chapter1.pdf

Chapter 1 - Local Area Network (LAN) Overview 1996, BICSI LAN Design Manual - CD-ROM, Issue 11

Defining LANs

Definition of a LANThe Institute of Electrical and Electronics Engineers (IEEE) defines a LAN as follows:

A datacom system allowing a number of independent devices tocommunicate directly with each other, within a moderately sizedgeographic area over a physical communications channel of moderatedata rates.

In its simplest form a Local Area Network (LAN) is a set of Personal Computers (PCs) andprinters connected together in a defined, limited geographic area. The connected PCs arereferred to as stations.Technically, two connectedPCs next to each other canbe considered a LANthiswould be a two-station LAN,the smallest possibleconfiguration.

FIGURE 1.1:A TWO-STATION LAN

Transmission Medium

PC 1 PC 2

Definition of a LAN, continued


Typical characteristics of a LAN environment: The stations on the network are peersany station can initiate data exchange with

any other station. Full connectivity among all stations. Fully administered by the owner. Runs over a shared transmission mediumoften, cabling. The network is confined to a small areaa single building or a cluster of buildings. The data rate is highseveral Mbps (million bits per second).


LANs vs. WANsThe words limited geographic area are used inthe definition of a LAN to highlight the fact that theL in LAN stands for Local. When computers areconnected across town or across cities, countriesor continents, the L becomes a W, indicating aWide Area Network, or WAN.A WAN is created every time computersare connected over long distances usingtelecommunications linkse.g.,telephone lines, satellites, microwave.

FIGURE 1.2:A WIDE AREA NETWORK


Purpose of a LANA LAN permits users to share resourceshardware, software or user-created files. Sharingof resources makes it possible to maximize the investments made in each resource.Ideally, distant resources should appear to be local to the user.

Objectives of an effective LANWhen implemented, LANs are expected to achieve certain basic objectives:

To improve employee productivity. To improve information management. To improve interaction between staff. To reduce/control costs. To provide for standardized hardware and software usage.

An effective LAN is one that meets objectives while fulfilling certain fundamentalrequirements:

Simplicity. Reliability. Transparency. Manageability.

Objectives of an effective LAN,continued


SimplicityThe technology should be easy to use.

Telephones are an example of such a technology. People are able to usea telephone with minimal training. It therefore becomes a useful toolrather than something to be avoided.

Working with the LAN must be as simple as working with a stand-alone PC. Otherwise,users may refuse to accept the LAN.Employee productivity is unlikely to increase if the LAN is difficult to use. Similarly,interaction between staff is unlikely to improve if users are unable to gain access to thesystem.

ReliabilityIf the LAN does not work reliably, people may not use it. Reliability is often seen as thesingle factor having the greatest impact on whether or not a LAN is accepted.A system is seen as being reliable if:

Devices are functional when needed. Access to distant shared devices and files is as fast as when these devices/files

are on the users local PC.

The voice network illustrates this concept of reliability. It works whenexpected, as expected.



TransparencyAn understanding of the technology used to implement the LAN should not be aprerequisite to using it. The technical aspect of making a LAN work should not be seenby the users. The art of good technology is that it isnt seen.

Few people understand how a telephone works, but they have littledifficulty using it.

Users should be able to work on a LAN in the same manner as they would with anon-LAN-attached PC.



ManageabilityThe LAN must be easy to manage and administer. This includes all aspects of the LANenvironment:

The hardware. The software.

The people.The planning and installation of the LAN determines how manageable it will be in thefuture.Factors to be considered at the earliest stages of LAN implementation include examining:

Possible points-of-failure. Future growth requirements. The expense of downtime.


Characteristics of a LANThere are four characteristics used to describe the architecture of a given LAN. Thesecharacteristics allow for comparisons among the many varied LAN environments.

Transmission medium.The type of cable/wireless system used to connect the network devices.

Topology.The physical appearance and/or manner of operation over the transmissionmedium when used to connect network devices.

Access control method.In a LAN environment, many devices share a common transmission medium butgenerally, only a few are able to use it at any given instant. Access controldetermines the way in which network devices are granted or denied access to thetransmission medium.

Transmission technique.The manner in which the transmission medium is used for communications.


Evolution of LANs

The concept of networkingA typical definition of a data network today would read as follows:

A data network is an interconnected system of computers,peripherals and software over which data, files and messages are sentand received.

A LAN is only one type of computer network. Before LANs came into being, there were othertypes of networks.The original messaging networkthe telephone systemdid not even involve computers.Computer networks came into prominence after the appearance of computers in thecorporate world in the 1950s. (Invented in the 1940s, modern-day computers were first putto use by U.S. and British national defense organizations and university laboratories.)

The concept of networking,continued


Key years in the history of computer networking are the following: 1950s Geographically dispersed university research mainframes are connected

for defense-related work in the United States. 1964 The first commercial mainframe network used for airline passenger

information and reservations in the United States. 1968 Nationwide air-traffic control in the United States. 1970s The Advanced Research Projects Agency Network (ARPANET) connects

computers from multiple vendors scattered across the United States. 1980 Xerox, Digital Equipment Corp. and Intel announce the Ethernet standard

for office networking.


Development of computer networks

The mainframe environmentThroughout the 1950s and 1960s, there was only one type of computer available forpurchasethe mainframe. Available since those early years of computing, the mostnotable feature of the mainframe is its centralized processing. All computing operationstake place in the central unit. A single mainframe can support hundreds of terminals(consisting of a keyboard and a display screen), all simultaneously issuing and receivingvarious instructions through constant communications with the mainframe.Mainframes have always been costly machines. In the early years, most companies couldonly afford a single mainframebut their offices may have been scattered across thecountry. In order for a users terminal to communicate with the mainframe, a way wasneeded to connect the two over various distancesthus, computer networking was born.

Development of computernetworks, continued


FIGURE 1.3:THE TYPICALMAINFRAMEENVIRONMENT

Mainframe computer Controller

Accounting department

Department of Human resources

Research and Development



The minicomputer environmentThe 1970s saw the introduction of minicomputers (minis)essentially scaled-downversions of mainframes. Minis are generally less powerful than mainframes, supportingfewer terminals in simultaneous use.The lower cost of minis:

Permitted smaller companies to afford their own computer. Allowed large companies to buy several low-cost minis as computing services

became necessary for a given department or region.From a networking perspective, minis are no different from mainframesterminals areconnected to a central unit. Itis more likely to see the minilocated in the same buildingas the terminals it supports,whereas a mainframe is morelikely to support a largenumber of remoteterminalscommunicatingwith the mainframe overtelephone lines.

FIGURE 1.4:THE TYPICALMINICOMPUTERENVIRONMENT

MinicomputerAccounting department

Department of Human resources

Research and Development

Minicomputer

Minicomputer



The Personal Computer (PC) environmentIn the 1980s PCs were introduced. Due to their much lower cost and useful software, theirpopularity grew quickly.Unlike mainframes and minis, PCs need no terminals to connect to a central unit. Theentire systemterminal, keyboard and processoris contained in a single box. The userhas more freedom when it comes to choosing, purchasing and installing software on thisPCno other individual or group needs to be involved in the process.PCs bring a great deal of freedom to individuals accustomed to dealing with programmersand systems analysts to obtain computer services (with all of the associated delays,misunderstandings and arguments). They can also create an administrative nightmare for

the organization if not properly designed and managed.

FIGURE 1.5: THE TYPICAL PC ENVIRONMENTDue to their lower cost, PCs were oftenpurchased in large numbers by every

department, group or individual in anorganization with an expense

budget. Some companiesattempted to control theprocess throughcentralized purchasing, butthey could not prevent

users from purchasing and usingthe software packages of their

choice.

Monitor

Keyboard

Computercontaining processingunit and disk drive(s)



Terminal emulationCompanies which owned mainframes and minicomputers saw PC users demand access tothese machines in order to retrieve and work with corporate data. This created the need toconnect the PCs to the mini or mainframe. Hardware and software products had to bepurchased for each PC to make it appear as an ordinary terminal to the mini or mainframe.This process is called Terminal Emulation. The PC emulates or pretends to be an ordinaryterminal.

The LAN environmentStand-alone PCs cause a variety of problems:

Each PC holds its own data, created by its owner, with no guarantee of accuracy;two people could create the same report and show different figures, making bothreports unreliable for decision making.

Security is a probleman intruder could walk up to any PC to gain access to itsfiles.

Users do not appreciate the need to make backup copies of their files until disasterstrikescausing losses in terms of time and money.

Users work with different software packages, or different versions of the samesoftware package, making their files difficult or impossible to share.

Expensive peripheralsplotters, laser printers or large hard disksare difficult tojustify for each PC in the company; users are required to go to a PC where theseperipherals are installed and often have to wait to use themlowering productivity.



As a solution to these problems, LANs began to appear in the mid-1980s. They offerbalance between the need for management control and the desire for personal freedom inchoosing business tools.In many respects, LANs strongly resemble the mini and mainframe environment:

LANs allow for the centralization of dataall reports draw their data from a singlelocation, guaranteeing consistency.

LANs enforce security by requiring passwords and access privileges in order towork with files.

LANs permit effective, global backupsbackups can take place automaticallywithout the need to have anyone stay after hours.

LANs make it possible for a company to quickly transfer all users to a new versionof a software package by installing the upgrade on the PC which stores the sharedprogram.

In comparison to the mini and mainframe environments, LANs were and continue to beless expensive. The level of competition in the PC industry helps to keep prices low.


Enterprise-wide computingThe goal of the computer industry is to provide an environment where all members of anorganization have access to the data and resources they need to perform a given task.Business-wide LAN systems, including corporate minicomputers and mainframes accessedthrough the LAN, provide transparent data access to the user in a secure environment.Todays computing technologies can be used with a three-stage process to achieve this goal.

Stage 1: Personal productivityAt this stage, individuals in the organization each have a PC equipped with personalproductivity tools such as spreadsheet or word processing software. The goal of theorganization is to encourage employees to become PC-literate by providing the tools,training and support.

Stage 2: Workgroup LANsAt the second stage, the PCs in a department are connected together to form a LAN.Peripheral device, program and file sharing concepts are explained and encouraged. Thegoal of the organization is to encourage employees to become LAN-literate by providingthe tools, training and support.

Enterprise-wide computing,continued


Stage 3: The integrated organizationThe final stage is to connect all departmental LANs together, whether they are in the samebuilding or far apart, to create a company-wide communications channel. Data sharing andcommunications take place between all employees and departments, across countries andtime zones. Such an environment requires a great deal of expertise on the part of theusers and especially the administrators of the LANs.

Future trendsThe trend towards many small computers instead of a single large one is expected tocontinue. The incremental growth and flexibility offered by this approach fits well with therapidly changing nature of todays organizations.As buyers move to a computing environment dominated by LANs, they are demanding thatvendors provide products which conform to accepted standards. Standards relevant to theLAN environment are discussed in a later chapter.


Benefits of a LAN

Shared resources

HardwareThe demand to be able to share expensive hardware led to the introduction of LANs in themid-1980s. Specifically, the introduction of the first PCs equipped with a hard diskprompted the need for the sharing of resources. These PCs, costing thousands of dollars,came equipped with a 5 Megabyte hard diskwhich was deemed to be too much storagespace for only one userso a method for sharing this hard disk had to be found. Theintention of these first LANs was to be able to share these large hard disks.The sharing of expensive peripheral devicessuch as color laser printers or high-speedoptical mass storage unitscontinues to be one of the benefits of a LAN.

Shared resources, continued


SoftwareIn addition to the ability to share hardware, a LAN provides the user of one station theability to access files and programs stored on another station. This program sharing cansave substantial sums of money.With stand-alone PCs, software must be purchased for every PC in the office. The cost ofbuying multiple copies of software is high. Software vendors, eager to avoid widespreadpiracy (unauthorized copying) of their products, have introduced versions of their softwaredesigned specifically for LANs. In such arrangements, the company pays for a licensewhich allows a given number of users to simultaneously access the program files.As the number of users who need access to the package grows, the company purchasesadditional licensesindividually or in bundles (e.g., 5-user or 10-user licenses).Sharing of software resources can also lead to the added benefit of standardizing on alimited number of software packagestwo word processing packages instead of tenandthe number of versions of a package.With the increased appearance of LANs in the typical office environment, a new categoryof software has been developedgroupware. Groupware is software specifically designedfor use on a LAN. It permits, for example, multiple users to work on a given document atthe same time, encouraging collaboration.

Shared resources, continued


FilesFiles can be shared by users on the LAN just as they share software programs. By placingfiles in a common areaa hard disk directory where more than one user has accessprivilegesfile sharing can take place. This common area is usually found on a file server.Since only one version of a file exists with file sharing, the likelihood of users updatingdifferent versions of the same filecreating inconsistencies and data redundancyiseliminated.File and program sharing have the added benefit of saving disk space. Files that areshareable need only be kept on one hard disk. In many cases, the saved disk space issubstantial.

Incremental growthLANs grow as users needs grow. If it is discovered that users sharing a printer are waitingtoo long for their jobs to print out, an additional printer can be added to the LAN. Money isspent only when needed.Contrast this with a minicomputer or mainframe environment where the bulk of the expenseis in purchasing the central computerthe terminals used for access to such machines aretypically less expensive than PCs. When the limits of the minicomputer or mainframe arereached, the company has little choice but to upgradeeither by purchasing options or byreplacing the unit with a more powerful one. Each option is costly and typically leaves thecompany with unused capacity for which if has already paidan inefficient use of resources.


Security

Access controlSecurity features of LANs permit the network administrator to allocate privileges amongusers in such a way as to permit data sharing to occur while respecting securityrequirements. Not all users will require access to all files and different users will requiredifferent levels of access to certain files.

For example, payroll files may need to be accessed by the Vice-President ofFinance, the Controller and two payroll clerksall of whom are users on theLAN. The network administrator must make certain that only theseindividuals, and no others, have access to this confidential information.There are also additional requirements beyond simple access to files. Thepayroll clerks may be authorized to add new payroll data, but not view orchange existing payroll records. The Vice-President and the Controller canview payroll records, but not change them.The network administrator is responsible for assigning access privileges sothat the Vice-President and the Controller each have read-only privileges,while the clerks are assigned write-only privileges.With such a setup, the clerks cannot look through the file and see thingsthey shouldnt and management cannot change the existing datagreatlyreducing the potential for abuse.

Security, continued


BackupsCentralized backups prevent a failed system from causing permanent damage to theorganization.

Audit trailsAudit trails provide details such as user name and time of access for some or all networkfiles. Being able to track who accessed individual files and when they did so discouragesabuse and greatly enhances management control.

Cost accountingCost accounting is provided through the use of chargeback software. Such softwarepermits individual users or entire departments to be allocated the cost for their share ofLAN use. Use of a chargeback system makes budgeting for the network a much simplertask.


Improved communicationsBy connecting PCs together, a new communications channel is created within the company.It functions alongside other, more traditional channels, such as the telephone, formalmeetings, lunchroom chats and company-sponsored events.

ProductivityBy using products such as electronic mail, scheduling and/or electronic bulletin boards,employees can use the LAN to better coordinate their activities.A LAN can also permit users located outside the office to access network resources.Remote access software allows users to remain in touch with colleagues and tasks.

Customer relationsImproving communications between employees reduces the organizations response timeto market changes. This provides a competitive advantage by allowing a quicker responseto the needs of the customer.The company can also use a LAN to forge closer links with its customers and/or itssuppliers by allowing access to some of its files or its LAN electronic mail system.

For example, customers and suppliers can access the companyswarehouse inventory file in read-only mode to prepare orders or shipments,and they can inform staff by sending messages through the companys LANelectronic mail system.


SummaryThere are many benefits a LAN can providesome easily measurable, others not so easy tomeasure. It remains the responsibility of managers to hire and develop those individuals whowill actively seek these benefits. By itself, a LAN is simply an information-delivery systemitrepresents the piping of a communications channel. Competitive advantage comes aboutthrough unique and innovative uses of the channel.


Components of a LAN

IntroductionA LAN is made up of three basic elements:

The hardware which is connected to form the LAN. The software (or programs) which is accessed through the LAN. The users, who create, work with and manage the various files.

Each of these elements can be divided into a number of components. Each componentrepresents only a part of the whole system and a LAN will only function properly if each partoperates according to specifications.A poorly designed network, one with mismatched components, often results in below-averageperformance. This becomes ever more critical as LANs become more sophisticated in theirabilities.


LAN hardwareA LAN can be thought of as a system composed of a series of building blocks. Theseblocks can be added and configured as needed. Though LANs come in a variety ofconfigurations and can connect from two to thousands of devices, it is possible to groupthese building blocks according to the role each plays in the LAN environment.The hardware components found on a typical LAN include the following items:

Transmission channel.

The medium which connects the network devices. Network Interface Cards (NICs) for attached devices. Servers.

File servers. Print servers.

Communications servers. Stations.

Local. Remote.

Hubs and switches.

LAN hardware, continued


Shared peripheral devices. Printers.

Hard disk drives. CD-ROM drives. Modems.

Following are a series of illustrations. Please note that any single LAN illustration can bemade too complex by trying to show all possible configurations. The following diagrams arean attempt to balance the need for detail with the need for comprehension. In each diagram,specific hardware components are identified.



FIGURE 1.6:LAN HARDWARECOMPONENTS


Local Station

Network Interface CardHub

File Server

Shared Hard Disk

Print Server

Shared Printer

Shared CD-ROM Drive

Communications Server

Shared Modem

Trans missionChannel

Remote Station


FIGURE 1.7:DEPARTMENTAL LAN


File Server

Hub

5th FloorAccountingDepartment

Non-dedicated Server(Combined Serverand Station)

Station

Station

Station

Station

Station

Station

Shared Printer

Telecommunications Closet

AccountingSoftware


FIGURE 1.8: CAMPUS LAN


IntrabuildingBackbone

Cable

InterbuildingBackbone Cable

Building 1

Building 2

File Server

File Server

File Server File Server

Station

SharedPrinter

Station

Hub


FIGURE 1.9: REMOTE ACCESS TO A LAN


File Server

Local Station

Shared Modem

CommunicationsServer

Remote Station

Remote Station


Transmission channelThe primary purpose of any LAN is the ability to transmit messages from one networkeddevice to another. Typically, such transmission channels are in the form of cablesphysically connecting devices, although certain wireless transmission channels areavailable.This physical infrastructure provides the foundation for all other devices and if it is notfunctional and stable, it can be guaranteed that none of the other components will be ableto function as desired.

The most common transmission channels are made up of some type of cabletwisted-pair, coaxial or optical fiber cableand corresponding connection hardware. Each of theseis discussed in some detail in the following section.A distinction must be made between the transmission channel used locally and that usedfor remote access to the LAN:

The local transmission channel is often limited to a single building, or at most to acluster of buildings closely co-located.

The transmission channel used for remote access to a LAN is often part of thepublic network.



Network Interface Cards (NICs)Networked stations require a means to connect to the transmission channel. They do sothrough a circuit board referred to as a Network Interface Card (NIC). The NIC allows adevice to be attached to a LAN and all LAN devices must be equipped with a NIC.The NIC plugs into an available expansion slot in the device to be networked, and thetransmission medium is attached to a connector on the NIC.

ServersServers manage the shared resources on the LAN. The server combines hardware andsoftware to offer (or serve) network resources.The server hardware may be an ordinary PC or a high-performance unit designedspecifically to be a server. The software running on the server will vary with the type ofserver. (It is possible for a LAN to have many different types of servers, each providing fordifferent types of services.)LAN servers often resemble host machines (mainframes or minicomputers) conceptuallyand diagramatically. There is one fundamental difference between the two:

In a host-based system, all processing takes places in the central host machine.The attached terminals lack a microprocessor and are therefore unable to processany information themselves. Processing is centralizedterminals are totallydependent on the host device for all processing functions.



A LAN server resembles a host machine in that it also provides shared functions. Itdiffers from a host machine in that it is not the only source of processing power.LAN stations are considered to be intelligent devices and are capable of processinginformation themselves. In the LAN environment, processing is distributed amongall of the intelligent devicesservers, stations and peripherals.



File ServersAll LANs typically have at least one type of servera File Server. The role of the fileserver is to provide and manage a shared storage area on the network.Although each computer on the LAN could create and control its own files and then makethose files available to others, it is preferable to provide a centralized storage andmanagement facility. It cannot be guaranteed that the individual PCs will always beavailable. Using a file server ensures shared files will be available to users when needed.File management allows multiple users access to files. The file server controlssimultaneous access to files, enforces access rights and restrictions, and provides adirectory structure that recognizes file names and supports the grouping of files.While the primary function of a file server is to manage the shared storage device(s), italso performs other valuable services. Some of these services include the following:

File server software tracks authorized users and maintains listings of privileges andauthorizations for these individual LAN users. These lists reflect the accessprivileges each user has to files and programs.

A file server makes it possible to provide a working environment independent of thestation. It allows a user to be able to work at different stations at different times (astation may fail, users may need to work at multiple physical locations, stationsmay be managed as a common pool, etc.). If all files and programs can be found ina central location, they can be accessed by any attached station.

Centralizing the location of files and programs simplifies the backupand ifrequired, the restorationof files. Backups must be performed regularly as adefense against storage device failure or user error.



Print ServersA print server acts as a centralized printing location. Typically, it is a PC or other deviceconnected to at least onebut often multipleprinters. It is able to handle the printingrequirements for a large number of stations.Print servers commonly use a technique known as spooling. A spool (SimultaneousPeripheral Operation On Line) is a combination of hardware and software that redirectsrequests destined for a printer (a relatively slow output device) to a hard disk (a muchfaster storage device).

When a print request comes in to the server, it is spooled onto the hard disk (i.e., itis written to the hard disk instead of being sent directly to the printer). Files areorganized on the hard disk in a first-in, first-out queue (although most print serversoftware allows files to be reordered). Files are retrievedin orderfrom the harddisk and printed.

Spooling files before printing overcomes two common problems: A print file stored on a hard disk can be read one block at a time and printed. This

is useful when the file to be printed is larger than the space available in mainmemory.

Print requests are queued (lined up) so users can continue with their work withoutwaiting for the printer to become available. This is important when multiple userswork with a single printer and many print requests may come in while a file isprinting.



Communications ServersAccess to distant (non-local) systems often requires the use of modems. Modems orequivalents are used whenever connections to the public network are necessary.Only a small percent of the total number of users on the LAN will require access to amodem at any given time. For this reason, a server with one or more attached modemsand corresponding telephone lines can support multiple LAN users.Use of a communications server is more economical than individual modems andtelephone lines with monthly charges for each user. Calling controls and security can beimplemented far more easily using communications servers.

StationsThe typical station on a LAN is a PC. In certain environments where security is a criticalissue, it is desirable to have a type of station known as a Diskless Station. These are PCswhich contain no hard disks or diskette drives. These devices function as any other stationbut the user is unable to store any files or programs locally.A common feature of all stations is their ability to function independentlythey do notneed to be part of the LAN in order to function. LANs do not depend on terminals thatrequire a host processor for operation.A station is considered to be local when it is connected to a server directly through thelocal transmission channel. A station is considered to be remote when it connects to aserver through a telecommunications link using a modem or equivalent.



HubsAlso referred to as a concentrator or switching node, a hub provides connections to andfrom multiple network devices.Hubs are useful for their centralized management facilities and for their ability to isolatestations from disruptions.Hubs are available in a variety of forms and functions:

They may be passive devices simply redirecting signals from one network device toanother.

They may be active devices where incoming signals are regenerated before beingsent to another device.

They may contain bridging, routing, network management and other sophisticatedmodules to actively control the LAN.

SwitchesSwitching hubs, or simply switches, are devices that provide dedicated circuits orconnections to individual LAN stations or LAN segments. They use a technique referred toas microsegmentation. This allows individual stations to have a direct transmissionchannel to the switch. The switch then handles all connections between different stationsneeding to communicate with each other.Switches provide a newer technology that can greatly reduce the traffic on the LAN. Byproviding a station with a dedicated transmission channel, there is no sharing of thebandwidth of the transmission channeleach station gets the full bandwidth.



Shared peripheral devicesA peripheral device is any devicesuch as a printer, hard disk drive, CD-ROM drive ormodemthat is connected to and controlled by a computer. Any or all of these devicescan be accessed by multiple users when connected to a LAN in the proper manner.A peripheral device can be shared by all users by connecting the device to a file server ora specialized server (such as a print server or communications server). A peripheral mayalso be attached to a users station. It can be accessed by other stationsbut only whenthe station to which it is attached is itself connected to the LAN.


LAN softwareOnce the physical building blocks of the LAN are put into place, the next step is to makethem functional. Software is needed for devices to function cooperatively and effectively onthe LAN.There are three categories of software found on a LAN:

The operating system of each attached server. The operating system of each attached station. Applications software accessed by LAN users.

LAN software, continued


Server operating systemsThe server operating system is considered to be the brains of the network. It controls themost critical aspects of network operations:

Network performance. Network management. File integrity. Access security.

Each of the file servers on a LAN is controlled by an operating system, which manages allactivities taking place inside that file server. Unlike a station, which has only one useraccessing its files at any time, a file server must handle simultaneous requests frommultiple users.From its position in the file server, the server operating system must satisfy stationdemands for programs, files, printing resources and communications services whilemaintaining network security. In this capacity, a network operating system found on a LANserver is very similar to the operating systems which run minicomputers and mainframes.



Station operating systemsAll PCs require an operating system to function. A station operating system is designed fora stand-alone PC and provides access to programs, files, printing resources andcommunications services found on that PC. When the PC becomes a station on a LAN,the PC operating system remains unaware of the change. It does not recognize that it nowhas access to LAN resources.

The station operating system must be made responsible for establishing the connectionwith the network and the file server and control communications flow between the stationand the file server.Often, the modification to the station operating system is software which is called a shellor shell software. The term comes from the role this software plays on the station. Afterit is installed, it covers the operating system running on the PC.When the user at the station requests a program or a file, sends a file to a printer or sendsa message to another station, the shell intercepts the request and examines it. If the shellfinds that the request can be handled by the station, it passes the request to the stationoperating system. If the request is for LAN resources, the shell sends it to the networkinterface card (NIC) in the station, which places the request on the transmission channeland sends it to the server.



Applications softwareApplications software is the term given to software used to perform a specific task. Themost common business applications are word processing, spreadsheet analysis anddatabase management.In a LAN environment, the program files necessary to run these applications are usuallyplaced on the file server to permit shared access. Note that applications software whichresides on the hard disk of a station is not considered LAN software because it cannot beaccessed by other users, even though the stations themselves may be connected. Bycontrast, an applications software that resides on a file server but can only be accessed byone individual for security reasons is considered to be LAN software because it can beaccessed by other users if the administrator grants them access privileges.

Client/server computingA more recent method of sharing software is called client/server computing. In client/server computing, the applications software is created and sold for use expressly on aLAN. Client/server software has two distinct partsthe client part which runs on theusers station and the server part which is installed on the file server.With traditional applications software, all of the files are installed on the file server. Whena user runs the software, all of the needed program files are transferred across thetransmission channel to the station. When the user requests data files to use with theprogram, those files must also be transferred.



In the client/server environment, when the user first makes a request for a program, onlythe client portion of the program is sent to the stationnot the entire program. This clientportion permits the user to make inquiries of data files. When the server receives aninquiry from a station, rather than send the entire data file to the station, it performs theinquiry locally and sends only the results to the station. This dramatically reduces thetraffic on the transmission channel.An additional benefit of client/server computing is data integrity. Since the data files neverleave the server, there is less likelihood of file corruption.

GroupwareA second type of application software has been introduced for the LAN environmentgroupware.As the name implies, groupware is software designed specifically for use in a LANenvironment by a group of individuals with common goals and responsibilities. This groupmay be one department, a project team or all employees in an organization.At its core, groupware manages the interactions between the members of the team bytracking their schedules, by providing electronic mail boxes for communication and bypermitting people to work on documents simultaneously. The software acts as a centraladministrator, allowing individuals to work on different parts of a project while trackingprogress as a whole.Groupware is particularly useful to teams whose members are geographically dispersedover many time zones. Instead of coordinating activities through ongoing long-distancephone calls and/or periodic meetings, the members use the groupware as their office.


The peopleAmong the most important elements of a LAN are the people. The purpose of a LAN is toallow the sharing of resources. This sharing is done by peoplemaking them an integral partof the structure.With any LAN there are two groups of people involvedthose who use the resources andthose who manage the resources.

The usersA user is defined as a person who makes use of the network resources. This person usesa station to access the server(s) and work with the resources stored there.Although the term user combines all of the individuals using a network, it is a variedcollection. Within the group will be individuals who are very knowledgeable about PCs,those who know how to use only a single application package, and everyone else inbetween. Due to the varying levels of competence, the LAN must be effortless to workwith. The easier a LAN is to use, the better the chance that people will actually make useof it.Making a LAN easy to use is a two-step procedure:1. Design and configure the LAN properlythis avoids having to make changes at a

later time, which is inconvenient and frustrating to users.2. Train users on LAN operationsthis helps users gain confidence in their ability to

work with the LAN.

The people, continued


The network administratorsThe network administrator is the individual responsible for maintaining the LAN. It isessential that the administrator have a good understanding of how the network is puttogether and how it functions. Responsibilities of an administrator include:

All aspects of maintenance and troubleshooting. Making final decisions regarding the manner of installation of new software. Reconfiguring the network for performance, security or changes. Addressing user inquiries. Keeping up-to-date with changes in the industry. Ensuring the proper use of LAN software and equipment. Maintaining standards and proper licensing.


LAN architectures

IntroductionIf asked to define the word architecture, most people would answer that it includes the designand all of the information needed to construct a building. Just as the architecture of a buildingdetermines its appearance and function, a LAN architecture defines the LANs appearanceand function.

The appearance of the LAN is primarily defined by its transmission channelinfrastructurein most cases, its cabling setup.

The function of the LAN is determined by the manner in which the devices on theLAN communicate with each other.

Introduction, continued


The heading architecture groups together the fundamental technical ingredients thattogether make a LAN. There are five components that are used to define the architectureof a LAN:

Transmission medium defines what is used to connect the devices on a LAN. LAN topology defines how the transmission medium is used to connect the

devices on the LAN. Access control defines the way in which a station on the LAN gains access to the

transmission medium. Transmission technique is the manner by which the signals or messages sent by

the station travel over the transmission medium. Transmission speed is how fast the signals or messages are able to travel over

the transmission medium.


Transmission mediaCommonly the term transmission medium (or simply medium) is referred to as cabling.However, this is no longer an accurate description. In places where a physical cableconnection is not feasible, devices can communicate using wireless systemsspecializedradio or infrared equipment.In all cases, the physical infrastructure of the LAN is critical to the successful operation of theLAN. Without it, devices would be unable to communicatewhich was the reason forinstalling a LAN in the first place.Cabling media can be classified into three categoriestwisted pair, coaxial and optical fiber.Each has its own characteristics, advantages and disadvantages.

Twisted-pair cablesA twisted-pair consists of two individual insulated copper wires physically twisted together.The two wires are twisted together to minimize unwanted electromagnetic signals frominterfering with or radiating from the pair.A wire pair acts as a single telecommunications path. Typically, a number of twisted-pairsare bundled into a cable by wrapping them in a protective sheath.

FIGURE 1.10:TWISTED-PAIR WIRE

Transmission media, continued


Unshielded twisted-pair (UTP)Historically, UTP cable was referred to as telephone wire. This is no longer an appropriatereference. Technical advances in the area of UTP cabling have transformed UTP into ahigh-quality channel capable of accommodating high-speed LAN systems.

FIGURE 1.11:UNSHIELDED TWISTED-PAIR CABLE

UTP Advantages It has a large installed base, and is

a familiar technology. It is relatively inexpensive. Most LAN systems are readily capable of running over UTP.

UTP Disadvantages High quality UTP requires specialized installation procedures. UTP is more sensitive to external electromagnetic interference than other media.

Cable jacket



Shielded twisted-pair (STP)Shielded twisted-pair cables have individual wire pairs covered with a layer of metallicshielding to further reduce interference-related problems.

FIGURE 1.12:SHIELDED TWISTED-PAIR CABLE

STP Advantage Provides better

performance thanUTP in environments withhigh noise levelshigh levels of unwanted electrical signals.

STP Disadvantages It is more labor intensive than UTP. Not all LAN systems work readily over STP.

Shielding around individual pairs

Overall shieldCable jacket



Coaxial cablesCoaxial cable is commonly referred to as coax. All coax consists of a central copper coresurrounded by a layer of insulating material. This insulation is enveloped by a metallic wiremesh or, in some cases, a solid metallic sleeve. All of this is then protected by an outerlayer of nonconducting material. Both the centralcore and the mesh or sleeve are capable ofconducting electrical signals.

FIGURE 1.13:COAXIAL CABLE

Coax Advantages It is less

susceptible tointerference thantwisted-pair cable.

It is theoretically capable of supporting higher data transmission rates than twisted-pair cable.

Coax Disadvantages There are many types of coax cables, each suited for one, or at most, a small

number of LAN systems. Due to its high metallic and insulation content, coax cable is usually more

expensive than other cable types.

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

AAAA

Central copper core

Metallic mesh or sleeve

InsulationNon-conducting outer layer



Optical fiber cablesOptical fiber cable contains glass fibers rather than copper wire. Signals are transmittedacross these fibers in the form of light pulses rather than electrical pulsesas is the casewith metallic cables (twisted-pair and coax).Optical fiber strands are thin filaments of glass consisting of an inner core and an outercladding. The diameter of the core varies with the type of optical fiber. Single-mode opticalfiber has approximately an 8.5 m core diameter and a commonly used type of multimodeoptical fiber has a core diameter of 62.5 m. The cladding diameter for both is 125 m.Signals are transmitted as light pulses through the core of the optical fiber. When theselight pulses strike the cladding they are reflected back to the corebecause the glassused in the cladding has a lower refractive index than the core. This prevents thetransmission signal from being lost.

FIGURE 1.14: OPTICAL FIBER CABLE

Optical fiber has become an important consideration in the design of LANs.

Cladding

Core



Optical Fiber Advantages Optical signals through glass encounter less loss than electrical signals through

copper. This translates into lower attenuation and higher bandwidth than copper,allowing transmission to occur over longer distances.

Total immunity to electromagnetic interference since signals are sent as light andnot as electricity. Extremely high-speed LANs are possible, especially at shortdistances.

Optical Fiber Disadvantage For lower-speed LANs, more expensive electronics are required than with copper

systems.



Wireless systemsAlthough the majority of LANs connect devices using a physical cable, there are instanceswhere it is difficult or impossible to install cablesuch as historical properties or whenthere is no right-of-way access between two adjacent buildings. In such cases, wirelesstransmission media is used to connect network devices. Wireless systems do notphysically connect network devices since the links between the devices are invisible. Theyare either infrared light or radio links.

Infrared linksConnecting devices using infrared light signals work essentially the same way that remotecontrols work with television sets. These systems consist of a base unit connected to theserver and device connections for the stations. The base unit has two optical nodesoneto receive signals from the station and one to send signals to the station.Since the system depends on infrared light to transmit, a requirement is that the base unitand the station connections are in a direct line of sight to each other. Alternatively, some ofthese systems use a reflective surface positioned between the base unit and the station toredirect the signal.



Radio linksThis second type of wireless media uses radio waves to transmit information between theserver and the stations. Most of these systems use spread-spectrum technology wheredata is transmitted at low density over a frequency range of 902 MHz to 928 MHztherange specified by the U.S. government for data communications.These systems also work with two componentsa Control Module connected to theserver and User Modules which connect to the network devices. The control module andthe user module do not have to be in direct line-of-sight with each other. The radio signalsare capable of passing through most office building doors and walls.

Wireless Advantage Can be used in situations where it is difficult or impossible to install cable.

Wireless Disadvantages Typically, such systems are not able to meet the performance demands of large or

busy networks. Transmission can only occur over limited distances.


TopologiesThe definition of the word topology states than an items topology defines its physicalappearance. For example, a topological map represents the physical appearance of the areashown. In many ways a LANs topology is the sameit is representative of the LANsphysical appearance.LAN topology is determined by how transmission channels are used to connect networkdevices. Typically, it refers to how the LAN is physically set up and the cabling strategy beingused. It is acknowledged that topology is the foundation of a LAN.It should be pointed out that within the context of LANs, the word topology takes on a dualmeaning. Both aspects are important to how the LAN will function.1. First, topology refers to the physical appearance of the LAN. This is known as the

physical topology.2. The second aspect refers to how the LAN functions. This logical topology is

determined by how the messages are transmitted from device to device.There are many instances where a LAN has a certain physical appearance but logicallytransmits its messages in a different manner. For this reason, it is necessary to make thedistinction between the physical topology and the logical topology of a LAN.The purpose of this section is to illustrate the physical appearances a LAN may take. Thelogical aspect will be discussed in a later section.There are three fundamental topologiesstar, bus and ring. From these three, a number ofhybrid topologies have developed, including tree, star-wired ring, clustered star andhierarchical star.

Topologies, continued


Star topologyIn a star topology, the hub or switch is placed in the logical center of the network. Theremaining network devices are connected to this central hub like the points on a star.

FIGURE 1.15: STAR TOPOLOGY

Each device has its own direct, dedicated line to the hub or switch.Any network device wanting to send a message to anothernetwork device does so through the central hub.The station sending the message sends it to thehub. The hub then routes the message to thespecified destination stationthis is known asswitching. Since the hub handles all the messageswitching, the stations on the network do notrequire any extra technology to route signals overthe transmission channel.

Star Topology Advantages Faults are easier to locate and isolate.

Provides a central location for managing the network.

Star Topology Disadvantage It may be vulnerable to breakdown as the network is essentially controlled by one

device.

Hub



Bus topologyA bus topology is a linear configuration. It places all of the network devices on one lengthof cable, similar to stops on a city bus route. The hubs, server, stations and peripheraldevices all use the same continuous length of transmission channel.

FIGURE 1.16: BUS TOPOLOGY

The ends of the transmission channel, in thisarrangement, are not connected to networkdevices. Ordinarily, problems would occurwhen the transmitted signal is sent along thecable and it reaches either of the ends. Forthis reason, each end of the cable isconnected to a terminator which safely stopsthe transmission signal.When a message is sent on this topology,the following takes place. The transmissionsignal leaves the sending device and travelsalong the cable in both directions. Thedevice for which the message is intended willrecognize the transmission and read themessage as it passes by.



Bus Topology Advantages It is easily adaptable to many environmentsit can be configured to suit most

situations. It is easily expanded by adding devices at any point along the cable.

Bus Topology Disadvantages It lacks central controlfinding a fault is difficult. If the cable is damaged or if either end of the cable loses its termination, the entire

network will fail.



Ring topologyA ring topology places all of the network devices in a circle. It uses one transmissionchannel to connect all devices. Each device is connected to the next one. The last deviceis connected to the firstclosing the circle.

FIGURE 1.17: RING TOPOLOGY

When a message is sent, it travels fromdevice to device around the circle. Thesending device sends its message towardsthe destination device. Each devicebetween the sender and the receiverlistens to the message as it passes by. Ifthe message is not intended for aparticular device, it resends the messageand the next one in the ring repeats theprocedure. This continues until the intendeddestination receives the message.



Ring Topology Advantage There is no reliance on a central deviceall messages pass through all devices.

Ring Topology Disadvantages Additional network devices can only be connected while the network is inoperative,

since breaking the ring would cause network failure. If any device fails, the entire network is affected.

Although the ring topology is considered as one of the three fundamental networktopologies, it has never been popular in its basic form. The more popular dual-ringtopology provides two paths between stationsa primary path and a backup path. In theevent of a failure in the primary path, the signal can be diverted to the backup path bystations on either side of the point of failurepreventing total network failure.



Hybrid topologiesHybrid topologies resulted from a need to meet specific requirements or industrytechnological advancements. While there are many variations of the three topologiesdiscussed above, certain hybrids achieved greater popularity than others.

Tree topologyThe tree topology is anextension of the bustopology. By adding cableextensions to the basic bustopology, a larger networkcan be achieved using lesscable. Each additional cableextends from the underlyingbus structure and supportsmultiple network devicesalong its length.

FIGURE 1.18:TREE TOPOLOGY



Star-wired ring topologyA star-wired ring is also referred to as a collapsed ring. In this configuration, the networkdevices are connected to each other as they are in a ring topology. The difference is thatthey are connected through a central unit which acts as a wire center. The transmissionmethod is the same as with the ring topology except now dual rings are present, oneprimary and the other backup, and all messages must first pass through the wire centerbefore moving to the next device. Themain improvement over the ringtopology is that the failure of a singledevice will no longer cause the wholenetwork to fail, due to active monitoringby the wire center.

FIGURE 1.19:STAR-WIRED RING TOPOLOGY



Clustered star topologyA clustered star is much like the tree topology exceptthat there are clusters of devices at the end of eachbranch. There exists an underlying bus configurationwhich supports cable extensions. Each of these cableextensions has a cluster of network devices at itsend.

FIGURE 1.20: CLUSTERED STAR TOPOLOGY


Hub

HubHub


Hierarchical star topologyA hierarchical star topology is anextension of the star topology. In thisconfiguration, departmental networkdevices are connected to a hub orswitch as in a star topology. Thesehubs or switches are then connectedto each other via a central hub, alsofollowing a star configuration. This isthe recommended topology forstructured cabling systems inbuildings and in campusenvironments.

FIGURE 1.21:HIERARCHICAL STAR TOPOLOGY

Hub

Hub

Central Hub


Access controlsAccess control is the way in which a device on the LAN gains access to the transmissionmedia. Since there are many devices on a network, a method needs to be established for anindividual device to gain access to the cable. Only then can it transmit its message. Accesscontrol represents how the devices get permission to communicate on the network.There are five basic ways in which a network device obtains use of the transmissionchannelpolling, token-passing, slotted ring, CSMA/CD (Carrier-Sense Multiple Access/Collision Detection) and switching.Some access schemes are naturally suited for certain kinds of topologies. However, the useof any given access method is not necessarily governed by the LAN topology.

PollingAs the name implies, polling requires that each device on the network be asked if it has amessage to transmit. To ensure that each device is given an equal opportunity to speak,polling must be under central control. It is therefore, most commonly found on networkswith a central controlling device such as that found in a star topology.When polling is used, the device gains access to the transmission channels as follows:

The central controlling device checks with, or polls, each station regularly to see ifit has a message to send.

If the station has a message to send, and the transmission channel is clear, thestation receives exclusive use of the channel and sends its message.

As soon as the station has sent its message, the channel is free for another deviceto use.

Access controls, continued


Token-passingToken-passing is a method that uses an electronic signal called a token. Possession of thetoken gives a device exclusive use of the transmission channel. The token travels alongthe channel and stops at each device. A device with a message to send will pick up thetoken and use it in order to send its message.When token-passing is used, the device gains access to the transmission channel asfollows:

A network device with a message to send captures the available token as it passesby on the channel.

The message is attached to the token. The message-bearing token continues to circulate on the channel. As the token stops at a device, it is checked to see if the message is for the

devicethis destination device will recognize its address and will read themessage.

The destination device then attaches an acknowledgment of receipt to the tokenwhich continues to circulate.

When the sending device eventually receives the acknowledgment, it clears thetoken so it may be used by another device.

The token-passing scheme is most commonly used in ring or bus topologies.



Slotted-ringThe slotted-ring access scheme is used only with a ring topology. It was developed so thatmore than one device would be able to use the network at a time. Instead of a single tokencirculating, a number of fixed-length slots circulate around the telecommunicationschannel.The procedure followed is much like that used in token-passing. The difference is thatmessages are now divided into packets. Each packet is the size of one of the slots.When slotted-ring is used, the device gains access to the transmission channel as follows:

The sending device deposits a packet into each empty slot that passes until theentire message has been sent.

Each message packet has a header attached to it identifying the destination deviceand whether or not the packet completes the message.

The destination device recognizes from the header that the packet is meant for itand copies the message.

If indicated from the header, it will wait to receive the rest of the message fromadditional packets.

When the slot returns to the sending device, it is emptied and released to begin thecycle again.



CSMA/CD (Carrier-Sense Multiple Access/Collision Detection)The CSMA/CD access scheme allows all network users equal access to thetelecommunications channel. It uses neither a central controller nor tokens to controlnetwork access. It is a scheme used most often with a bus topology.The procedure for sending messages is twofold. The first part is referred to as carriersensing and the second is collision detection.

Carrier sensingCarrier sensing is the part of the operation where a device with a message to sendmonitors the transmission channel to see whether or not it is clear. It is checking to seewhether or not another device is transmitting a message.The role of carrier sensing is as follows:

When the sending device detects a clear channel, it transmits its messagemarking it with the destination address.

The destination device accepts the message and returns an acknowledgment ofreceipt to the sending device.

When the sending device receives the acknowledgment, the transmission iscomplete.



Collision detectionA collision occurs when two or more devices attempt to send a message at the same timeand the messages interfere with each other on the transmission channel.The role of collision detection is as follows:

The sending device will wait a specified amount of time for the acknowledgment. If the acknowledgment is not received in this period of time, the sender assumes

that the message was not received because a collision occurred in the transmissionchannel.

When the sending device detects a collision, it stops transmission. The sending device waits for a random amount of time and retransmits the

messagetypically, such an attempt can be repeated many times. If collisions are still detected after many attempts, the user is informed that the

network is too busy to transmit.Throughout this time, all idle network devicesthose not transmitting messagesarecontinuously monitoring the transmission channel to see if any messages are directed atthem or if the channel is available for transmission.



SwitchingWhile not strictly an access control scheme, switching provides a mechanism where astation does not have to share a transmission channel. Switching provides a dedicatedtransmission channel to each port of a switching hub. Each transmission channel can havemultiple stations attached to it, but in high traffic environments, each station can beassigned its own dedicated channel to the switching hub. The switching hub is responsiblefor providing communications between the channels.Traditionally, if a network is experiencing excessive trafficslow performancethenetwork is split into smaller segments, each with its own hub and with fewer attachedstations. A switching hub performs this type of segmentation inside a single chassis. It hasa number of ports, each of which is a dedicated LAN segment.When switching is used, stations access the transmission channel and communicate asfollows:

The sending station puts its data onto the transmission channel. The switching hub handles the connection to other stations. The switching hub handles intersegment traffic via an internal matrix switch. When a packet arrives at the switch, its destination address is noted and a

connection is made to the destination station.

The packet is then switched to the destination station. Subsequent packets are relayed through the switch automatically.


Transmission techniquesA transmission technique defines how the signal is actually transmitted over the channel. Itdetermines how the message travels on the medium.The various transmission mediatwisted-pair, coax, optical fiberare capable oftransmitting data at various frequencies. The range of frequencies they can use is called thebandwidth. A transmission technique describes how a device uses the available bandwidth totransmit signals to another device.The two transmission techniques used in the LAN environment are known as basebandtransmission and broadband transmission.

Transmission techniques,continued


Baseband transmissionBaseband transmission is the more commonly used method in the LAN environment. Thistechnique allocates the entire bandwidth to a single channel. Baseband transmission isbest suited for networks covering a small geographic area, such as a LAN.Baseband transmission handles only digital data and is capable of high-speedtransmission.

Baseband Transmission Advantages It is a less expensive technology to implement than broadband. It is relatively easy to design. It allows for easy reconfiguration and addition of stations.

Baseband Transmission Disadvantage There is only one pathway, which means only a single stream of data can exist on

the channel at any instant in time.

Transmission techniques,continued


Broadband transmissionBroadband transmission divides the bandwidth into multiple channels. Since there aremany channels available for transmission, more then one device can transmit at a time.Simultaneous transmissions can, therefore, occur without collisions.Broadband transmission can handle large amounts of data at one time. As well, it is notlimited to handling only digital data transmission. This method of transmission can supportanalog traffic, making it capable of handling traditional voice and video signals.Broadband transmission was designed for transmission over long distance with a channellength measured in kilometers or miles. A network using this technology can, therefore,cover a much larger geographic area than one using baseband technology.

Broadband Transmission Advantages It can handle the simultaneous transmission of many network devices. It can handle many types of signals over long distances.

Broadband Transmission Disadvantages It is costly to implement. It requires the mastery of complex equipment. The network is difficult to maintain

and reconfigure.


SpeedThe speed is a measure of how fast the signals travel over the transmission media.Speed is, in reality, not a totally separate component. It is actually dependent on the fourareastransmission media, topologies, access controls and transmission techniquesalready discussed.Typically, transmission speed ranges from a low of 235 Kbps (kilobits per second) for anAppleTalk LAN to 155 Mbps (megabits per second) for an ATM (Asynchronous TransferMode) LAN. Speeds in the Gbps (gigabits per second) are possible, but these are notcurrently used in LAN environments.


Basic network design

IntroductionWhen designing a LAN, many decisions need to be made. The previous section discussedthe five components of LAN architecturetransmission media, topologies, access controls,transmission techniques, and speed. While making wise selections in each of these areas iscritical, decisions about which transmission medium to choose or how to physically configurethe LAN is best made after deciding which underlying LAN structure best suits theenvironment.Essentially, it must first be decided what role each of the network devices will play in theLAN. Broadly defined, LANs come in two varietiespeer-to-peer and centralized server.Each of these environments has its own characteristics, advantages and disadvantages andwill be the more appropriate choice for a given situation.


Peer-to-peer LANsA peer-to-peer LAN is one in which any PC can contribute to or share network resources. Insuch a LAN environment, there is a great deal of freedom regarding the location of sharedfiles, programs and peripherals.As the name implies, all stations on this type of a LAN are peersmeaning they can work asequalscontributing and using files, programs, printers and other peripheral devices. Thereare cases where some stations in a peer-to-peer network will have more responsibilities thanothers.There are three configurations that a station on a peer-to-peer LAN can assumea no-serverconfiguration, a non-dedicated server configuration and a dedicated server configuration.

No-server configurationIn this configuration, a station on a peer-to-peer LAN is set up solely to access files,programs and peripheral devices found on other stations. The station itself does notcontribute any resources of its own to the network.

Non-dedicated server configurationIn this configuration, a station both accesses files, programs and devices on other stationsand at the same time allows those other stations to access some or all of its files andattached devices. A station configured in this manner is called a non-dedicated serveritis capable of servicing requests while at the same time allowing a user to work with itskeyboard, screen, files and attached devices.

Peer-to-peer LANs, continued


Dedicated server configurationIn this configuration, a station will do nothing but allow other stations to access its filesand attached devices. Such a machine will not permit a user to work with its keyboard andscreen since its only purpose is to respond to requests coming over the network from otherstations. A station set up in this manner is called a dedicated serverit is dedicated todoing nothing else but serving requests.

On a peer-to-peer LAN, one will see the non-dedicated server type of configuration mostoften, with each station both accessing and contributing to network resources.

Design issuesPeer-to-peer LANs are easy to set up but difficult to administer due to the changing role ofeach station. On any given day, the user of a station may allow others access to a specificdirectory on their machine and the next day remove this privilege.If a station on a peer-to-peer network is being repaired, or if someone in the office decidesto shut it off, other users are affected.As well, backups become a major issue. In essence, everyone on a peer-to-peer networkmust make their own daily backups to avoid a group disaster if a hard disk fails.

Peer-to-peer LANs, continued


Another problem area is security. At best, security is difficult. If a peer-to-peer LAN isbeing considered, certain questions must be asked before a final decision is made.

Who has access to which files? How can the organization prevent certain data on a station from accidentally

becoming available to other users? What prevents someone from walking up to a station, copying its files onto

diskettes, and walking away? Who will have overall responsibility for security?

These issues illustrate the overall problem with peer-to-peer LANsthey are difficult toadminister. For this reason they are not nearly as common as centralized server LANs.Peer-to-peer LANs are found in environments where the PC users are very experienced.They must be able to handle the administrative and technical details imposed on each ofthem. Offices where software developers or writers for computer-related publications workare two such environments.


Centralized server LANsA centralized server LAN is defined as a LAN in which one or more PCs are serverstheycontribute files, programs and peripheral devices to the network for shared use. All otherPCs are stationsthey have access to the files, programs and peripheral devices on theserver(s). Please note:

The server(s) can be dedicated or non-dedicated. The stations can have their own hard disks and attached peripheral devices but

these will not be available to other stations on the networkonly those on theserver are sharable.

There are similarities between a centralized server LAN and a peer-to-peer LAN. A dedicated server corresponds to a dedicated server station on a peer-to-peer LAN. A non-dedicated server corresponds to a non-dedicated server station on a peer-to-

peer LAN. A station corresponds to a no-server station on a peer-to-peer LAN.

This forces the question to be askedWhat is the difference between a peer-to-peer LANand a centralized server LAN?

There is one major differenceon a peer-to-peer LAN, any user can change theconfiguration of their own station from one day to the next, simply by issuing theappropriate commands at the station.

For example, a user can change the status of their peer-to-peer station from a non-dedicated server station to a non-server station. This would deny other usersaccess to that stations files and peripheral devices.

Centralized server LANs,continued


Design issuesOn a centralized server LAN, once it has been established that a given PC will be a server(dedicated or non-dedicated), it may take several hours or days and some expertise toconfigure and put the server into service on the LAN. Changes will require similar effort.There is a clear line which separates server and station. This makes administration amuch easier task. Servers are identified and stable. It is clear which PCs contain theshared files and therefore, need to be backed up.Similarly, security is simplified. The servers on a centralized server LAN can be physicallyisolatedplaced in a locked room with keys given only to the network administrator(s).Once physical access to the server is restricted, access to its files through the network canbe controlled by assigning passwords and directory/file access rights to individuals. Thisshould not be taken to mean that centralized server networks are automatically moresecure and easier to administer than peer-to-peer LANs. They can be, but this is entirelydependent on management. Poor or nonexistent use of the security and administrativeresources provided on a LAN will only invite abuse and disastersaccidental ordeliberate.



Security featuresTwo common security features on a LAN are login security and password security.

Login securityLogin security controls the users access to the network by installing the followingfeatures:

Users can only login at specific hours and days.This feature controls the days of the week and the hours of the day that a user canaccess the network.For example, it is possible to limit a users access to the network to Mondaythrough Friday from 8:30 A.M. to 6:30 P.M. This ensures that the user cannotaccess any files on the weekends or after working hours.

Users can only login at specific stations.All stations are assigned a number by the network operating system. This featurerestricts the physical station(s) where an individual user can work.For example, it is possible to restrict users to their own desktop stations. Thismeans that the users can only access the network from their own work areas.



Users can have a limited number of simultaneous logins.This feature controls the number of simultaneous active sessions a user can have.

For example, limiting the user to one simultaneous login specifies that the user canlog into the network once. If the user wants to log into the LAN again from anotherstation then they must first log out of the network from the first station.

Unauthorized user (intruder) detection/lockout option.This feature tracks the number of times an unauthorized individual tries to accessthe network. An unauthorized individual is anyone who tries to gain access to thenetwork with an incorrect password.After the threshold of unsuccessful attempts is exceeded, the intruders user nameis barred from accessing the network for a predetermined period of timefrom afew minutes to many days.

Automatic expiration date on accounts.This feature prevents a user from accessing the network after a specific date. Suchusers can be employees hired for temporary work or students enrolled in a courseneeding PC lab accounts for the semester.



Password securityPassword security controls access to the LAN by forcing users to provide passwordsbefore they can use network resources. These passwords can have the followingrestrictions placed on them.

Passwords must be a certain length.Setting a minimum length for a password (e.g., at least seven-characters long)makes it more difficult for someone to guess a users password.

Users are forced to change their passwords on a regular basis.Such a restriction requires a user to change their password after a certain period oftime (e.g., once a month). This makes it more difficult for an intruder to obtain andkeep using a users password.

Unique passwords are required (ones not used in the past).Each time a new password is required, it must be different from a set number ofprevious passwords. Typically, it is not possible for users to use one of their lastten passwords.

Limit the number of grace logins (logins allowed after the password hasexpired).A grace login allows a user access to the network after their password has expired.A message then warns the user to change the password. By limiting such gracelogins, the users are reminded to observe strict password maintenance.


Overview of LAN communications

IntroductionIn this discussion of LAN communications the objective is to examine how devices on a LANare able to communicate with each other. Specifically, three areas need to be covered:

Communications signaling The way in which signals are generated by the sending devices.

Communications addressing The way in which a signal finds its way to the correct destination.

Communications processing The path the signal takes once it arrives at its destination.


Communications signalingThe lowest-level function of LANs is the generation of digital dataall information processedby a PC is in a binary format consisting of a series of 0s and 1s. There are two generaltechniques for doing thisgenerate the data as a digital signal or as an a

chapter1.pdf

Documents

lan environment

bicsi lan design manual

awide area network

lana lan permits users

wide area networkchapter

peersany station

connected pcs

certain basic objectives