transport/network layerprotocols include the transmission controlprotocol/internet protocol (TCP/IP) used on the

Internet, and the sequenced packet exchange/internet packet exchange (SPX/IPX) protocolused on many Novell and Microsoft LANs.

IP and IPX Header Formats


141/181

1-141

VersionIP

HeaderLength

DestinationAddress

SourceAddress

HeaderChecksum

Typeof

Service

PacketLength

Identification FlagsFragment

Offset

Timeto

LiveProtocol

Checksum

Source

Socket

Source

Node

Source

NetworkLength

Transport

Control

Packet

Type

Destination

Network

Destination

Node

Destination

Socket

IP Header

IPX Header


142/181

A Generic Server Table


143/181

1-143

MktgSvr

AcctSvr

SWDevSvr

HRSvr

MfgSvr

RanDSvr

ExOffL

Unused

File

File

File

SQL

File

File

Printer

1

2

0

1

1

0

1

A12B634A

DDC3958B

DD84A124

A12C583D

29837CAB

834AB3E7

5602C0A1

Server Name Address Server Type Hops


144/181

Possible Contents of a NetworkDirectory


145/181

1-145

Directory

Server

User

Printer

Disk volume

Network address

Full name

Location

Print server

Server name

Manufacturer

Location

Telephone

E-mail address

Operator name

Capacity

Name

Login name

Address

Name

Print queues

Name

Interface

Object Properties


146/181


147/181


148/181

LAN Operating System Functions


149/181

1-149

Optimized I/O One of the main services provided by a server is disk access. Disk

access consists of three components: seek, latency, and transfer. I/O

optimization attempts to reduce one or more of these disk access

components.

Disk Configurations One of the functions of an OS is to implement a file system. This

involves allocating and deallocating disk space and keeping track of

space allocated to each file.

Partitioning

sometimes it is beneficial to divide a single disk drive in two or

more partitions; each partition can be managed separately

LAN Operating System Functions(cont.)


150/181

1-150

Single Disk Volume

a volume is a logical disk (a partition or collection of partitions) or

physical disk that has been formatted and can be used to store data by

an OS.

Multiple Disk Volumes or Volume Sets

Most LAN OSs allow multiple partitions or disks to becombined to form a single logical partition. A volume created

from multiple partitions is called a volume set.

RAID Level 0Striping without parity

Another capability provided with some LAN OSs is called a

Redundant Array of Inexpensive Disks (RAID) Level 0 or

striping without parity. Multiple partitions on different disks

can be combined to proved a single logical disk; striping with

parity differs from a volume just described in that data is

written to all partitions simultaneously.

Fault Tolerance


151/181

1-151

A LAN with fault tolerance allows the server tosurvive some failures that would ordinarily bedisabling. Fault tolerance usually is provided by acombination of backup hardware components

and software capable of using the backuphardware.

A level of fault tolerance also can be provided byusing redundant arrays of inexpensive disks

(RAID). There are six levels of RAID, but for faulttolerance we are concerned only with RAID Level1 and RAID Level 5.

Mirrored Disk Drives


152/181

1-152

Controller 1 Controller 2

File 1 File 1

File 2 File 2

DuplexedControllers

1 2


153/181

A Fault-Tolerant Duplexed Server


154/181

1-154

Dedicated High-SpeedConnection

Duplexed Servers

Disk Drive Disk Drive

Mirrored Disk Drives

Logical and Physical Printers


155/181

1-155

In discussing the operation of a printingenvironment, we need to make adistinction between logical printing and

physical printing. The applications jobstream is sent to a logical printer. Logicalprinters collect print streams from one ormore applications. Logical printers are

connected to zero, one, or severalphysical printers that Microsoft call aprinting device.

LAN Printing Configurations


156/181

1-156

Because printing requirements varyamong companies using LANs, a widerange of implementation options exist.Among these are:

Private printers attached to a users computer and not shared

Shared printers attached to file servers

Shared printers attached to users workstations

Shared printers attached to computers that are dedicated to the

printing task

Shared printers attached to special-purpose print servers

Shared printers attached directly to the LAN via an on-board LAN

adapter and print server

Some Spooler Options


157/181

1-157

Collect printed output

Hold hobs in disk queue afterprinting

View jobs on hold in print queue

Delete jobs from print queue

Se/change job priorities

Start/stop printers

Print banners

Print statistical reports

Direct print jobs to designatedprinters

Hold jobs in disk queue beforeprinting

Set number of print copies

Attach/detach printers from printqueue

Add/delete printers

Start/stop spooler process

Close print jobs based on time-out interval

Backup Software


158/181

1-158

The software used to perform the backups is asimportant as the hardware. Backup software isresponsible for reading the files being backed upand writing them to the backup device.

Backup devices often come with a backup/restoreprogram (both capabilities are contained on oneprogram), and most LAN system softwareincludes a backup/restore module.

Some LAN administrators choose to purchase aseparate, more functional backup system thanthe LAN or backup device versions.


159/181

1-159

Client/ServerArchitecture

Section Preview


160/181

1-160

What is client/server?

Client/server history

Client/server advantages anddisadvantages

Client/server technology

Client/server standards Client/server database

implementations


What Is Client/Server?


161/181

1-161

A client is a software module that relies ona server software module for some of an

applications processing requirements.

The server performs the work necessaryto complete the request and returns theresult to the client.

Client/server computing is: a modular approach to application design

an architecture for software deployment and sharing

a hardware-independent approach to application design

Client/Server Computing: AModular Approach


162/181

1-162

The client/server approach to applicationdesign takes advantage of the fact thatapplications typically perform a variety of

functions. Three of the most commonfunctions are presentation, business, anddatabase logic.

If this type of architecture is adopted,changes in one functional area can beisolated from the remainder of the system.

Single Process Client/ServerConfiguration


163/181

1-163

PresentationLogic

BusinessLogic

DatabaseLogic

Application TerminalDisk drives

Client/Server Computing: SoftwareSharing


164/181

1-164

Client/server computing can takeadvantage of the fact that someapplication logic is common to multipleprograms.

Because most applications usedatabases, the logic these applicationsuse to access and manipulate data ismostly the same. With client/servercomputing, these applications can sharethis logic rather than have it embeddedwithin each program.

Client/Server Computing:Hardware-Independent Approach


165/181

1-165

Client/server provides a way of craftingapplication systems for single computersas well as for networks. Therefore,

client/server applications can be spreadover a wide variety of hardwareconfigurations.

The primary requirement is an ability forthe client and server to communicate.

Client/Server on a Network


166/181

1-166

Bus.Logic

Pres.Logic

Requests and Responses

Clients

Database(SQL)Server

File/PrintServer

Client/Server an HistoricalPerspective


167/181

1-167

In the early days of computing, a programmer often neededto address the issues of hardware interfaces as well as theissues of business logic. This required more skills from theprogrammer, and development time was extensive becausehardware-oriented logic often had to be duplicated in

multiple programs. With the advent of compilers and operating systems, the

application programmers responsibilities were reduced.

Compilers allowed code to be written in high-levellanguages, like COBOL and C++.

ON-line transaction processing (OLTP) provided anotherstep toward client/server computing. The implementationused in leading OLTP systems was the precursor ofmodern client/server computing environments.

Client/Server Models


168/181

1-168

Database Logic Business Logic Presentation Logic



Database Server Model

Server Client

Split Business Logic Model

Server Client

Business Logic

ClientServer

Presentation Client Model

Local Procedure Call Example


169/181

1-169

Procedure P2 (input1, output1)

Procedure P1 (input1, input2, ouput1)Call P2 (param1, param2)

Call P1 (param1, param2, param3)

Program

P2

P1

Main

LocalProcedureCalls

Remote Procedure Call Example


170/181

1-170

Procedure P2 (input1, output1) P2

Server ComputerClient Computer

Call P1 (param1, param2,param3)

Procedure P1 (input1,input2, ouput1)Call P2 (param1, param2)

Local

ProcedureCall

P1

Main

RemoteProcedure

Call

Program

SQL Server


171/181

1-171

With SQL server technology, the databaseprocessing is done on the server. In file servertechnology, the database processing is theresponsibility of the client. The advantages of the

SQL server technology are as follows: Less data is transmitted over the network.

Database processing that can be optimized for that task is performed on the

server.

If database processing demands increase, more powerful SQL servers can

be installed to accommodate the increased workload.

Processing is distributed over multiple computers, each of which can be

optimized for the tasks it performs.

Relationship Between SQL Serverand Client


172/181

1-172

BusinessLogic

Database

DriverVendor A

DatabaseDriver

Vendor B

Com.Interface

Client Computer

Vendor A

Vendor B

Database(SQL)

Servers


173/181


174/181

Client/Server Standards


175/181

1-175

Client/server standards have been formulated to allow opensystems and provide a mechanisms for interoperability ofclients and servers. The standards being formed include:

how clients and servers communicate with each other

how clients find the server or servers that perform their needed functions

Interfaces between clients and servers have come to becalled middleware. Middleware is responsible for makingthe connection between clients and servers.

Examples of middleware and its standardization efforts are:

distributed computing environment (DCE) specifications

object request broker (ORB)

open database connectivity (ODBC)

A Look to the Future


176/181

1-176

Operating system and application softwareevolution is placing ever higher demands fordisk, memory, and processing. These increasingdemands together with a dependence on desktopcomputing have the following consequences:

Newer, higher capacity computers are needed to replace older, lowercapacity computers.

Existing computers need upgrading to accommodate new software.

Installing and supporting distributed applications can be expensive.

Managing centralized systems is easier than managing decentralized

systems.

In an effort to trim the costs associated withsoftware evolution, some companies are adoptingthe use of thin clients and server clusters.

A Look to the Future (cont.)


177/181

1-177

A thin client is a stripped-down microcomputer.The function of the thin client is to provide basicinput and output much like the terminals in theearlier centralized host computer architectures.

A cluster servers consists of two or more servers

operating in concert with each other.

The benefits of the then client architecture are: Lower costs because the thin clients are less expensive to buy

Obsolescence is not as likely for thin clients because they do not need to

keep pace with expanding software needs Software distribution and maintenance is simpler when restricted to the

servers

Management and control of applications and data is easier in a centralized

architecture

Thin Client Architecture


178/181

1-178

High-speed Interface Between Servers

Server Cluster

Database(SQL)

Servers

File/PrintServers

ApplicationServers

TerminalServer

ThinClients


179/181

Assignment/Exercise 5/6: Local/Wide AreaNetworks

What is the Institute for Electrical and ElectronicEngineers (IEEE) Committee 802? What are its


180/181

Engineers (IEEE) Committee 802? What are its

functions and sub-committees? Describe the following (potential) wide area network

physical/data link layer connection technologies: plain old telephone system

Leased line (64kbps, T1/E1, fractional T1/E1, etc.)

Integrated Services Digital Network (ISDN) Digital Subscriber Lines (DSL)

Frame Relay

Asynchronous Transfer Mode (ATM)

What is a repeater, a bridge, a switch, a router and a

gateway? In what OSI layer does each of thesedevices functions?

Submit: a 2-4 page write-up of your findings (be briefbut concise!)

Due: 19/26 Jan. 2002

Assignment/Exercise 5/6: Local/Wide AreaNetworks

Familiarize with USENET Newsgroups related to localk h i


181/181

area networks, that is:

comp.dcom.lans.ethernet comp.dcom.lans.fddi

comp.dcom.lans.novell

comp.dcom.lans.token-ring

comp.dcom.lans.misc

etc

Tip: For starters, read the Frequently Asked Questions(FAQ) for each area.

Where can you access USENET News? A lot of sites in

the Internet:

Documents

CS 55 - Local Area Networks