You are on page 1of 159

Syllabus for Advanced Networks

For M.Sc. (I.T.)


Chapter 1: Basics of Networking
Introduction Applications of a Network Analog and Digital Techniques Serial and Parallel Transmission Asynchronous and Synchronous Transmission.

Chapter 2: Internet & Wide Area Network


Internet asics Internet in India T!P"IP Telnet #orld #ide #e$ %yperte&t Transfer Protocol #e$ Ser'ers rowsers Search (ngines #e$ Ser'ers %T)* +senet ,irewalls - Intranets.

Chapter 3: Local Area Networks


*ocal Area Networks International .rgani/ation for Standardi/ation 0IS.1 IS. .SI 2eference )odel T!P"IP 2eference )odel T!P"IP 2eference )odel !haracteristics and uses of *AN *AN Protocols *AN Standards

Chapter 4:

AN ! etropolitan Area Network"


0Distri$uted 3ueue Dual

)etropolitan Area Network 0)AN1 D3D

us1 Structure Data Transfer I D3D Asynchronous Transfer )ode PSTN Structure !ircuit - Packet Switching Asynchronous - Synchronous Transfer

)ode Introduction to AT) enefits of AT) AT) Technology !lasses of ser'ices in AT).

Chapter #: $thernet
(thernet (thernet *imitations 45 )$ps Switched (thernet 455 )$ps (thernet Arcnet *AN I ) Token ring *AN ,DDI

Chapter %: Integrated &er'ices (igital Network


Introduction to ISDN Types of ISDN ,unctions of ISDN ISDN Standards ,uture Applications of ISDN

Chapter ): &torage Area Networks


Introduction enefits )anagea$ility .pen Standard Platforms (ase of Integration Ad'anced Application !apa$ilities Ad'anced Storage )anagement.

Chapter *: C+rrent ,rends in Co-p+ter Networks Bl+e ,ooth: Introduction . %istory System !hallenges Security The
asic Structure lue Tooth for (m$edded Internet Need for lue Tooth.

WA/: Introduction *imitations #AP forum De'ices used in #AP


.S !ompati$le with #AP #AP protocols Introduction to #)* Security in #AP #AP - Internet Introduction to (6mail Push Technology Push ,rame #ork ,uture in #AP.

WLAN: Introduction Application )a7or


Technologies Need to Deploy #*AN.

enefits #*AN

!hapter 4 ASI!S ., N(T#.29IN: Introduction Application of a Network Analog - Digital Signals Serial - Parallel Transmission Asynchronous - Synchronous Transmission. andwidth

Network:
A computer network is a collection of de'ices that can store and manipulate electronic data and is interconnected in such a way that network users can store< retrie'e and share information. A network may $e 'ast< comprising of hundreds of computers spread across continents= it may link together mainframes minicomputers and micros< printers< fa& machines and pagers= its users may $e host of indi'idual enthusiasts or firms= or the network may consist of not more than two machines connected with the sole purpose of sharing a printer or hard disk. In the near future< numerous other types of de'ices will $e network connecta$le< including interacti'e T>s< 'ideophones< na'igational and en'ironmental control systems. The larger network systems are generally referred to as #ide Area Networks. Some are run $y single organi/ations< with perhaps the $iggest $eing the world6wide area network run $y I ) for its own use< linking its many research esta$lishments and sales organi/ations. In +9 many of the leading chain stores and supermarkets ha'e networks that span the whole country with e'ery store feeding data $ack to the central organi/ation. A computer network is a resource< which ena$les the $usinesses to gather< analy/e< organi/e and disseminate the information that is essential to their profita$ility. The rise of intranets and e&tranets is an indication of the crucial importance of computer networking to $usinesses. Intranets and e&tranets are pri'ate $usiness networks that are $ased on Internet technology.

,he I-portance of Co-p+ter Networks:


Information and communication are two of the most important strategic issues for the success of e'ery enterprise. Today e'ery organi/ation uses a su$stantial num$er of computers and communication tools< to communicate with other departments and participate in information retrie'al programs< effecti'e usage of information technology< computer networks are necessary. These networks are a kind 0one might call it paradigm1 of organi/ation of computer systems produced $y the need to merge computers and communications. At the same time they are the means to con'erge the two areas= the unnecessary distinction $etween tools to process and store information and tools to collect and transport information can disappear. !omputer networks can manage to put down the $arriers $etween information held on se'eral 0not only computer1 systems. .nly with the help of computer networks can a $orderless communication and information en'ironment $e $uilt. !omputer networks allow the user to access remote programs and remote data$ases either of the same organi/ation or from other enterprises or pu$lic sources. !omputer networks pro'ide communication possi$ilities faster than other facilities. ecause of these optimal information and communication possi$ilities< computer networks may increase the organi/ational learning rate< which many authors declare as the only fundamental ad'antage in competition.

What

akes +p A Network0

The most important components are< o$'iously< the computers. A design or engineering office may well ha'e a network composed largely of high6resolution graphic terminals to run their !AD software< with a smattering of P!@S for routine word processing and accounting.

/eripherals: %ard disk dri'es and tape streamers< printers and plotters<
modems and mice. #ith a network< usually fewer peripherals are needed than with the

B same num$er of separate computers< for each user will ha'e access to e'ery peripheral that is attached to the network. !a$les are needed to create the physical links $etween the computers. Special networking software or Netware is also essential. This pro'ides a means of identifying and addressing each component< and controls the flow of data around the system. #hen a file is sent to $e printed< it is the Netware that ensures that it reaches the right printer=

Applications of Co-p+ter Networks


a1 &haring of applications: This allows all network users to share the same application< sa'ing disk space< $ecause the application only needs to $e installed on one of the computers. 21 &haring of (ata2ases: Second aspect $eing multi6user access and modify to the same data$ase at the same time is definitely $etter than ha'ing the same data$ase in all the computers and periodically com$ine all the modifications together. it is particularly useful for companies like $anks and tra'el agencies. c1 &haring 3eso+rces: It allows each user to ha'e access to the peripheral de'ices like printers and scanners. It is certainly cheaper than each terminal ha'ing its own peripheral de'ice. d1 /ersonal Co--+nications: It allows users to communicate with each other< sending computer files to another user= 7ust $y clicking a $utton and it impro'es company@s efficiency. +sers can send messages quickly without any to mo'ement. e1 Cost $ffecti'e 3eso+rce &haring: y selecting the right mi& of printers and allowing each network user an appropriate access to them< one could ha'e enough printing power to take care of the

E needs of all users= one can ensure that< a network ena$les to share any networka$le equipment or software and reali/e the same $enefits that one can en7oy from sharing printer. .n a network< users can share modems= data storage de'ices< such as hard disks and !D62.) dri'es= data $ackup de'ices< such as tape dri'ers= (6mail systems= fa& machines= and all networka$le software. #hen you compare sharing these resources to purchasing them for each computer< the cost sa'ings can $e enormous.

Analog & (igital &ignals


Analog (ata: Data that is in the form of continuously 'aria$le physical quantities. Analog &ignaling: An analog signal is one that 'aries in a continuous manner such as 'oice or music. Analog ,rans-ission: Transmission of a continuously 'aria$le signal as opposed to a discretely 'aria$le signal. Physical quantities such as temperature are continuously 'aria$le and so are descri$ed as CanalogD. (igital (ata: Information represented $y a code consisting of a sequence of discrete elements. (igital (ata Network: A network specially designed for transmission of data< where'er possi$le in digital form< as distinct from analog networks such as telephone systems< on which data transmission is an e&ception.

F The purpose of computer networks is to ena$le users to manipulate data so that it can $e stored< retrie'ed and shared. To understand how a'aila$le technology ena$les us to do this< we need to define a few terms and understand some $asic concepts. !omputers in a network must CcommunicateD with each other to ha'e the desired $enefits of the network. These signals can $e either CanalogD or CdigitalD. Digital signals< on the other hand< are distincti'ely different. Digital signals ha'e 'ery few 'alues. (ach signal is unique from a pre'ious digital 'alue and unique from the one to come. In effect< a digital signal is a snapshot of a condition and does not represent continual mo'ement. The most o$'ious e&ample of digital data is that communication on6$oard a computer. Since a computer@s memory is simply a series of switches that can either $e on or off< digital data directly represents one of these two conditions. #e typically represent this on and off status with 4s and 5s where 4 represents an ConD $it and 5 represents CoffD. oard$and networks incorporate technology similar to that of ca$le tele'ision. Data< whether it is 'ideo< audio< or digital< is transmitted on the wire at certain frequencies. The typical medium is coa&ial ca$le. Digital technology is generally utili/ed e&clusi'ely for $ase$and networks. These networks de'ote the entire ca$le to network transmission. !omparing analog and digital signals< ad'antages lie on either end of the spectrum. Analog signals suffer far less from attenuation o'er long distances. This rather makes sense. Since digital data can only $e a 4 or 5.

H Digital de'ices are lot less sophisticated< meaning that they are fairly easy to manufacture and cost6effecti'e. Digital de'ices are more resilient to ()I and make more efficient use of the ca$ling $andwidths than analog systems do.

/arallel ,rans-ission:
Parallel Transmission is the technique that sends each $it simultaneously o'er a separate line. Normally parallel Transmission technique is used to send data a $yte 0F$its o'er eight lines1 at a time to a high speed printer or other locally attached peripherals.

&erial ,rans-ission:
The standard method of AS!II transmission where $its are sent< one at a time< in sequence. (ach E6$it AS!II character is preceded $y a start $it and ended with a parity $it and stop $it. A group of SNA networks connected in series $y gateways is called as Serial Network. Serial Transmission is a technique in which each $it of information is sent sequentially on a single channel< rather than simultaneously as in parallel transmission. Serial Transmission is the normal mode of data communications. Parallel Transmission is often used $etween computer and local peripheral de'ices.

(ata ,rans-ission:
.ne ma7or difficulty in data transmission is that of synchroni/ing the recei'er with the sender. Two approaches e&ist to sol'e the pro$lem of synchroni/ationG these are asynchronous transmission and synchronous transmission.

45

As4nchrono+s ,rans-ission:
In this approach< synchroni/ation is implemented at character le'el and each indi'idual character is transmitted along with the necessary control information to allow this to take place. The control information consists of additional $its added to each character< STA2T ITS which indicate that it is a$out to cease. +sually< the stop $its are of the same polarity as the !hannel idle state. The initial change in the state of polarity< from the idle state to the first $it< is known as the STA2T P+*S(. !learly< this ena$les the recei'er@s clock to $e synchroni/ed with the transmitter@s clock.

As4nchrono+s trans-ission has 'ario+s ad'antages5 the ad'antages 2eing:


4. .ne principal ad'antage is that each indi'idual character is complete in itself therefore if a character is corrupted during transmission< its successor and predecessor will $e unaffected. 8. Particularly suited for applications where the characters are generated at irregular inter'als e.g. data entry from the key $oard.

,he ad'antages of as4nchrono+s trans-ission s4ste- are:


4. Successful transmission ine'ita$ly depends on the recognition of the start $its clearly these can $e easily missed or occasionally spurious start $ut can $e generated $y line interference. 8. %igh proportions of the transmitted $its are unique for control purposes and thus carry no useful information. ;. As a result of the effects of distortion the speed of transmission is limited.

44

Asynchronous serial transmission is normally used for speeds of up to ;555 $its per second for simple< single character error detection.

&4nchrono+s ,rans-ission:
In this system the message is transmitted 'ia single channel. %owe'er< in this instance it is imperati'e to note that there is no control information associated with indi'idual characters. The characters are grouped together in $locks of some fi&ed si/e and each $lock transmitted is preceded $y one or more special synchroni/ation characters< which can $e recogni/ed $y the recei'er. AS!II pro'ides a control character.

,he ad'antages of as4nchrono+s trans-ission are:


4. The amount of central information which requires to $e transmitted is restricted to only a few characters at the start of each $lock. 8. The system is not as prone to distortion as asynchronous communication and can thus $e used at higher speeds.

&4nchrono+s trans-ission also s+ffers fro- a few detri-ental attri2+tes the4 are:
4. If an error does occur rather than 7ust a single character the whole $lock of data is lost. 8. The sender cannot transmit characters simply as they occur and consequently has to store them until it as $uilt up a $lock< thus the system is unsuita$le for applications where characters are generated at irregular inter'als.

48

Bandwidth
In the simplest sense< $andwidth refers to the amount of information that can $e transferred $etween computers. The $and width is the speed at which the physical connection can mo'e data< and it actually constrains we$ access or access across the network more than the speed of your computer. ,or e&ample a 4?.? k$ps modem can recei'e only a$out 4.8 4.? kilo$ytes of data per second< e'en if there is no other traffic on the network. andwidth descri$es the amount of data a network can transport in a certain period of time. In other words< $andwidth is a capacity for rate of transfer< usually e&pressed in $its per second. )any networks today are $ased on a technology called (thernet< which has a standard $andwidth of 45 )$ps. 45 mega $ites of data can mo'e through any gi'en spot on the network. And the new ,ast (thernet has transmission speeds of 455 )$ps. As technology continues to e'ol'e< e'en more ad'anced networks ha'e $een de'eloped that offer transmission rates greater than 4 :$ps.

6+estions
4. (&plain the $asis of networksI #hy networking is essentialI 8. #hat are the important applications of networkingI ;. (&plain analog - digital techniques in detailI ?. #hat do you mean $y serial and parallel transmission and e&plain them in detailI A. #hat is Asynchronous and Synchronous Transmission and discuss them in detailI

4;

!hapter 8 INT(2N(T - #ID( A2(A N(T#.29 0#AN1 Internet asics Internet in India Internet Protocols Telnet #orld #ide #e$ %yper Te&t Transfer Protocol #e$ Ser'ers rowsers

Search (ngines +senet ,ire #alls Intranets

4?

IN,$3N$,
Internet is network of computers that offer access to people and information. .'er B5 million people use Internet< and the num$er is e&pected to increase o'er 485 million within a few years. The kind of information freely a'aila$le from internet includes :o'ernment documents< scientific data< ho$$yist lists< $usiness and personal information< ad'ertising data$ases and much more. The kinds of communication that can $e a'ailed on the internet include the followingG 4. (&changing short social notes. 8. :etting the latest news around the world. ;. !onducting $usiness negotiations. ?. !olla$orating on scientific research. A. (&changing information with others who ha'e similar ho$$ies or interests. B. Transferring computer files.

Internet in India
In India Internet was started to ser'e the educational institutions to help in their research work. In 4HFB< IIT was linked up with Indian Institute of Science $y (2N(T< which later connected with foreign uni'ersities. .n 4Ath August 4HHB :o'ernment called >SN* and started its dial6up ser'ices as first Internet Ser'ice Pro'ider. Now< there are

4A more than 455 ISPs gi'en license to pro'ide Internet Ser'ice. Some of them are )TN*< Satyam !omputers *td.< #intech< etc.

Internet /rotocols
,C/7I/
Transfer !ontrol Protocol and Internet Protocol are two sets of rules that allow computers and networks to communicate effecti'ely. They regulate the flow of data and make sure that it reaches its destination safe and sound. T!P and IP goes hand6in6had to ena$le the safe deli'ery of data o'er a network< the data is split into a num$er of smaller packets. T!P"IP attaches a header to the data packet< which contains information like the address< its origin< length of the packet and so on. IP< on the other hand< works like a postal department and ensures that once the data packets reach the recei'er@s end< they are re6assem$led in the same sequence they were $roken up and are ready for the application they are meant for. IP works as routing agent falls under the network layer which has function of making decision for transmitting data across de'ice not connected to each other.

La4ers of ,ra'el
The two protocols T!P"IP are stacked o'er each other and occupy the network layer and the transport layer. These layers are a part of 'irtual model of networking called .P(N SJST() INT(2!.NN(!TI.N 0.SI1 model. The .SI model consists of Physical< Data link< Network< Transport< Session< Presentation and Application layers. The physical layer transmits data from one location to another and is made up of physical aspects of the network like ca$les and connectors. The data link layer ensures error6free transmission of data and consists of networking cards< modems< etc. The

4B function of the network layer is to make routing decisions for transmitting data across de'ices that are not connected to each other. As IP is a routing agent< it falls under this layer. The transport layer comes ne&t and its primary function is to ensure error free transmission of data. Transfer !ontrol Protocol or T!P falls under this layer. The remaining layers such as the Session< Presentation and Application layers from the application group< which synchroni/es links across programs and con'erts network data to user reada$le formats.

,C/
As Transport layer protocol< T!P accepts message information from the applications< and di'ides it into multiple segments< and encapsulates each segment into a datagram. (ach datagram is passed o'er to the network layer protocol 0IP1 for further transmission and routing. At the recei'er@s end< T!P reassem$les the data and distri$utes it to the concerned application program. T!P transmits data in the form of packets that comprise of a header and a data $lock. The header consists of information like the address of the packet< its origin< the length of the packet and more. The data $lock carries the payload< which is the te&t or pictures that we down load or $rowse off the Net.

Internet /rotocol
Internet protocol or IP works like postal department. It routes data packets to the address mentioned in the header and fragments them. These are then marked so that the fragmentation sequence is maintained and are reassem$led upon reaching their destination. The routing of data grams o'er a network can occur o'er different paths and the possi$ility of some data grams arri'ing out of sequence is not ruled out. In addition< as

4E data grams flow $etween 'arious networks< they also face physical limitations in terms of the amount of data that can $e transferred o'er a particular network. IP is also attached to a small header on the data packet< which pro'ides information a$out the handling of the datagram< identification of fragmented data grams and the like. The 'ersion field contains a ?< $it code that identifies the IP protocol used to create the datagram. The identification field pro'es the identity of a datagram. In case the datagram has $een further fragmented< the fragment offset field specifies the other offset of the datagram. The flag field contains information a$out the nature of fragmentation. It pro'ides information a$out the current fragment and also gi'es the total num$er of the fragments. The header also has a field called KTime to *i'e@ or TT* that defines the num$er of routers a data packet can encounter en route to its destination computer< there$y a'oiding chocked networks. IP operates on gateway machines that mo'e data from the department to the organi/ation< then to the region and finally across the world.

I/ A((3$&& & 83L


To connect from one machine to another machine on the Internet< we need to know its IP address< which is an identifier for a particular machine on a particular network. These are referred as IP num$ers or Internet addresses. The IP address is represented $y four decimal num$ers separated $y dots and is $asically di'ided into the host computer section.

Classification to Internets:

4F

!lass AG This comprises of 'ery large networks with millions of nodes. They ha'e their
IP addresses ranging from 4.5.5.5 to 48E.5.5.5. The first num$er $efore the dot defines the network with the remaining three sections assigned to hosts.

!lass

G These are smaller networks and can ha'e only a$out BA<555 nodes. Their IP

addresses range from 48F.5.5.5 to 4H4.5.5.5. The first two num$ers are allocated to the network and the remaining two num$ers for the hosts.

!lass !G These are much smaller networks< which support a ma&imum of 8A? nodes.
The IP addresses range from 4H8.5.5.5 to 88;.5.5.5. In this case< the first three num$ers denote the network and the last one denotes the host. To make the operation simpler< ordinary names are assigned to each address using the Domain Name System. (ach Domain Name corresponds to a numeric IP address. The Internet uses the IP address to identify the network and the node and send data to the same. ,or e&ampleG As you type in the address www.yahoo.com form your $rowser< the Internet actually connects you to the IP num$er 85?.E4.855.E? 0#hich is the domain name for the we$ site1.

8nifor- 3eso+rce Locaters !83L"


+2* is the way to represent site name on the #orld #ide #e$. +2*s are similar to postal addresses or telephone num$ers which are used to represent the destinations. )ost +2* consists of ; partsG 4. Ser'ice Name 8. %ost Name ;. 2equest

4H The most common ser'ice names you use in +2*s are ChttpD< CtogetherD< CftpD and CnewsD. These refer to #e$ ser'ers< :opher ser'ers< ,TP ser'ers and +senet news ser'ers< respecti'ely. A few +2*s do not ha'e a host name.

(o-ain Na-e &er'ers


Internet works on the num$ering system. These num$ers are called IP. #hen we connect to the net we ha'e seen a set of ? num$ers $eing dialed i.e. for each address on the Internet there is a unique set of these ;8 $it num$ers. Domain Name Ser'ers are the ser'ers< which maintain a distri$uted list of all domains against Internet Protocol address. (arlier to Domain Name Ser'ers there was a system of ha'ing a host ta$le maintained $y S2I6NI!. It was updated twice a week to include new sites. System would download the copy of this ta$le through ,TP. There are two types of ser'ers as $elowG 4. 2esol'er 8. DNS There are a num$er of ser'ers< which maintain the addresses of sites. #hen $rowser needs the address of any site< resol'er queries the nearest name ser'er< replies immediately if it knows the answer or it asks another ser'er. Thus e'ery ser'er has two roles to playG 4. As a ser'er for name ser'er. 8. Super ser'er to e&tend functionality. All we$ sites are arranged in E $ranches namely arpa< com< edu< net< go'< mil< org. ,ollowing this are 8;B country name a$$re'iations like C.inD for India. This helps to locate the site easily. The IP addresses of name ser'ers at each of the domain name tags are maintained $y 45 root ser'ers. #hen a DNS fields a query that it cannot answer 4. It sends a query to root ser'er 8. 2oot ser'er says it does not know $ut a machine at say 4HA.HA.8A4.45 might know

85 ;. DNS sends a query to the a$o'e machine ?. Ser'er at 4HA.HA.8A4.45 knows the answer A. DNS returns answer to your P!.

&pecial feat+res of (N&


!acheG Name ser'er caches all IP address for domain names that were requested
recently. So that if requested again it responds immediately.

*oad

alancingG *arge sites like www.msn.com can ha'e multiple addresses

for same domain name. Name ser'ers currently return all IP addresses lea'ing P! to choose at random. ut some name ser'ers will now e'aluate all addresses to find out he one with least load.

9,/
,ile Transfer Protocol< which is the standard system for mo'ing files on the Internet. #ith it< it is possi$le to send or recei'e files to and from a machine on the Internet. The machine to which we are connecting must ha'e ,TP ser'er and its address generally starts with Kftp@ code. ,TP requires that we $e directly linked to the machine in question. So we are unlikely to use ,TP to transfer your files unless we or our client has dedicated ,TP ser'er permanently connected to the Net. As ,TP computer knows the name of the Internet ser'ices we are calling from< we only need to type KusernameL@ followed $y return. .nce logged on< we can get access to pu$lic accessi$le software. %ere one can get all sorts of files inside them. Thus< numerous ,TP ser'ers all o'er the world allow the people any where on the Internet to log in and download whate'er files ha'e $een placed on the ,TP ser'er.

84

,$LN$,
Telnet is a way of connecting to another machine on the Internet< and using it as if it were our own. In most cases we need to log6in and details of how to do this may $e displayed after we ha'e connected to it. A useful telnet site< especially if we are una$le to access the full graphical splendor of the #e$ isG telnet.w;.org. This is a te&t $ased we$ $rowsing system. The telnet command is a user interface to a protocol called< not surprisingly< T(*N(T. oth computers in order for the telnet program< to work must use the T(*N(T protocol. +NIM computer on the Internet uses the T(*N(T protocol< so this is rarely an issue. Telnet has many uses on the InternetG 4. %undreds of li$rary catalogs are a'aila$le only through direct connection to the li$rary@s computers. +sing telnet ser'er long distance charges of dial on directly to those computers= some don@t e'en allow direct dialing. 8. If you ha'e accounts on more than one computer on the Internet< you can log into the one closest to you and use telnet to log into the others. ;. 2esearchers colla$orating across the country can log into a single computer to run 7oint e&periments.

,:$ W;3L( WI($ W$B


The #orld #ide #e$ is an architectural frame work for accessing linked documents spread out o'er thousands of machines all o'er the Internet. Its enormous popularity stems from the fact that it has a colorful graphical interface that is easy for

88 $eginners to use< and pro'ides an enormous wealth of information on almost< e'ery concei'a$le su$7ect. The #e$ $egan in 4HFH at !(2N< the (uropean center for nuclear research. !(2N has se'eral accelerators at which large team of scientists from the participating (uropean !ountries carry out research in particle physics. The #e$ grew out of the need to ha'e these large teams of internationally dispersed researchers colla$orate using a constantly changing collection of reports< $lueprints< drawings< photos< and other documents.

:4per ,e<t ,ransfer /rotocol !:,,/"


The %TTP is a method used to make hyper te&t documents reada$le on the #orld #ide #e$. #e$ ser'ers and clients speak to each other using %TTP< so end users don@t need to know anything a$out its intricacies. %TTP is a stateless protocol< meaning that the client and the ser'er programs speak to each other only once and that a connection is not retained. A we$ client program sends a single request to the we$ ser'er for information< and the we$ se'er responds with a single reply.

We2 &er'ers:
Is a software program that sits on your ser'er 0The physical machine that is designed to store and ser'e we$ pages1. Any ser'er configured to communicate using T!P"IP uses ports. Not serial or parallel ports like the ones on the $ack of your computer< $ut the ones that look different and ser'e the same purpose. (.g. Port F5 is the default we$ ser'er port and all hyper te&t transfer protocols.

8; As %TTP request comes to the ser'er it checks the appropriate permissions and then either transmits the page or if the permissions are not adequate< it sends an error message. The #e$ ser'er has certain restrictions to what it can process. #hen the ser'er recei'es a request for a page ha'ing em$edded scripts< it cannot process these $y itself. It therefore uses additional software that performs au&iliary processing called middleware< which is written in Perl< ! or !NN. )ost middleware re'ol'e around interaction with the data$ase.

,4pe of We2 &er'ers


9,/ &er'ers:
,TP or ,ile Ser'er Transfer Protocol ser'ers are Internet computers that use this protocol and pro'ide data to Internet users for downloading. ,irstly the user should log in to ,TP ser'er< where the access rights are deri'ed for different directories.

(N& &er'ers:
('ery computer in the Internet is pro'ided with a specific IP address consisting of four num$ers like 48;.48;.48;.48;. Since remem$ering the num$ers is difficult< the Internet users want to reach a specific computer on the ### can also use description. A DNS ser'er includes a data$ase of IP Address. )apping #e$ is referred $y Internet software to retrie'e the IP address.

=opher &er'ers:
In the past few years many new tools for searching for information on the Internet ha'e de'eloped. .ne of them is :opher Ser'er. It is a !lient"Ser'er system that teaches you to na'igate through the Internet.

8senet News &er'ers:

8? It is a system where messages a$out any su$7ect can $e posted and other people on the Internet can reply to them. This topic includes politics< science< religion etc. There are more than A555 acti'e new groups<

Browsers
rowser is a kind of program< which can understand the hyper te&t protocol and present it into te&tual or graphical 'iew. Some of the popular $rowsers are Internet (&plorer< Netscape Na'igator< .pera< )osiac etc.

&earch $ngines:
Searching on the net comprise the enormous and e&hausti'e task of connecting to each ser'er and finding the requisite information on it. Da'id ,ilo and Oerry *ang< studying electronics at Stanford +ni'ersity decided to do something a$out it. They decided to de'elop a uni'ersal data$ase using which one could find information in a quick and simple way. !onsequently the worlds first search engine JahooP was $orn in April 4HH?. Today there e&ist a 'ariety of search engines< all of them they would possi$ly co'er almost a ma7or part of the information on the we$ at any gi'en time. Search engines are composed of data$ases that comprise inde&ing schemes< a query processor and Kspiders@. Spiders are programs that are designed to look up we$ pages which are listed in e'ery data$ase< follow up on each and e'ery link and update their data$ases to reflect the updated information. The records in these data$ases consist of the +niform 2esource *ocater or more simply< the dares where the we$site or page is located< the title of the page and the keywords for that page along with a short summary of the site in a few lines.

:ow do &earch $ngines Work0

8A A Search engine continuously sends so6called Kspiders@< a special kind of program< which starts in a homepage of a ser'er and pursues all links stepwise. #ord indices are created from indi'idual pages and the data$ase us updated. 4. In some search engines< the operators make entries using forms. Depending on the system< the data is released only after editorial processing. 8. To search for data< the search criteria are entered in the form pro'ided $y the search engine. The query is forwarded to the data$ase. ;. The result displays a list with all pages hat correspond to the search criteria. At the same time< the entries are displayed as links. Jou can reach the corresponding pages with a click.

&earch site t4pes:


Search sites are $asically of two types< search directories and search engines. Search directories are lists of we$ sites organi/ed into categories and su$6categories. Search directories are created manually rather than $eing automated. Their co'erage is far less than that of search engines $ut comprise recommendations and re'iews of sites. Search engines are huge computer generated data$ases containing information on millions of we$ sites. They ha'e programs called spiders that automatically look up we$sites and update their data$ases. Alta'ista< %ot$ot< *ycosQ< InfoseekQ< (&citeQ and #e$crawlerQ are search engines 0Q these are hy$rid sites i.e. they are search engines as well as offering search directories1

We2 &er'ices:
The we$ ser'ice is that facility to pro'ide the user with or without charge with some limitations. %otmail was one of the first success stories on the Net. It promises you free we$6$ased e6mail account that you can access from anywhere in the world. #e can send and recei'e (6mail= through there is a strong limit of 8 ) . The popular ser'ices areG

8B (6mail Ser'ice ,a&ing Pager Ser'ice (lectronic !ard ,ree we$ pages www.hotmail.com www.tcp.in or www.tacstems.com"fa&.htm www.icq.com www.48;greetings.com www.geocities.com or www.&oom.com

:4per ,e<t

ark+p Lang+age !:, L"

It is language for descri$ing how documents are to $e formatted. The term C)arkupD comes from the old days when copy editors actually marked up documents to tell the printer6in those days< )arkup languages thus contain e&plicit commands for formatting. ,or e&ample< in %T)*< R S means start $oldface RT S means lea'e $old face mode. Documents written in a markup language can $e contrasted to documents produced with #JSI#J: 0#hat Jou See Is #hat Jou :et1 word processor< such as )S6#ord or )S6Perfect. These systems may store their files with hidden em$edded markup so they can reproduce them later< $ut not all of them work this way. $ecomes possi$le for any we$ $rowser to read and reformat any we$ page. y em$edding the markup commands within each %T)* file and standardi/ing them< it

Commonly used HTM Ta!s"


R%T)*S R"%T)*S RTITA*S R"TITA*S R .DJS R" .DJS RPS R"PS RS!2IPTS R"S!2IPTS eginning and end of we$ page. The te&t $etween these tags does not appear on the eginning and end of the paragraph. eginning and end of the paragraph. )akes the $eginning and end of a script section.

page $ut $ecomes the title of the page that appears in the $rowser title $ar.

The script itself is not displayed on the page only the result is dictated.

8&$N$,

8E

A newsgroup is a worldwide discussion forum on some specific topic. People interested in the su$7ect can Csu$scri$eD to the newsgroup. Su$scri$ers can use a special kind of user agent< a newsgroup< to read all the articles posted to the newsgroup. People can also post articles to the newsgroup. (ach article posted to a newsgroup is automatically deli'ered to all the su$scri$ers< where'er they may $e in the world. Deli'ery typically takes $etween a few seconds and a few hours< depending how far off the $eaten path the sender and recei'er are. Alt is to the official groups as a flea market is to a department store. It is a chaotic< unregulated mishmash of newsgroups on all topics< some of which are 'ery popular< and most of which are worldwide. The !omp groups were the original +S(N(T groups. !omputer scientists< computer professionals and computer ho$$yists populate these groups. (ach one features technical discussions on a topic related to computer hardware or software. The Sci and humanities groups are populated $y scientists< scholars< and amateurs with an interest in physics< chemistry $iology< Shakespeare< and so on. The news hierarchy is used to discuss and manage the news system itself. System administrators can get help here. The hierarchies co'ered so far ha'e a professional< somewhat academic tone. Soc< which has many newsgroups concerning< politics< gender< religion< 'arious national cultures and genealogy. Talk co'ers contro'ersial topics and is populated $y people who are strong on opinions< weak on facts. Air is a complete alternati'e tree that operates under its own rules.

8F

In nearly all cases< when the newsreader is started< it checks a file to see which newsgroups the user su$scri$ers to. It then typically displays a one6line summary or each as6yet6unread article in the first newsgroup and waits for the user to select one or more for reading. The selected articles are then displayed one at a time. News readers also allow users to su$scri$e and unsu$scri$e to newsgroups. !hanging a su$scription simply means editing the local file listing which newsgroups the user is su$scri$ed to. News readers also handle posting. The user composes an article and then gi'es a command or clicks on an icon to send the article on its way. #ithin a day< it will reach almost e'eryone in the world su$scri$ing to the newsgroup to which it was posted. The sociology of +S(N(T is unique< to put it mildly. Ne'er $efore has it $een possi$le for thousands of people who do not know each other to ha'e world wide discussions on a 'ast 'ariety of topics. A moderated newsgroup is one in which only one person< the moderator< can post articles to the newsgroup. All postings to a moderated newsgroup are automatically sent to the moderator< who posts the good ones and discards the $ad ones.

&:$LL ACC;8N, ///7&LI/ ACC;8N,


Shell account descri$es the authori/ation to access another computer at the operating system le'el. Shell accounts are useful to the user who needs data in te&tual format. In shell account< user has no direct IP6*ink 'ia S*IP"PPP.

///7&LI/
!omputer using the T!P"IP !ommunication protocol to another T!P"IP computer o'er a modem or a serial line< $oth computers must $e running on an additional protocol. This can either $e PPP 0point to point1 or S*IP 0special *ine IP1. oth protocols perform

8H the same task $ut they are not interopera$le 0i.e. $oth ends of the connection must $e running on either PPP or S*IP1. PPP was deri'ed in 4HH4 $y I(T, 0Internet (ngineering Task ,orce1

9irewalls
,or Networks integrated with the Internet< there is a need to ensure safety to our network. A study re'ealed that out of the 8A5<555 attacks on the Department@s computer systems< a$out BA percent succeed. To minimi/e such pro$lems< the companies need to add a fire wall $etween the network and the Internet. The firewall consists of hardware such as routers and host systems software. A firewall is $asically a data packet $etween trusted and un6trusted networks. Any kind of network that uses T!P"IP for data transmission depends on source address< and the port num$er. A firewall uses these addresses and port num$ers to control the flow of data packets $etween the trusted and un6trusted network.

Classification of 9ire Walls:


Packet filter< Application pro&y or Application gateway Packet Inspection ,irewall.

4.

Application filter firewallG It is the fastest and simplest of the three and is

also one of the earliest. +sually the 2outer 0hardware1 $ased< in this system a packet filter compares the header information source and destination address< and port num$er6of each incoming or outgoing packet against a ta$le of access control rules.

;5 8. Application pro&y firewallG Pro&y firewalls are $uilt on the principle that security can $e relia$le only if there is no direct connection $etween the trusted and un6trusted networks. An application firewall works $y e&amining what application or ser'ice 0such as e6mail or file transfer1 a data packet is directed to. If the ser'ice is a'aila$le to that packet< then it is allowed to pass through. ;. Packet inspection firewallG The content of the packets is also considered. This inspection of packet can $e either $ased on its Kstate@ or Ksession@. In !ase of state filtering< the firewall only allows the incoming packet if it can $e matched with an out$ound request 0or@ in'itation1 for that packet. In case of session filtering< the network station in tracked. .nce the trusted user terminates the session< all incoming packets with identity pertaining to that session are re7ected.

3eal like firewalls:


There are two types in which a firewall can $e set upG Dual homed gatewayG %ere there is only one firewall with two connections< one for trusted network another for un6trusted network. Demilitari/ed /oneG %ere two firewalls are used. The first firewall has one connection leading to un6trusted network and second leading to host systems that can $e accessed through untrusted network. The area $etween the firewalls is called demilitari/ed /one.

IN,3AN$,:

;4 It is a network connecting as an affiliated set of clients using standard internet protocols< especially T!P"IP and %TTP. It is also defined as an IP6$ased network of nodes $ehind a firewall< or $ehind se'eral firewalls connected $y secure< possi$ly 'irtual< networks.

Ad'antages of Intranet:
&trea-lining 2+siness process:
Intranets are phenomenally powerful tools to streamline $usiness process. ,rom decision support< customer ser'ice and product engineering to distri$uted channel operations< from sales force automation and e&ecuti'e information systems< $usiness applications $ased on intranets can su$stantially impro'e the efficiency of comple& operations. This is possi$le $ecause intranet applications are typically much less e&pensi'e to de'elop and deploy< and much easier to use than applications $ased on older proprietary platforms. enefiting from a uni'ersal client interface the #e$ $rowser6intranet $usiness applications can $e deployed and managed from a central location. At the same time< standard $ased protocols and de'elopment technologies ena$le separate departments across a company to create intranet solutions that remain compati$le and compliant with company wide systems and process.

,acilitating Information DisseminationG


A key $enefit of the intranet technology is its a$ility to pro'ide up6to6date information quickly and cost6effecti'ely to the entire user community. An intranet puts 'ital information at the fingertips of employees< regardless of their location or the location of the information. Information disseminated on an intranet ena$les a high degree of coherence for the entire company $ecause communications are consistent. A @news@ section of an intranet< for e&ample< can include recent company press releases regarding management strategies< partnerships< and new products. A

;8 finance section can keep employees informed of 'ital financial reports and forecasts. A customer section can allow customers to check the status of an order or repair. >endors can su$mit in'oices online and check procurement status. y gi'ing people the a$ility to access time6critical information< intranets impro'e the decision6making process $y empowering indi'iduals with the knowledge necessary for faster and $etter informed $usiness decisions. Intranets allow the centrali/ation of information< which makes it easier to maintain and keep data up to data. The $enefit to the end6user is the simplicity and speed of information access. ,or e&ample< the interacti'e capa$ility ena$led $y hyper6te&t links makes it easy for users to gather all the information they need from #e$ pages quickly< 7ust $y clicking on a related icon or $utton. Pro'iding instant and secure access to $usiness6critical information sa'es time and increases producti'ity< and pu$lishing information online eliminates the production< duplication and distri$ution costs associated with paper.

$nriching Co--+nications and Colla2oration:


Intranet technologies ena$le teams to share knowledge and information regardless of their locations or time /ones. (ngineering groups can share research data< design concepts< schedules and other pro7ect materials for comments and re'iews during a de'elopment process. Training groups can distri$ute training schedules and multimedia computer6associated training courses using #e$6$ased technologies. Pro7ect terms can take ad'antage of intranet newsgroups and threaded discussion to communicate issues and solutions< and can use online chat technology when real6time interaction is required. #ith intranet teleconferencing< participants can share conference materials in a 'ariety of formats< including te&t< graphics< audio and 'ideo.

3uestions
4. (&plain the $asic principles of InternetsI 8. :i'e a $rief o'er'iew of Indian Internet ScenarioI ;. (&plain the concepts of T!P"IPI

;; ?. #hat are Telnet - #orld #ide #e$I riefly e&plain themI A. ring out the highlights of %yperte&t Transfer Protocol< #e$ Ser'ers and rowsersI B. #hat are Search (ngines - #hy it is importantI E. #hat are ,irewalls and IntranetsI (&plain themI

!hapter ; *.!A* A2(A N(T#.29 Introduction International .rgani/ation for Standardi/ation T!P"IP 2eference model The Network .perating System !lient Ser'er Network .perating System !lassification of *AN

;?

Local Area Network


*AN is an interconnection of computers and peripheral de'ices within a limited geographical area utili/ing a communication link and operating under some from of standard control. *AN is a computer network confined to a $uilding or a cluster of $uildings= it is typically personal to an organi/ation and is installed for the e&clusi'e use of an office or factory of a gi'en organi/ation.

International ;rgani>ation for &tandardi>ation !I&;":


IS. is made up of o'er 4B5 technical committees with o'er 8<;55 su$ committees across the glo$e. )ost of these committees work with national standards organi/ations from se'eral countries. All told< there are o'er EA of these national groups. IS. has pro'ided to networking the .SI model 0.pen Systems Interconnection1. It is $asically contains details all and the functions of networking and pro'ides a framework in which all 'endors around the world can create systems that can communicate with one another. Is a standard attempts to define the structure of a network as a E layer hierarchy each of which has a well defined function. The main aim of .SI standard is to define the way that a network node should look from the outside< i.e. from other network nodes. This ena$les the interconnection of networks< which differ in terms of the implementation in internal organi/ation and operation. A $rief description of se'en layers of the .SI model is gi'en $elow. 4. /h4sical La4er: Is the le'el at which the interchange of electrical signals< which represents data and control information takes place. This includes a

;A specification of electrical and mechanical characteristics of the physical connection. 8. (ata Link La4er: Takes the $are $it6le'el communication system pro'ided $y the physical layer and superimposes onto this a means for transmitting data and control information. The protocol used may $e character oriented< where control characters are used to delimit the 'arious fields of the $asic transmission $lock< or may relay upon positional significance. Acknowledgement of receipt of data and error control is $oth implemented at this le'el with the facility of retransmission if necessary. ;. Network La4er: Takes the packet si/e data $locks< which are handed down from the transport layer and attaches to these the address and routing information< which completes the packet. The choice of routing algorithm is ar$itrary and so routing can $e fi&ed or adapti'e< in which case packets are routed according to current network traffic loads. ?. ,ransport La4er: Pro'ides a relia$le data transmission and reception ser'ice for the session layer. The data is transmitted in the most efficient way that is suita$le for the needs of the session layer. This may $e an error free 'irtual connection with acknowledgements on a per packet $asis for secure data e&change. It could also $e a transmission ser'ice with no guarantee of deli'ery< which may $e suita$le for certain< types of traffic< digital 'oice for instance. The transport layer takes data from the session layer and splits it up in to pies< the si/e of packet data field. A. &ession La4er: Pro'ides a ser'ice to esta$lish< to maintain and terminate a connection with a process of a remote host computer. This layer should pro'ide a relia$le ser'ice to the presentation layer and ha'e the a$ility to reesta$lish a connection< should one of the lower layers in the hierarchy fail. Session layer should $e a$le to negotiate with the remote machine o'er

;B certain connection< parameters. These may include the type of

communication to $e employed< how the integrity of session connection is to $e controlled. B. /resentation La4er: Pro'ides a set of ser'ices to the application layer< which can $e used to process the data e&changed across the session connection. E. Application La4er: Is the highest layer in the network hierarchy. This layer protocol interacts directly with the application software wanting to transfer data across the network. All the other layers in the hierarchy e&ist for the sole purpose of satisfying.

;E

,C/7I/ 3eference

odel:
Network Access Internet Transport Process

+nlike IS. model this model uses ? layers. These are

La4er 4: Is the highest layer of T!P"IP concerned with the application process the user
requires.

La4er 3: CTransportD layer uses the Transmission !ontrol Protocol 0T!P1 to pass the
message from the user process to the internet 0IP1 layer The transport layer is where a long message is su$di'ided in smaller CpacketsD in preparation for east in CtransportingD. These packets are properly called datagrams. At the other end< this layer reassem$les the CPacketsD it recei'es into their correct order and puts the original message $ack together for the application to use.

La4er 2: IP< the Internet Protocol is responsi$le for routing indi'idual datagrams across
the interconnected networks.

La4er 1: Network Access: The $ottom layer< here is where the data link to the
physical media is prepared according to the desired type of connection. The completed message is often referred to as a frame The CtrailerD is added at this point.

;F Is finally con'erted into an electromagnetic signal $y special D!( hardware and placed on the physical medium.

,he ad'antage of I&; -odel o'er ,C/7I/ -odel can 2e stated as follows:
)ore carefully thought out and more CmodernD %as se'en layers< as compared to the four used in T!P"IP +sed more as a Creference modelD is the standard $y which others are often compared.

Characteristics & 8ses of LAN


A *AN typifies a distri$uted en'ironment and finds applications in a num$er of areas. Some e&amples areG 4. .ffice automation 8. ,actory automation

;H ;. Distri$uted !omputing ?. ,ire and Security Systems A. Process !ontrol B. Document Distri$ution.

,he characteristics of the ideal LAN can 2e s+--ari>ed as follows:


:igh &peed: Data rates of currently a'aila$le *ANs co'er a wide range. The slowest
transfer data at around 455 k$ps while the fastest ha'e data rates of up to 455 k$ps.

Low Cost: )any applications of *ANs in'ol'e low cost microprocessors systems= it is
desira$le that connection of such systems to a *AN should $e economic. Another factor that influences the cost of a *AN is the wiring< which must $e installed. There are $oth the costs of the wirer and its installation to consider. *ANs use 'ery ine&pensi'e ca$le such as twisted pair telephone wire.

:igh relia2ilit47Integrit4: Since *AN is s set of multiple interconnected systems= it


offers a good $ackup capa$ility in the e'ent of one or two systems failing in the network. This enhances the relia$ility and a'aila$ility of the systems to users.

Installation fle<i2ilit4: *AN offers fle&i$ility in locating the equipment. )ost


computers on a *AN are physically placed at the user ta$le< which is most con'enient for working and impro'es producti'ity significantly.

$<panda2ilit4: +nlike a large centrali/ed system< a *AN may e'ol'e with time. It
may $e put into operation with a small in'estment< and more systems.

?5

$ast of Access: The connection pattern of a *AN is normally a simple topological


form such as a ring or a tree and this has implications for the routing of packets on a *AN.

,he other ad'antages of LAN are as follows:


4. *AN pro'ides a resource6sharing en'ironment. All the *AN users may share e&pensi'e peripherals< hosts and data$ases. 8. *AN adhering to a certain standard< permits multi6'endor systems to $e connected to it. ;. In *AN< the systems are generally so chosen as to meet most of the user requirements locally and the network is used only for resource and information sharing purposes.

Co-ponents of LAN
Workstations: In *AN< a workstation refers to a machine that will allow users
access to a *AN and its resources while pro'iding intelligence on $oard allowing local e&ecution of applications. It may allow data to $e stored locally or remotely on a file ser'er. .$'iously< diskless workstations require all data to $e stored remotely< including that data necessary for the diskless machine to $oot up. (&ecuta$le files may reside locally or remotely as well< meaning a workstation can run its own programs or those copied off the *AN.

&er'ers: A ser'er is a computer that pro'ides the data< software and hardware
resources that are shared on the *AN. A *AN can ha'e more than one ser'er= each has its unique name on the network and all *AN users identify the ser'er $y its name.

?4 4. (edicated &er'er: A ser'er that functions only as a storage area for data and software and allows access to hardware resources is called a dedicated ser'er. Dedicated ser'ers need to $e powerful computers. 8. Non?(edicated &er'er: In many *ANs< the ser'er is 7ust another work station. Thus< there is a user networking on the computer and using it as a workstation< $ut part of the computer also dou$les up as a ser'er. Such a ser'er is called a non6dedicated ser'er. Since< it is not completely dedicated to ser'ing. *ANs do not require a dedicated ser'er since resource sharing amongst a few workstations is proportionately on a smaller scale. ;. ;ther ,4pes of &er'ers: In large installations< which ha'e hundreds of workstations sharing resource< a single computer is often not sufficient to function as a ser'er.

&o-e of the other ser'ers ha'e 2een disc+ssed here +nder:


4. 9ile &er'er: A file ser'er stores files that workstations can access and it also decides on the rights and restrictions that the users need to ha'e while accessing files on *AN. 8. /rinter &er'er: A Printer ser'er takes care of the printing requirement of a num$er of workstations. ;.

odern &er'er: It allows *AN users to use the modern to transmit long
distance messages. Ser'er attached to one or two modems would ser'e the purpose.

?8

CLI$N,&
A client is any machine that requires something from a ser'er. In the more common definition of a client< the ser'er supplies files and sometimes processing power to the smaller machines connected to it. (ach machine is a client. Thus a typical ten P! local area network may ha'e one large ser'er with all the ma7or files and data$ases on it and all the other machines connected as clients. This type of terminology is common with T!P"IP networks< where no single machine is necessarily the central repository.

N;($&:
Small networks that comprise of a ser'er and a num$er of P!. (ach P! on the network is called a node. A node essentially means any de'ice that is attached to the network. ecause each machine has a unique name or num$er 0so the rest of the network can identify it1< you will hear the term node name or node num$er quite often.

N$,W;3@ IN,$39AC$ CA3(&

?; The Network Interface card< or *AN adapter< functions as an interface $etween the computer and the network ca$ling< so it must ser'e two masters. Inside the computer< it controls the flow of data to and from the 2andom6Access )emory 02A)1. .utside the computer< it controls the flow of data in and out of the network ca$le system. An interface card has a speciali/ed port that matches the electrical signaling standards used on the ca$le and the specific type of ca$le connector. .ne must select a network interface card that matches your computer@s data $us and the network ca$le. Token ring *ANs require token ring NI!s< (thernet *ANs require (thernet NI!s< etc. The peripheral component interface $us has emerged as a new standard for adapter card interfaces. It is ad'isa$le to $ut P!I6equipped computers and using P!I *AN adapters where'er possi$le. Software is required to interface $etween a particular NI! and an operating system.

??

Connectors:
!onnectors used with TP included 2O644 and 2O6?A modular connectors in current use $y phone companies. .ccasionally other special connectors< such as I )@s Data !onnector< are used. 2O644 connectors accommodate ? wires or 8 twisted pairs< while 2O6?A houses F wires or ? twisted pairs.

,he Network ;perating &4steThe Network .perating System software acts as the command center< ena$ling all of the network hardware and all other network software to function together as one cohesi'e< organi/ed system. In other words< the network operating system is the 'ery heart of the network.

Client &er'er Network ;perating &4ste-s:


.n a !lient Ser'er Network< the network operating system is installed and runs on a computer called the network ser'er. The ser'er must $e a specific type of computer. A client6ser'er operating system is responsi$le for coordinating the use of all resources and ser'ices a'aila$le from the ser'er on which it is running. The client part of a client6ser'er network is any other network de'ice or process that makes requests to use ser'er resources and ser'ices. To log in< a user enters a log in command and gi'es his user name and password. If the user name and password are 'alid< the ser'er CauthenticatesD the user and allows him access to all network ser'ices and resources to which he has $een granted rights. The .S manages 'arious ser'er resources< which include hardware such as hard disks< 2A)< printers and equipment used for remote communications< such as modems. The network file system is also a ser'er resource. The network operating system pro'ides many ser'ices< including coordinating file access and file sharing< managing ser'er memory< managing data security< scheduling tasks for processing coordinating printer

?A access< and managing inter network communications. The most important functions performed $y a client ser'er operating system are ensuring the relia$ility of data stored on the ser'er and managing ser'er security.

Ad'antages of a client7ser'er network:


4. !entrali/ed 2esources and data security are controlled through the ser'er. 8. Scala$ility Any or all elements can $e replaced indi'idually as needs increase. ;. ,le&i$ility New technology can $e easily integrated into system. ?. Interopera$ility All components 0client"network"ser'er1 work together. A. Accessi$ility Ser'er can $e accessed remotely and across multiple platforms.

(isad'antages of a client7ser'er network:


4. (&pense 2equires initial in'estment in dedicated ser'er. 8. )aintenance *arge networks will require a staff to ensure efficient operation. ;. Dependence #hen ser'er goes down< operations will cease across the network.

/eer?to?/eer Network ;perating &4ste-s:


(na$le networked computers to function as $oth a ser'er and a workstation. In a peer6to6peer network< the operating system is installed on e'ery networked computer= this ena$les any networked computer to pro'ide resources and ser'ices to all other networked computers. Peer6to6peer operating systems ha'e $oth ad'antages and disad'antages when compared to client6ser'er operating systems. They pro'ide many of the same resources and ser'ices so do client ser'er operating systems< and under the right circumstances< can pro'ide good performance.

?B

Peer6to6peer networks pro'ide fewer ser'ices than client6ser'er operating systems. Also< the ser'ices they pro'ide are a great deal less ro$ust than those pro'ided $y mature< full6featured client6ser'er operating systems and the performance of peer6to6 peer networks commonly decreases significantly.

?E

Ad'antages of a peer?to?peer
4. (asy to connect a computer or peripheral to a linear $us. 8. 2equires less ca$le length than a star topology.

(isad'antages of a peer?to?peer network


4. Decentrali/ed No central repository for files and applications. 8. Security Does not pro'ide the security a'aila$le on a client"serer network.

Classification of LAN:
Network topologies: A network topology is the way the ca$ling is laid out. This
doesn@t mean the physical layout< $ut how the logical layout looks when 'iewed in a simplified diagram.

B+s ,opolog4: In this topology all de'ices share a common wire to transmit and
recei'e data. This approach is 'ery economical< as single ca$le is cheaper to purchase than se'eral indi'idual ca$les. Additionally< a single ca$le is easier to install than se'eral ca$les. These apparent ad'antages of the $us topology are offset< $y the difficulty in

?F trou$le shooting a pro$lem in this layout. Trou$le shooting in $us topologies may require a good pair of sneakers. .n the ends of the common ca$le< a de'ice a called a terminator is utili/ed to a$sor$ signals that ha'e tra'ersed the entire length of the $us. Since e'ery one shares the same ca$le no two machines can transmit at once or the $its of data from each will collide destroying $oth pieces of information. This e'ent is called a collision and o$'iously too many of them can $e disastrous to traffic flow on a network. A data reflection can occur any time an electronic signal encounters a short or an open. The end result is the same reflected data collides with the CgoodD data on the *AN and traffic flow is impacted.

Ad'antages of a Linear B+s ,opolog4


4. (asy to connect a computer or peripheral to a linear $us. 8. 2equires less ca$le length than a star topology.

(isad'antages of a Linear B+s ,opolog4


4. (ntire network shuts down if there is a $reak in the main ca$le. 8. Terminators are required at $oth ends of the $ack$oned ca$les. ;. It is difficult to identify the pro$lem if the entire network shuts down. ?. Not good as a stand6alone solution in a large $uilding.

&,A3 ,opolog4:
Star topology deri'es its name from the arrangement of de'ices so that they radiate from a central point. At the central point we usually see a de'ice generically called a hu$. 9ey $enefits of the star topology is the hu$ unit which may 'ary in function from a simple signal splitter to one that amplifies and keeps statistics on data tra'eling through them. Star topology a popular choice in the networking market place. %u$s that amplify signals coming through are called acti'e hu$s or multi6port repeaters.

?H Star topologies do require more ca$le than a simple $us topology< $ut most use a relati'ely ine&pensi'e type of ca$le called twisted pair ca$ling which helps control costs of wiring. The hu$s themsel'es require e&pense and the le'el of that e&pense is direct attri$uta$le to how comple& a hu$ is needed.

Trou$le shooting is $it easier than

us topology. At the 'ery least< one may

disconnect de'ices from a central hu$ to isolate a pro$lem as opposed to 'isiting each indi'idual machine. It@s o$'ious how the central hu$ de'ice offers ad'antages< $ut there is one draw$ack. The hu$ itself represents a single point of failure. If you lose a hu$< you effecti'ely lose all workstations attached to it.

Ad'antages of a &tar ,opolog4


4. (asy to install and wire. 8. No disruptions to the network when connecting or remo'ing de'ices. ;. (asy to detect faults and to remo'e parts.

(isad'antages of a &tat ,opolog4


4. 2equires more ca$le length than a linear topology.

A5 8. If the hu$ fails< nodes attached are disa$led. ;. )ore e&pensi'e than linear $us topologies $ecause of the cost of the concentrators.

3ing ,opolog4:
It descri$es the logical layout of token ring and ,DDI networks. In this a ring is created to which each de'ice is attached. A special signal called a token tra'el around this ring 'isiting each machine letting it know that it is that machine@s turn to transmit. Since the token 'isits e'ery node< e'ery one gets the chance to transmit< creating a 'ery CfairD *AN. The simplistic e&planation $elies the true comple&ity of ring topology systems a'aila$le today. Token ring *ANs< and their ,DDI cousins< are the most sophisticated fault6tolerant< and consequently< the most e&pensi'e systems a'aila$le in the current market place.

The logical creation of a ring allows information on such a *AN to tra'el in one direction. Since only one de'ice is allowed to transmit at a time< collisions are not a pro$lem on ring systems. Typical ring system NI!s contain the a$ility to perform what is known as signal regeneration< this means information recei'ed $y them is copied and retransmitted at a higher amplification. Since e'ery piece of data tra'eling around a ring

A4 must 'isit each de'ice. The signal gets regenerated numerous times. This feature allows for greater distances $etween nodes and increased chances that good data will completely tra'erse the ring.

,ree ,opolog4:
A tree topology com$ines characteristics of linear $us and star topologies. It consists of groups of star configured workstations connected to a linear $us $ack$one ca$le.

Ad'antages of a tree topolog4:


4. Point6to6point wiring for indi'idual segments. 8. Supported $y se'eral hardware and software 'endors.

(isad'antages of a tree topolog4


4. .'erall the type of ca$ling used limits length of each segment. 8. If the $ack$one line $reaks< the entire segment goes down. ;. )ore difficult to configure and wire than other topologies.

&,A3?WI3$( 3IN=1

A8 A star6wired ring topology may appear 0e&ternally1 to $e the same as a star topology. Internally< the multi6station access unit of a star6wired ring contains wiring that allows information to pass from one de'ice to another in a circle or ring.

C;N&I($3A,I;N& W:$N C:;;&IN= A ,;/;L;=A:


one41 A liner $us network may $e the least e&pensi'e way to install a network= Length of ca2le needed1 The linear $us network uses shorter lengths of ca$le. 9+t+re growth1 #ith a star topology< adding another concentrator easily does e&panding a network. Ca2le t4pe1 The most common ca$le is unshielded twisted pair< which is most often used with star topologies.

LAN Access Control


Collision &ense +ltiple Access 7 Collision (etection !C& A7C("1
In $us topology systems< all de'ices are attached to a common wire. As mentioned in a pre'ious section< this means that only one de'ice may use the common wire at a time. Since se'eral de'ices may need to use the wire at once< machines are said to $e contending for the media.

A;

(thernet systems use a channel access method known as !S)A"!D< short for !arrier Sense )ultiple Access " !ollision Detection. Though this seems a lot of words< the meaning is quite simple. !arrier Sense means that each de'ice checks the *AN $efore it starts transmitting to see if some other de'ice is using the media then. If another signal 0containing a CcarrierD1 was present< than the de'ice attempting to send would wait until the *AN is clear. Then it transmits its data. The collision detection part means that each workstation listens to make sure that only one signal is present on the *AN. In the e'ent there are two then o$'iously the data from one de'ice has collided with that of another. *ocal Talk *ANs used $y )acintosh P!s also use !S)A contention schemes< $ut these machines incorporate a technology called time6di'ision multiple&ing to allow a'oidance of collisions. In fact< *ocal Talk Systems are said to $e !S)A"!A systems< with !A standing for !ollision A'oidance. The ma7or ad'antage of contention systems is that de'ices may transmit whene'er they like 7ust as long as the *AN is free. !onsequently the o'er headed of de'ices waiting on the opportunity is generally low. %owe'er< as traffic increases in a contention system< collisions can $ecome e&cessi'e< impacting the o'erall performance of the network. The capacity of the *AN may $e far underutili/ed in this e'ent. The other ma7or disad'antage is that contention systems do not follow an easily predicta$le pattern of performance degradation as traffic increases.

,oken /assing &che-e


This technology is used for token ring systems. Its incorporation along with complementary fault6tolerance capa$ilities yields a *AN with a fair amount of sophistication< managea$ility and relia$ility. In this channel accesses a small signal called a token which regularly 'isits each de'ice. The token gi'es permission for the de'ice to transmit if it needs to. If transfer of

A? data is needed< the de'ice recei'es a set amount of time to $roadcast its data. #hen it is done< the machine then retransmits the token to another machine gi'ing that recipient permission to transmit< and so the system continues. This mechanism ensures opportunity for all de'ices to gain access to the *AN. ha'e enhanced access to the *AN if warranted. As traffic demand increases on a token *AN< the o'erall throughput of data rises as well< until a point is reached where the networks simply cannot accommodate anymore. The function in this case is somewhat like a waterwheel. The wheel itself recei'es water from a sluice. Jou may increase the capacity of the wheel< $ut the sluice can only hold so much water. Throughput characteristics of token *ANs are so predicta$le< $ecause of the characteristics of traffic demand. These systems are ideal for hea'y traffic situations. !omple&ity of such a *AN does come at some cost. Token systems require o'erhead to carry out their many functions including fault6tolerance. Token ring systems are additionally considera$ly more e&pensi'e than (thernet systems. ,actors weighing in deciding which system to choose should include traffic demand and $udgetary restraints. ecause of its predicta$le $eha'ior< token scheme *ANs offers the ad'antage of priorities< where a certain group of de'ices may

LAN CABLIN=
!a$le is the medium thorough which information usually mo'es from one network de'ice to another. There are se'eral types of ca$le< which are commonly used with *ANs. The type of ca$le chosen for a network is related to the network@s topology< protocol< and si/e. +nderstanding the characteristics of different types of ca$le and how they relate to other aspects of a network is necessary for the de'elopment of a successful network. The 'arious types of ca$les are as followsG

,WI&,$( /AI3 CABLIN=:

AA

Twisted pair ca$ling is the current popular fa'orite for new *AN installations. The marketplace popularity is primarily due to twisted pair@s 0TP@s1 low cost in proportion to its functionality. The construction of TP is simple. Two insulated wires are twisted around one another a set num$er of times within one foot of distance. If properly manufactured< the twists themsel'es fall in no consistent pattern. This is to help offset electrical distur$ances< which can affect TP ca$le such as radio frequency interference 02,I1 and electromagnetic interference 0()I1. These CpairsD of wires are then $undled together and coated to form a ca$le.

Twisted pair comes in two different 'arieties6 shielded and unshielded. Shielded twisted pair 0STP1 is often implemented with *ocal Talk $y Apple and $y I )@s token ring systems. STP is simply TP ca$ling with a foil or mesh wrap inside the outer coating. This Special layer is designed to help offset interference pro$lems. The shielding has to $e properly grounded< howe'er< or it may cause serious pro$lems for the *AN. TP ca$ling has $een around a while and is a tried and true medium. It hasn@t $een a$le to support high6speed data transmissions until relati'ely recently howe'er. New de'elopment is focusing on achie'ing 455 )$ps throughput on +TP without costing the user an arm and a leg. A copper 'ersion of fi$er optic@s ,DDI< called !DDI< will continue to mature while standardi/ation is worked out for 455 )$ps (thernet systems $y the mid H5s. !opper ca$le will not allow the speeds attaina$le with fi$er optic ca$le. %owe'er< the standard for fi$er stipulates *AN speeds of only 455 ) $ps< for $elow the fi$er optic ca$le@s actual capacity.

AB

Twisted pair is grouped into certain classifications $ased on quality and transmission characteristics. I ) calls the classifications CtypesD. +TP $y itself is often grouped $y CgradesD.

8N&:I$L($( ,WI&,$( /AI3 =3A($&


:rade 4 Suita$le for 'oice transmission and data transfer upto 4 )$ps. :rade 8 !apa$le of carrying data at ? )$ps. :rade ; !arries data at upto 45 )$ps. :rade ? 2ated at 85 )$ps. :rade A Support speeds at upto 455 )$ps.

,WI&,$( /AI3 CABL$ ,A/$& !IB


Type 4 shield. Type 8 Type ; Type A Type B

&,AN(A3(&"

STP< two pair< 88 gauge< solid conductors< and $raided6 Type 4 ca$le with additional four pairs of +TP. +TP< 88 or 8? gauge< 8 twists per foot< four pairs. ,i$er optic ca$le used to link )A+s. Two pair< stranded 0not solid1 8B gauge< patch ca$les.

Type F Two pair< 8B gauge< and untwisted $ut untwisted $ut shielded ca$le

,WI&,$( /AI3 CABL$: A#$ANTA%&S


4. Ine&pensi'e 8. .ften a'aila$le in e&isting phone system ;. #ell tested and easy to get.

#ISA#$ANTA%&S
4. Suscepti$le to 2,I and ()I 8. Not as dura$le as coa&. ;. Doesn@t support as high a speed as other media.

C;ABIAL CABL$:

AE !oa&ial ca$le or 7ust Ccoa&D en7oys a huge installed $ase among *AN sites in the +S. It has fit the $ill perfectly for applications requiring sta$le transmission characteristics o'er fairly long distances. It has $een used in A2! net systems< (thernet systems and is sometimes used to connect one hu$ de'ice to another in order systems. !onstruction6wise coa& is little more comple& than TP. It is typically composed of a copper conductor that ser'es as the CcoreD of the ca$le. This conductor is co'ered $y a piece of insulating plastic< which is co'ered $y a wire mesh ser'ing as $oth a shield and second conductor. P>! or other coating then coats this second conductor. The conductor within a conductor sharing a single a&is is how the name of the ca$le is deri'ed.

!oa&ial ca$le@s construction and components make it superior to twisted pair for carrying data. It can carry data farther and faster than TP can. These characteristics impro'e as the si/e of the coa& increases. There are se'eral different types of coa& used in the network world. (ach has its own 2: specification that go'erns si/e and impedance< the measure of a ca$le@s resistance to an alternating current. Different ca$le can differ widely in many important areas.

C;

;N C;ABIAL CABL$ ,A/$& 8&$( IN N$,W;3@IN=:

Type !ommon +sage Impedance 2:6F Thick (thernet A5 .hms 2:644 road$and *ANs EA .hms

AF 2:6AF Thin (thernet A5 .hms 2:6AH Tele'ision EA .hms 2:6B8 A2! net H; .hms !.AMIA* !A *(

A#$ANTA%&S
4. ,airly resistant to 2,I and ()I 8. Supports ,aster data rates than twisted pair ;. )ore dura$le than TP

#ISA#$ANTA%&S
4. !an $e effected $y strong interference 8. )ore costly than TP. ;. ulkier and more rigid than TP

9IB$3 ;/,IC&:
,i$er has come into importance on its own as the premier $ounded media for high6speed *AN use. ecause of fi$er@s formida$le e&pense< howe'er< we are not likely to see it at the local workstation any time real soon. ,i$er optic is unsophisticated in its structure< $ut e&pensi'e in its manufacture. The crucial element for fi$er is glass that makes up the core of the ca$ling. The glass fi$ers may $e only a few microns thick or $undled to produce something more si/a$le. It is worth noting that there are two kinds of fi$er optic ca$le commercially a'aila$le6single mode and multimode. AT use single mode in the telecommunications industry and T or +S sprint to carry huge 'olumes of 'oice data. )ultimode is what we use in the *AN world.

AH

The glass core of a fi$er optic ca$le is surrounded $y and $ound to a glass tu$e called CcladdingD. !ladding adds strength to the ca$le while disallowing any stray light wa'e from lea'ing the central core. A plastic then surrounds this cladding or P>! outer 7acket which pro'ides additional strength and protection for the inwards. Some fi$er optic ca$les incorporate 9e'lar fi$ers for added strength and dura$ility. 9e'lar is the stuff of which $ulletproof 'ests are made< so it@s tough. ,i$er optic is lightweight and is utili/ed often with *(Ds 0*ight (mitting Diodes1 and I*Ds 0In7ection *aser Diodes1. Since it contains no metal< it is not suscepti$le to pro$lems that copper wiring encounters like 2,I and ()I. Plus< fi$er optic is e&tremely difficult to tap< so security is not a real issue. The $iggest hindrance to fi$er is the cost. Special tools and skills are needed to work with fi$er. These tools are e&pensi'e and hired skills are e&pensi'e too. The ca$le itself is pricey< $ut demand will ease that $urden as more people in'est in this medium. Attempts ha'e $een made to ease the cost of fi$er. .ne solution was to create synthetic ca$les from plastic as opposed to glass. #hile this ca$le worked< it didn@t possess the near capa$ilities of glass fi$er optic< so its acceptance has $een somewhat limited. The plastic fi$er ca$les are constructed like glass fi$er only with a plastic core and cladding. The $andwidth or capacity of fi$er is enormous is comparison with copper ca$ling. )ultimode fi$er can carry data in e&cess of A giga$its per second. Single mode fi$er used in telecommunications has a theoretical top speed in e&cess of 8A<555 :$ps. That much data is the equi'alent of all the catalogued knowledge of man transmitted through a single small glass tu$e in less than 85 seconds. The standard go'erning implementation of fi$er optic in the marketplace is called the fiber distributed data interface standard or ,DDI. ,DDI specifies the speed of the *AN< the construction of the ca$le< and distance of transmission guidelines. ,DDI $eha'es 'ery much like token ring< only much faster. An added feature for ,DDI is a

B5 $ackup ring in case the main ring fails. This fault tolerance along with the fault tolerance already incorporated in token ring technology makes ,DDI *ANs pretty resilient.

LAN /rotocols:
A protocol is a set of rules that go'erns the communications $etween computers on a network. These rules include guidelines that regulate the following characteristics of a network= access method< allowed physically topologies< types of ca$ling< and speeds of data transfer.

$thernet:
The (thernet protocol is $y far the most widely used. It uses an access method called !S)A"!D 0!arrier Sense )ultiple Access " !ollision Detection1. This is a system where each computer listens to the ca$le $efore sending anything through the network. If the network is clear< the computer will transmit. If some other node is already transmitting on the ca$le< the computer will wait and try again when the line is clear< sometimes< two computers attempt to transmit at the same instant< when this happens a collision occurs. (ach computer then $acks off and waits a random amount of time $efore attempting to retransmit. #ith this access method< it is normal to ha'e commissions. This protocol allows for linear $us< star< or tree topologies. Data can $e transmitted o'er twisted pair< coa&ial or fi$er optic ca$le at a speed of 45 )$ps.

9ast $thernet: ,or an increased speed of transmission< the (thernet protocol


has de'eloped to new standard that supports 455 )$ps. This is commonly called ,ast (thernet. ,ast (thernet requires the use of different< more e&pensi'e network concentrators"hu$s and network interface cards.

B4

Local ,alk: *ocal talk is a network protocol that was de'eloped $y Apple
!omputer< Inc. for )acintosh !omputers. The method used $y local talk is !S)A"!S 0!arrier Sense )ultiple Access with !ollision A'oidance1. It is similar to !S)A"!D e&cept that a computer signals its intent to transmit $efore it actually does so. *ocal Talk adapters and special twisted pair ca$le can $e used to connect a series of computers through the serial port. The )acintosh operating system allows the esta$lishment of a peer6to6peer network without the need for additional software. The *ocal Talk protocol allows for linear $us< star< or tree topologies using twisted pair ca$le. A primary disad'antage of *ocal Talk is seed< its speed of transmission is only 8;5 9$ps.

,oken 3ing:
I ) de'eloped this protocol in the mid 4HF5s. The access method used in'ol'es token passing. In token ring the computers are connected so that the signal tra'els around the network from one computer to another in a logical ring. A single electronic token mo'es around the ring from one computer to ne&t. If a computer does not ha'e information to transmit< it simply passes the token on to the ne&t workstation. If a computer wishes to transmit and recei'es an empty token< it attaches data to the token. The token then proceeds around the ring until it comes to the computer for which the data is meant. At this point< the recei'ing computer captures the data. The token ring protocol requires a star wired ring using twisted pair or fi$er optic ca$le. It can operate at transmission speeds of ? )$ps or 4B )$ps.

9((I:
,i$er Distri$uted Data Interface 0,DDI1 is a network protocol< used primarily to interconnect two or more local area networks< often o'er large distance. The access method used $y ,DDI in'ol'es token passing. ,DDI uses a dual ring physical topology. Transmission normally occurs on one of the rings= howe'er< if a $reak occurs< the system keeps information mo'ing $y automatically using portions of the second ring to create a

B8 new complete ring. A ma7or ad'antage of ,DDI is speed. It operates o'er fi$er optic ca$le at 455 )$ps.

'rotocol
(thernet ,ast (thernet *ocal Talk Token 2ing ,DDI

Cable
Twisted !oa&ial< ,i$er Twisted Pair< ,i$er Twisted Pair Twisted Pair ,i$er

S(eed
Pair< 45 )$ps 455 )$ps 8; )$ps ? )$ps 4B )$ps 455 )$ps

To(olo!y
*inear us< Star< Tree Star *inear us or Star Star #ired 2ing Dual 2ing

LAN &tandards:
Institute of (lectrical and (lectronic (ngineers 0I(((1G The I((( has done nota$le work in the standards area of networking. This organi/ation is huge with o'er ;55<555 mem$ers cosists up of engineers< technicians< scientists< and students in related areas. The computer society of I((( alone has o'er 455<555 mem$ers. I((( is credited with ha'ing pro'ided definiti'e standards in *ocal Area Networking. These standards fall under a group of standards known as the F58 pro7ect. The F58 standards were the culmination of work performed $y the su$committee starting in 4HF5. The first pu$lished work was F58.4< which specified a framework for *AN@s and inter6networking. This was followed in 4HFA with specific *AN6oriented standards titled F58.86F58.A. )ost of the work performed $y the F58 pro7ect committee re'ol'es around the first two layers of the .SI model initiated $y the IS.. These layers in'ol'e the physical medium on which we mo'e data and the way that we interact with it. In order to $etter define these functions< the I((( split the Data *ink *ayer of the .SI model up into two separate components. F58 I((( committee responsi$le for setting standards concerning ca$ling< physical topologies< logical topologies and physical access methods for networking products. The !omputer Society of I(((@s F58 pro7ect committee is di'ided into se'eral su$6committees that deal with specific standards in these general areas. Specifically the Physical *ayer and the Data *ink *ayer of the IS.@s .SI model are addressed.

B; F58.4 This work defines an o'erall picture of *ANs and connecti'ity. F58.4 This set of standards specifically address the network management. F58.4D Standards for $ridges used to connect 'arious types of *ANs together were set up with F58.4D. F58.8 !alled the *ogical *ink !ontrol 0**!1 standards< this specification go'erns the communication of packets of information from one de'ice to another on a network. F58.; Defines the way data has access to a network for multiple topology systems using !S)A"!D. A prime e&ample is (thernet and Star *AN Systems. F58.? Standards de'eloped for a token passing scheme on a $us topology. The primary utili/ation of this specification was the )anufacturing Automation Protocol *ANs de'eloped $y :eneral )otors< operates at 45 )$ps. F58.A This standard defines token ring systems. It in'ol'es the token passing concept on a ring topology with twisted pair ca$ling. I )@s token ring system uses this specification= the speed is either ? )$ps or 4B )$ps. F58.B )etropolitan Area Networks are defined $y this group. )ANs are networks that are larger than *ANs typically falling within A5 9ilometers. They operate at speeds ranging from 4 )$ps up to a$out 855 )$ps. F58.E These are standards concerning $road$and *ANs. F58.F This group sets up standards for *ANs using fi$er optic ca$ling and access methods. F58.H This specification co'ers 'oice and digital data integration.

B?

F58.45 These mem$ers set standards for interopera$le security. F58.44 #ireless *ANs are the su$7ect of this particular su$committee@s works. oth infrared and radio *ANs are co'ered.

Ad'antages of LAN o'er

ini and

ainfra-e Co-p+ters

)ainframe computers or )ini computers ha'e a huge processing power. )any users are attached to the !P+ with the help of Kdum$ terminals@. Though< the processing power and num$er of people interacting with the computer is great economically< such computing power would $e 'ery e&pensi'e. The ad'antage of the )ainframe and )ini system are rather rigid gi'ing 'ery little room for the fle&i$ility in design and approach. *AN on the contrary is modular which can $e altered as per the user requirement. ,le&i$ility is another ad'antage of the networked P!s. The setup and operations of )ainframe and )ini systems are rather rigid gi'ing 'ery little room for the fle&i$ility in design and approach *AN on the contrary is modular which can $e altered as per the user requirement. Scala$ility is 'ery difficult and time and money consuming for )ainframes and )inis while *AN is ideally suita$le for this. Skilled and highly qualified engineers are required for the operations of )ainframe and )inis whiles users themsel'es can manage *ANs without any pro$lem. ('en the installation and commissioning is e&tremely easy for *ANs.

6+estions

BA 4. ring out the concept of *ocal Area NetworkI the techniques of .SI modelI ;. #hat is T!P"IP reference modelI ?. #hat are *AN Protocols - *AN standardsI (&plainI A. Discuss the !haracteristics - +sers of *ANI

8. (&plain the CInternational Standard .rgani/ation 0IS.1 functions and $ring out

!hapter ? )(T2.P.*ITAN A2(A N(T#.29 Introduction !ircuit Switching - Packet Switching Synchronous and Asynchronous Transfer )ode AT) !lasses of Ser'ices

BB

etropolitan Area Network


Is $asically a $igger 'ersion of *AN and uses similar technology. It co'ers a group of near $y corporate offices or a city and might $e either pri'ate or pu$lic. It can support $oth data and 'oice and might e'en $e related to the local ca$le tele'ision network. The technology aspect of )AN is that there is a $roadcast medium< to which all the computers are attached. ,or networks co'ering an entire city< I((( defined one )AN called D3D 0Distri$uted 3ueue Dual us1< as standard F58.B.

BE The $asic geometry F58.B is that two parallel uni6directional $usses make through the city with stations attached to $oth $usses in parallel. (ach $us has a head end< which generates a steady stream of A; $yte cells. (ach cell tra'els down stream from the head end. #hen it reaches the end it falls off the $us. Traffic that is destined for a computer to the right of the sender uses the upper $us. Traffic to the left uses the lower one.

(ach cell carries a ?? $yte payload field< and it also holds two protocol $its< $usy set to indicate that a cell is occupied< and request< which can $e set when a station wants to make a request. %ere stations queue up in the order till they $ecome ready to send and transmit in ,I,. order. The $asic rule is that stations are polite. This politeness is needed to pre'ent a situation in which the station nearest to the head end simply grasp all the empty cells as they come $y and fills them up< star'ing e'ery down stream.

BF To simulate the ,I,. queue< each station maintains two counters< 2! - !D. 2! 02equest !ounter1 counts the num$er of downstream request pending until the station itself has a frame to send at that point 2! is copied to !D< 2! is reset to 5< and now counts the num$er of request made after the station $ecame ready. ,or simplicity in the discussion $elow it is assumed that a station can ha'e only one cell ready for transmission at a time.

(ata ,ransfer in (6(B:


To send the cell< a station must first make a reser'ation $y setting the request $it in some cell on the re'erse $us. As this cell propagates down the re'erse $us< e'ery station along the way notes it and increments it@s 2!. Initially all the 2! counters are 5 and no cells are queued up as shown in the figure. Then station D makes a request< which causes station !< < and A< to increment their 2! counters< after that D makes a request copying its current 2! 'alue in !D. At this point the head end on $us A generates an empty cell. As it passes $y < that station sees that its !D S 5< so it may not use the empty cell 0when a station has a cell queue< !D represents its position in the queue< with 5 $eing front of the queue1. Instead it decrements !D. #hen the still empty cell gets to < that station sees that !D U 5< meaning that no one is ahead of it on the queue< so it inserts its data into the cell and sets the $usy $it. In this way stations queue up to take turns without a centrali/ed queue manager. )any carriers throughout the entire cities are now installing D3D )$ps. systems. Typically they run for up to 4B5 9) at speeds of ??.E;B

BH

A,

!As4nchrono+s ,ransfer

ode"

,he /&,N !/+2lic &witched ,elephone Network":


In AT - T system which can $e looked at as a general model the telephone system has fi'e classes of switching offices. !alls are generally connected at the lowest possi$le le'el. Thus< if a su$scri$er is connected to end office 4 calls another su$scri$er connected to end office 4 the call will $e completed to that office. %owe'er calls from customer attached to end office 4 to a customer attached to end office 8 will ha'e to go to toll office 4. %owe'er a call from end office 4 to end office ? will ha'e to go to primary office 4< and so on. #ith a pure tree< there is only one minimal route that could normally $e taken. The telephone companies noticed that some routes were $usier than other e.g. There were many calls from New Jork to *os Angeles. 2ather than go all the way up the hierarchy< they simply installed direct trunks for the $usy routes. ,ew such lines are shown in the figure as dashed lines. As a consequence many calls can now $e routed

E5 along many paths. The actual route choosed is generally the most direct one< $ut if the necessary trunks along it are full< the alternati'e is chosen.

Circ+it &witching and /acket &witching


Two different switching techniques are used in the telecommunications systems namely !ircuit Switching and Packet Switching.

Circ+it &witching:
#hen a computer places a telephone call< the switching equipment within the telephone system seeks out a physical copper path all the way from the senders telephone to the recei'er telephone< this technique is called !ircuit Switching and is shown fig 0i1.

E4

An important property of circuit switching is the need to setup an end6to6end path $efore any data can $e sent. The elapsed time $etween the end of dialing and the start of ringing can easily $e 45 seconds< more on long distant or international calls. During this time inter'al the telephone system is hunting for a copper path. ,or many computer applications long setup time are undesira$le. .nce the setup is completed the only delay for data is the propagation time for the electro magnetic signal a$out A msec. per thousand kms. As a result of the esta$lished path there is no danger of congestion i.e. once the call is put through< you ne'er get $usy signals< although you might get one $efore the connection has $een esta$lished due to lack of switching or trunk capacity.

/acket &witching:
In packet Switching fi&ed length $locks or packets or information is sent o'er the transmission line. y making sure that no user can monopoli/e any transmission line for 'ery long packet switching networks are well suited for handling interacti'e traffic. The further ad'antage of packet switching is that the first packet of a )ulti6packet message

E8 can $e forwarded $efore the second one has fully arri'ed< reducing delay and impro'ement throughput.

'arameter
Dedicated CcopperD path andwidth a'aila$le Potentially #asted andwidth Store6and6forward transmission (ach packet follows the same route !all Setup #hen can congestion occur !harging

C)rcu)t * Sw)tc+ed
Jes ,i&ed Jes No Jes 2equired At setup time Per minute

'acket * Sw)tc+ed
No Dynamic No Jes No Not 2equired Not packet Per packet

&4nchrono+s & As4nchrono+s ,ransfer

ode:

Analog signals at digiti/ed in the end office $y a de'ice called !odec 0!ode Decoder1< reducing F6$it num$er. The !odec makes F555 samples per second 048A micro second per sample1 $ecause the Nyquist Theorem says that this is sufficient to capture all the information from the ? k%/ telephone channel $andwidth. At a lower sampling rate< information will $e lost= at a higher one< no e&tra information would $e gained. This technique is called P!) 0Pulse !ode )odulation1. .ne method that is in wide use in North America and Oapan is the T4 carrier. The T4 carrier consists of 8? 'oice channel multiple& together. +sually< the analog signals are sampled on a round ro$in $asis with resulting analog screen $eing fade to the !odec rather than ha'ing 8? separate !odecs and then merging the digital output. .ne T4 frame is generated precisely e'ery 48A micro second. This rate is go'erned $y a master clock. AT) in contrast has no requirement that cells rigidly alternate among the 'arious sources. !ells arri'e randomly from different sources with no particular pattern.

E;

,he 2enefits of A,

are the following:

4. %igh performance 'ia hardware watching 8. Dynamic $andwidth for $ursty traffic ;. !lass6of6ser'ice support for multimedia ?. Scala$ility in speed and network si/e A. .pportunities for simplification 'ia >! architecture B. International standards compliance The high6le'el $enefits deli'ered through AT) ser'ices deployed on AT) technology using International AT) standards can $e summari/ed as followsG %igh performance 'ia hardware switching with tera$it switches on the hori/on. Dynamic $andwidth for $ursty traffic meeting application needs and deli'ering high utili/ation of networking resources. )ost applications are or can $e 'iewed as inherently $ursty= data applications are *AN6$ased and are 'ery $ursty< 'oice is $ursty since $oth parties are either speaking at once or all the time= 'ideo is $ursty since the amount of motion and required resolution 'aries o'er time. !lass6of6ser'ice support for multimedia traffic allowing applications with 'arying throughput and latency requirement to $e met on a single network. Scala$ility in speed and network si/e supporting link speeds of T64"(64 to .!648 0B88 )$ps1 today and into the multi :$ps range $efore the end of the decade. common *AN"#AN architecture allowing AT) to $e used consistently from one desktop to another. .pportunities for simplification 'ia switched >! architecture. This is particularly for *AN6 ased traffic< which today is connectionless in nature. The simplification possi$le through AT) >!s could $e in areas such as $illing< traffic management< security< and configuration management.

E? International Standards compliance in central office and customer6premise en'ironments allowing for multi6'endor operation.

A,

,echnolog4:
In AT) networks< all information is formatted into fi&ed length cells consisting

of ?F $ytes 0F $its per $yte1 of payload and A $ytes of cell header. The fi&ed cell si/e ensures that time6critical information such as 'oice or 'ideo is not ad'ersely affected $y long data frames or packets. The header is organi/ed for efficient switching in high6speed hardware implementations and carries payload6type information< 'irtual6circuit identifiers< and header error check.

EA AT) is connection oriented. .rgani/ing different streams of traffic in separate cells allows the user to specify the resources required and allows the network to allocate resources $ased on these needs. )ultiple&ing multiple streams of traffic on each physical facility com$ined with the a$ility send the streams to many different destinations ena$les cost sa'ings through a reduction in the num$er of interfaces and facilities required constructing a network. A, standards defined two t4pes of A, 0>!!1. $1 A 'irtual channel connection 0or 'irtual circuit1 is the $asic unit< which caries a single of cells< in order< from user to user. A collection of 'irtual circuits can $e $undled together into a 'irtual path connection. A 'irtual path connection can $e created from end6to end across an AT) network. In this case< the AT) network does not route cells $elonging to a particular 'irtual circuit. All cells $elonging to a particular 'irtual path are routed the same way through the AT) network< thus resulting in faster reco'ery in case of ma7or failures. An AT) network also uses 'irtual paths internally for purposes of $undling 'irtual circuits together $etween switches. Two AT) switches may ha'e many different 'irtual channel connections $etween them< $elonging to different users. These can $e $undled $y the two AT) switches into a 'irtual path connection. This can ser'e the purpose of a 'irtual trunk $etween the two switches. This 'irtual trunk can then $e handled as a single entity $y< perhaps< multiple intermediate 'irtual path cross connects $etween the two 'irtual circuit switches. >irtual circuits can $e statistically configured as permanent 'irtual circuits 0P>!s1 or dynamically controlled 'ia signaling as switched 'irtual circuits 0S>!s1. They can also $e point6to6point or point6to6multipoint< thus pro'iding a rich set of ser'ice connections:

a1 >irtual path connections 0>P!1 which contain 'irtual channel connections

EB capa$ilities. S>!s are the preferred mode of operation $ecause they can $e dynamically esta$lished< thus minimi/ing reconfiguration comple&ity.

A,

CLA&&$& ;9 &$3CIC$&:
AT) is connection oriented and allows the user to dynamically specify the

resources required on a per6connection $asis 0per S>!1. There are the fi'e classes of ser'ice defined for AT) 0as per AT) ,orum +NI ?.5 specification1. The 3os parameters for these ser'ice classes are summari/ed in the following ta$le.

ATM Serv)ce Classes"


Ser'ice !lass !onstant it 2ate 0! 21 3uality of ser'ice parameter This class is used for emulating circuit switching. The cell rate is constant with time. ! 2 applications are quite sensiti'e to cell6delay 'ariation. (&amples of applications that can use ! 2 are telephone traffic< >aria$le it 2ate6Non62eal Time 'ideo conferencing< and tele'ision. This class allows users to send traffic at a rate that 'aries with time depending on the a'aila$ility of user information. Statistical multiple&ing is pro'ided to make optimum use of network resources. )ultimedia >aria$le it 2ate62eal Time e6mail is an e&ample of > 26N2T. This class is similar to > 26N2T $ut is designed for applications that are sensiti'e to cell delay 'ariation. (&amples for real time > 2 are 'oices with speech acti'ity detection 0SAD1 and interacti'e compressed A'aila$le it 2ate 0A 21 'ideo. This class of AT) ser'ices pro'ides rate $ased flow control and is aims at data traffic such as file transfer and e6mail. Although the standard does not require the cell transfer delay and cell6loss ration to $e guaranteed or minimi/ed= it is desira$le for switches to minimi/e

EE delay and loss as much as possi$le. Depending upon the state of congestion in the network< the source is required to control its rate. The users are allowed to declare a minimum cell rate< which is guaranteed to the +nspecified it 2ate 0+ 21 connection $y the network This class is the catch6all CotherD class< and is widely used today for T!P"IP.

6+estions
4. #hat is )etropolitan Area NetworkI (&plainI 8. #hat is D3D structureI ;. #hat is Asynchronous Transfer )odeI Discuss. ?. (&plain the !oncept of !ircuit and Packet SwitchingI A. (&plain in detail the functions of Asynchronous and Synchronous Transfer )odeI B. #hat is AT) Technology - (&plain its $enefitsI

!hapter A !.))(2!IA* *AN SJST()S (T%(2N(T

EF

The (thernet *AN Standard (thernet *imitations 45 )$ps Switched (thernet 455 )$ps ,ast (thernet The Arc Net *AN The I ) Token 2ing *AN ,i$er Distri$uted Data Interface.

W:A, I& $,:$3N$,0


(thernet is a type of network ca$ling and signaling specifications 0.SI )odel layers 4 VphysicalW and 8 Vdata linkW originally de'eloped $y Mero& in the late 4HE5s. It is the least e&pensi'e high6speed *AN alternati'e (thernet adapter cards for a P! range from XB5 to X485. They transmit and recei'e data at the speed of 45 million $its per second through up to ;55 feet of telephone wire to a Chu$D de'ice normally stacked in a wiring closet. The hu$ adds less than XA5 to the cost of each desktop connection. Data is transferred $etween wiring closets using either a hea'y coa& ca$le 0CThicknetD1 or fi$er optic ca$le.

EH

In'ention of $thernet
Dr. 2o$ert )etcalfe at Mero& PA2! created the (thernet. %e reali/ed that he could impro'e on the Aloha system of ar$itrating access to a shared communications channel. %e de'eloped a new system that included a mechanism that detects when a collision occurs 0collision detect1. The system also includes Clisten $efore talkD< in which stations listen for acti'ity 0carrier sense1 $efore transmitting< and supports access to a shared channel $y multiple stations. Put all these components together. (thernet channel access protocol is called !arrier Sense )ultiple Access with !ollision Detect 0!S)A"!D1. )etcalfe also de'eloped a much more sophisticated $ackoff algorithm< which in com$ination with the !S)A"!D protocol< allows the (thernet system to function all the way upto 455 percent load. In late 4HE8< )etcalfe and his Mero& PA2! colleagues de'eloped the first e&perimental (thernet system to interconnect the Mero& Alto. The Alto was a personal workstation with a graphical user interface< and e&perimental (thernet was used to link Altos to one another< and to ser'ers and laser printers. The signal clock for the e&perimental (thernet interfaces was deri'ed from the Alto@s system clock< which resulted in a data transmission rate on the e&perimental (thernet of 8.H? )$ps. To $ase the name on the word CetherD as a way of descri$ing an essential feature of the systemG the physical medium carries $its to all stations< much the same way that the old Cluminiferous etherD was once through to propagate electromagnetic wa'es through space. Thus< (thernet was $orn.

;peration of $thernet1
(ach (thernet6equipped computer< also known as a station< operates independently of all other stations on the network< there is no central controller. All

F5 stations attached to an (thernet are connected to a shared signaling system< also called the medium. (thernet signals are transmitted serially< one $it at a time< o'er the shared signal channel to e'ery attached station. To send data a station first listens to the channel< and when the channel is idle the station transmits its data in the form of an (thernet frame< or packet. All stations on the network must contend equally for the ne&t frame transmission opportunity. This ensures that access to the network channel is fair< and that no single station can lock out the other stations. Access to the shared channel is determined $y the )edium Access !ontrol 0)A!1 mechanism em$edded in the (thernet interface located in each station.

Access and collisions


(thernet uses a protocol called C& AC(. This stands for C!arrier Sense< )ultiple Access< !ollision DetectD. The C)ultiple AccessD part means that e'ery station is connected to a single copper wire 0or a set of wires that are connected together to form a single data path1. The C!arrier SenseD part says that $efore transmitting data< a station checks the wire to see if any other station is already sending something. If the *AN appears to $e idle then the station can $egin to send data.

Need for Coll+sion (etect


An (thernet station sends data at a rate of 45 mega$its per second. That $it allows 455 nanoseconds per $it. *ight and electricity tra'el a$out one foot in a nanosecond. Therefore< after the electric signal for the first $it has tra'eled a$out 455 feet down the wire< the station has $egun to send the second $it. An (thernet ca$le can run for hundreds of feet. If two stations are located< say< 8A5 feet apart on the same ca$le< and $oth $egin transmitting at the same time< then they will $e in the middle of the third $it $efore the signal from each reaches.

F4

C!ollision DetectD part. Two stations can $egin to send data at the same time<
and their signals will CcollideD nanoseconds later. #hen such a collision occurs< the two stations stop transmitting< C$ack offD< and try again later after a randomly chosen delay period. (thernet can $e $uilt using a repeater. A repeater is a simple station that connected to two wires. Any data that it recei'es on one wire repeats $it6for6$it on the other wire. #hen collisions occur< it repeats the collision as well. The connection to the desktop uses ordinary telephone wire< the hu$ $ack in the telephone closet contains a repeater for e'ery phone circuit. Any data coming down any phone line is copied onto the main (thernet coa& ca$le< and any data from the main ca$le is duplicated and transmitted down e'ery phone line. to $e carried ;55 feet on ordinary wire. Any system $ased on collision detect must control the time required for the worst round trip through the *AN. As the term C(thernetD is commonly defined< this round trip is limited to A5 microseconds 0millionths of a second1. AT a signaling speed of 45 million $its per second< this is enough time to transmit A55 $its. At F $its per $yte< this is slightly less than B? $ytes. The repeaters in the hu$ electrically isolate each phone circuit< which is necessary if a 45 mega$it signal is going

=i'en 2elow are so-e of the de'ices +sed:


3ecei'es and then i--ediatel4 retrans-its each 2it1 It has no memory and does not depend on any particular protocol. It duplicates e'erything< including the collisions. 3ecei'es the entire -essage into -e-or41 If a collision or noise damages the message< then it is discarded. If the $ridge knows that the message was $eing sent $etween two stations on the same ca$le< then it discards it. .therwise< the

F8 message is queued up and will $e retransmitted on another (thernet ca$le. Its actions are transparent to the client and ser'er workstations. Acts as an agent to recei'e and forward -essages1 The router has an address and is known to the client or ser'er machines. Typically< machines directly send messages to each other when they are on the same ca$le< and they send the router messages addressed to another /one< department< or su$6network. 2outing is a function specific to each protocol. ,or IPM< the No'el ser'er can act as a router. ,or SNA< an APPN Network Node does the routing. T!P"IP can $e routed $y dedicated de'ices< +NIM workstations< or .S"8 ser'ers. There is a speciali/ed de'ice that finds pro$lems in an (thernet *AN. It plugs into any attachment point in the ca$le< and< sends out its own 'oltage pulse. The effect is similar to a sonar CpingD. If the ca$le is $roken then there is no proper terminating resistor. The pulse will hit the loose end of the $roken ca$le and will $ounce $ack. The test de'ice senses the echo< computes how long the round trip took< and then reports how far away the $reak is in the ca$le. If the (thernet ca$le is shorted out< a simple 'olt meter would determine that the proper resistor is missing from the signal and shield wires. Again< $y sending out a pulse and timing the return< the test de'ice can determine the distance to the pro$lem.

$thernet 9ra-e and $thernet Address


The heart of the (thernet system is the (thernet frame< which is used to deli'er data $etween computers. The frame consists of a set of $its organi/ed into se'eral fields. These fields include address fields< a 'aria$le si/e data field that carries from ?B to 4<A55 $ytes of data< and an error checking field that checks the integrity of the $its in the frame to make sure that the frame has arri'ed intact. The first two fields in the frame carry ?F6$it address< called the destination and source address. The I((( controls< the assignment of these addresses $y administering a

F; portion of the address field. The I((( does this $y pro'iding 8?6$it identifiers called C.rgani/ationally +nique IdentifiersD 0.+Is1< since a unique 8?6$it identifier is assigned to each organi/ation< in turn< creates ?F6$it address using the assigned .+I as the first 8? $its of the address. This ?F6$it address is also known as the physical address< hardware address or )A! address. #hen (thernet frame is sent onto the shared signal channel< all (thernet interfaces look at the first ?F $it field of the frame< which contains the destination address. The interfaces compare the destination address of the frame with their own address. The (thernet interface with the same address as the destination address in the frame will read in the entire frame and deli'er it to the networking software running on that computer. All other network interfaces will stop reading the frame when they disco'er that the destination address does not match their own address.

=i'en 2elow is a diagra- of $thernet fra-es1

Pream$le

S,D

Destination

Source

*(N

Data

PAD

!2!

8 or B

8 or B

564A55

56?B

$thernet 9ra-es

%igh6*e'el Protocols and (thernet Addresses


!omputers attached to an (thernet can send application data to one another using high6le'el protocol software< such as the T!P"IP protocol suite used on the worldwide Internet. The high6le'el protocol packets are carried $etween computers in the data field of (thernet frames. The system of high6le'el protocols carrying application data and the (thernet system are independent entities that cooperate to deli'er data $etween computers.

F? To make things< work< there needs to $e some way t o disco'er the (thernet addresses of other IP6$ased stations on the network. ,or se'eral high6le'el protocols< including T!P"IP< this is done using yet another high6le'el protocol called the Address 2esolution Protocol 0A2P1. AS an e&ample of how (thernet and one family of high6le'el protocols interact< let@s take a quick look at how the< A2P protocol functions. Also known as the logical topology< to distinguish it from the actual physical layout of the media ca$les. The logical topology of an (thernet pro'ides a single channel 0or $us1 that carries (thernet. !onnector 0Transcei'er1

Printer P.!. #orkstation

us 0 ack$one1 P.! #orkstation P.! ,ile Ser'er us Terminator )ultiple (thernet segments can $e linked together to form a large (thernet *AN using a signal amplifying and retiming de'ice called a repeater. Through the use of repeaters< a gi'en (thernet system of multiple segments can grow as a Cnon6rooted $ranching tree. CThis means that each media segment is an indi'idual $ranch of the complete signal system. ('en though the media segments may $e physically connected in a star pattern< with multiple segments attached to a repeater< the logical topology is still that of a single (thernet channel that carries signals to all stations.

FA The notion of CtreeD is 7ust a formal name for systems like this< and a typical network design actually ends up looking more like a comple& concatenation of network segments. .n media segments that support multiple connections< such as coa&ial (thernet< you may install a repeater and a link to another segment at any point on the segment. .ther types of segments known as link segments can only ha'e one connection at each end this is descri$ed in more detail in the indi'idual media segment chapters. DNon?rootedE means that the resulting system of linked segments may grow in any direction< and does not ha'e a specific root segment. )ost importantly< segments must ne'er $e connected in a loop. ('ery segment in the system must ha'e two ends< since the (thernet system will not operate correctly in the presence of loop paths. There are media segments linked with repeaters and connecting to stations. A signal sent from any station tra'els o'er that station@s a signal sent from any station tra'els o'er that station@s segment and is repeated onto all other segments. This way all other stations hear it o'er the single (thernet channel.

$<tending $thernet with :+2s


(thernet was designed to $e easily e&panda$le to meet the networking needs of a gi'en site. To help e&tend (thernet systems< networking 'endors sell de'ices that pro'ide multiple (thernet ports. These de'ices are known as hu$s< since they pro'ide the central portion or hu$< of a media system. There are two ma7or kinds of hu$sG repeater hu$s and switching hu$s. (ach port of a repeater hu$ links indi'idual (thernet media segments together to create a larger network that operates as a single (thernet *AN. The total set of segments and repeaters in the (thernet *AN must meet the round trip timing specifications. The second kind of hu$ pro'ides packet switching< typically $ased on $ridging ports. The important thing to know at this point is that each port of a packet switching hu$ pro'ides a connection to an (thernet media system that operates as a separate

FB (thernet *AN. +nlike a repeater hu$ whose indi'idual ports com$ine segments together to create a single large *AN< a switching hu$ makes it possi$le to di'ide a set of (thernet media systems into multiple *ANs that are linked together $y way of the packet switching electronics in the hu$. A gi'en (thernet *AN can consist of merely a single ca$le segment linking some num$er of computers< or it may consist of a repeater hu$ linking se'eral such media segments together. All (thernet *ANs can themsel'es $e linked together to form e&tended network systems using packet switching hu$s. #hile an indi'idual (thernet *AN can typically support anywhere from a few up to se'eral do/en computers< the total system of (thernet *ANs linked with packet switches at a gi'en site may support many hundreds or thousands of machines.

FE

($9INI,I;N& AN( &,AN(A3(&


The I((( was assigned the task of de'eloping formal international standards for all *ocal Area Network technology. It formed the CF58D committee to look at (thernet< Token 2ing< ,i$er .ptic< and other *AN technology. The o$7ecti'e of the pro7ect was not 7ust to standardi/e each *AN indi'idually< $ut also to esta$lish rules that would $e glo$al to all types of *ANs so that data could easily mo'e from (thernet to Token 2ing or ,i$er .ptics. The I((( was careful to separate the new and old rules. It recogni/ed that there would $e a period when old DIM messages and new I((( F58 messages would ha'e to coe&ist on the same *AN. It pu$lished a set of standards of which the most important areG F58.; %ardware standards for (thernet cards and ca$les. F58.A %ardware standards for Token 2ing cards and ca$les. F58.8 The new message format for data on any *AN. The F58.; standard further refined the electrical connection to the (thernet. It was immediately adopted $y all the hardware 'endors. %owe'er< the F58.8 standard would require a change to the network architecture of all e&isting (thernet users. Apple had to change its (ther talk< and did so when con'erting from phase 4 to phase 8 Appletalk. D(! had to change its D(!N(T. No'ell added F58 as an option to its IPM< $ut it supports $oth DIM and F58 message formats at the same time. The T!P"IP protocol used $y the Internet refused to change. Internet standards are managed $y the I(T, group< and they decided to stick with the old DIM message format indefinitely. This produced a deadlock $etween two standards organi/ations that has not $een resol'ed.

FF I ) waited until the F58 committee released its standards< and then rigorously implemented the F58 rules for e'erything e&cept T!P"IP where the I(T, rules take precedence. This means that N(T (+I 0the format for N(T I.S on the *AN1 and SNA o$ey the F58 con'entions. So C(thernetD suffers from too many standards. The old DIM rules for message format persist for some uses 0Internet< D(!N(T< and some No'ell1. The new F58 rules apply to other traffic 0SNA< N(T (+I1. The most pressing pro$lem is to make sure that No'ell clients and ser'ers are configured to use the same frame format.

LI I,A,I;N& ;9 $,:$3N$,:
.ld style (thernet $us wiring is prone to ca$le failure and quickly consumes allowed distances due to the aesthetic wiring needs. (thernets fail in three common ways< 4. A nail can $e dri'en into the ca$le $reaking the signal wire. 8. A nail can $e dri'en touching the signal wire and shorting it to the e&ternal grounded metal shield. ;. ,inally< a station on the *AN can $reak down and start to generate a continuous stream of 7unk< $locking e'eryone else from sending. (thernet is particularly suscepti$le to performance loss from such pro$lems when people ignore the CrulesD for wiring (thernet. There are practical limitations to the distance of a shared medium and the num$er of workstations you can connect to it.

FH

The electrical characteristics of the ca$le also dictate *AN limitations. Network designers ha'e to find a $alance $etween the type of ca$le used< the transmission rates< signal loss o'er distance and the signal emanations. All these factors must stay within physical $ounds and restrictions specified $y 'arious standards and go'ernment $odies. Delay is another factor. .n (thernet networks< workstations on either end of a long ca$le may not e'en detect that they are transmitting at the same time< thus causing a collision that results in corrupted data. Sometimes late collisions occur when two de'ices transmit at the same time< $ut due to ca$ling errors 0most commonly< e&cessi'e network segment length to repeaters $etween de'ices1 neither detects a collision. The reason this happens is $ecause the time to propagate the signal from one end of the network to another is longer than the time to put the entire packet on the network< so the two de'ices that cause the late collision ne'er see that the other@s sending until after it puts the entire packet on the network. A network suffering a measura$le rate of late collisions 0on large packets1 is also suffering loss on small packets.

1F

2ps &WI,C:$( $,:$3N$,

The following list descri$es the different 'arieties of 45 )$ps Switched (thernetG 45 ase 8 is 45)%/ (thernet running o'er thin< A5 .hm $ase$and coa&ial ca$le. 45 ase8 is also commonly referred to as thin6(thernet or !heapernet. 1FBase# is 45 )%/ (thernet running o'er standard 0thick1 A5 .hm $ase$and coa&ial ca$ling. 1FBase f is 45)%/ (thernet running o'er fi$er6optic ca$ling.

H5 1FBase, is 45)%/ (thernet running o'er unshielded< twisted6pair ca$ling. 1FBroad 3% is 45)%/ (thernet running through a $road$and ca$le.

:ow does 1F

2ps $thernet work0

The P%J is the actual transcei'er that can $e a separate de'ice or it can $e integrated on the network card. The transcei'er interface is called A+I 0Attachment +nit Interface1. #hen a network card doesn@t contain the interface there will $e a Su$6D4A female connector. .n the ca$le will $e a transcei'er with a male connector. This means that an A+I6ca$le will A*#AJS $e male6female. The difference $etween the different 45 )$ps topologies is in the P%J part. This section connects directly to the ca$le and is responsi$le for e'erything that is medium depended likeG line encoding< transmission 'oltages< S3(< etc. #ith A+I there are two ways power can $e pro'ided to the units. Jou either ha'e a positi'e or negati'e polarity.

&ignal 6+alit4 $rror !&6$":


The Signal 3uality (rror signal is also called Kheart$eat@ and is a kind of keep ali'e notification $etween the transcei'er and the (thernet de'ice. S3( can $e .N or .,, $etween a transcei'er and a workstation or file ser'er. It )+ST $e set .,, $etween a transcei'er and a 2epeater. )a& Speed 645 )$ps !a$le Standard6(thernet !oa& !a$le !onnectors6N6Type Terminators6A5 .hm )a&.length of a segment6A55m"4B?oft

H4 )a&.num$er of taps per segment6455 )a& .num$er of stations per network6458? )in.distance $etween taps68.Am"F.;ft )a&.length of transcei'er ca$le6A5m"4B?ft )a&.num$er of repeaters6? Topology6 us.

1FF

2ps 9A&, $,:$3N$,


There are two *AN standards that can carry (thernet frames at 4556)$ps. #hen the I((( standardi/ation committee met to $egin work on a faster (thernet

system< two approaches were presented. .ne approach was to speed up the original (thernet system to 455 )$ps< keeping the original !S)A"!D medium access control mechanism. This approach is called 455 AS(6T ,ast (thernet. Another approach presented to the committee was to create an entirely new medium access control mechanism< one $ased on hu$s that controlled access to the medium using a Cdemand priorityD mechanism. This new access control system transports standard (thernet frames< $ut it does it with a new medium access control mechanism. This system was further e&tended to allow it to transport token ring frames as well. As a result< this approach called 455>:6Any*AN.

1FF? 2ps

edia &4ste-s

!ompared to the 456)$ps specifications< the 4556)$ps system results in a factor of ten reductions in the $it6time< which is the amount of time it takes to transmit a $it on the (thernet channel. This produces a tenfold increase in the speed of the packets o'er the media system. %owe'er< the other important aspects of the (thernet system include the frame format< the amount of data a frame may carry< and the media access control mechanism< are all unchanged.

H8

The ,ast (thernet specifications include mechanism for Auto6Negotiation of the media speed. This makes it possi$le for 'endors to pro'ide dual6speed (thernet interfaces that can $e installed and run at either 456)$ps or 4556)$ps automatically. The I((( identifiers include three pieces of information. The first item< C455D< stands for the media speed of 4556)$ps. The C AS(D stands for C$ase $and<D #hich is a type of signaling. media system. ase$and

signaling simply means that (thernet signals are the only signals carried o'er the

The third part of the identifier pro'ides an indication of the segment type. The CT?D segment type is a twisted6pair segment that uses four pairs of telephone6 grade twisted6pair wire. The CTMD segment type is a twisted6pair segment that uses two pairs of wires and is $ased on the data graded twisted6pair physical medium standard de'eloped $y ANSI. The C,MD segment type is a fi$er optic link segment $ased on the fi$er optic physical medium standard de'eloped $y ANSI and that uses two strands of fi$er ca$le. standards are collecti'ely known as 455 AS(6M. The TM and ,M medium

Co-ponents 8sed for a 1FF? 2ps Connection


The physical medium is used to carry (thernet signals $etween computers. This could $e any one of the three 4556)$ps media types.

/h4sical La4er (e'ice


This de'ice performs the same general function as transcei'er in the 456)$ps (thernet system. It may $e a set of integrated circuits inside the (thernet port of a

H; network de'ice< therefore in'isi$le to the user< or it may $e a small $o& equipped with an )II ca$le< like the out$oard transcei'er and transcei'er ca$le.

edi+- Independent Interface


The )II is an optional set of electronics that pro'ides a way to link the (thernet medium access control functions in the network de'ice with the Physical *ayer De'ice 0P%J1 that sends signals onto the network medium. The )II is designed to make the signaling differences among the 'arious media segments transparent to the (thernet chips in the network de'ice. The )II electronics may $e linked to an out$oard transcei'er through a ?56pin )II connector and a short )II ca$le.

/+tting it All ,ogether


,or a typical station connection the D(T 0computer1 contains an (thernet interface which forms up and sends (thernet frames that carry data $etween computers attached to the network. The (thernet interface is attached to the media system using a set of equipment that might include an out$oard )II ca$le and P%J with its associated )DI 0twisted6pair 2O?A6style 7ack or fi$er optic connector1. The interface or repeater port might also $e designed to include the P%J electronics internally< in which case all you will see is )DI for whate'er physical medium the interface or port was designed to support.

A3CN$, C;N9I=83A,I;N
The A2! net 0Attached 2esource !omputing Network1 is a $ase $and. Token passing network system that offers fle&i$le star and $us topologies at a low price. Transmission speeds are 8.A )$its per second. A2! net uses a token6passing protocol on

H? a token $us network topology. A2!net is showing its age and is no longer sold ma7or 'endors.

IB

,;@$N 3IN=
In 4HFA I ) announced its ma7or entry into *AN field with the I ) Token ring.

The I ) topology permits se'eral rings to $e attached through the $ridges. A $ack$one ring then connects the $ridges. The $ridge will pro'ide a cross6ring network $y copying frames that are forwarded from one ring to another. The $ridges also pro'ide for speed translations if rings are operating at different data rates. )ore e'er each ring still retains its own capacity and will continue operating in the e'ent another ring on the $ridge fails. I ) Token ring approach pro'ides resiliency to station and link failure.

HA

9IB$3 (I&,3IB8,$( (A,A IN,$39AC$ !9((I"


,DDI is a high6speed networking technology de'eloped $y the ANSI 0American National Standards Institute1 M;TH.A committee. It was originally designed for fi$er6 optic ca$les $ut now supports copper ca$le o'er short distances. The standard is commonly used for *AN and campus en'ironment. ,DDI has a data rate of 455 )$its"sec and uses a redundant dual ring topology that supports A55 nodes o'er a

HB ma&imum distance of 455 kilometers. Such distances also qualify ,DDI of use as a )AN 0)etropolitan Area Network1. The dual counter6rotating rings offer redundancy 0fault tolerance1. if a link fails or the ca$le is cut< the ring configures itself and the network keeps operating. (ach station contains relays that 7oin the rings in case of a $reak or $ypass the station if it is ha'ing pro$lems. ,DDI has $een used e&tensi'ely as a network $ack$one topology. *AN segments attach to the $ack$one< along with minicomputers< mainframes and other systems. Small networks that consist of a few *AN segments and hea'y traffic produced $y high6 performance workstations< graphics file transfers< or other internetwork traffic will $enefit from ,DDI.

9((I Config+ration
The topology is called a physical ring of trees $ut logically the entire network forms a ring. The two ,DDI rings are known as the primary ring and the secondary ring. oth may $e used as a transmission path or one may $e set aside for use as a $ack up in the e'ent of a $reak in the primary ring. There are three types of de'ices that can attach to the ringG DAS 0dual attached station1 connected to $oth rings< such as a critical ser'er and other pieces of equipment. DA! 0dual attached concentrator1 6 connected to $oth rings and pro'ides a connection point for stations. SAS 0single attached station1 attached to the primary ring 'ia connector. If a computer attached to an ,DDI concentrator fails< the concentrator ensures the ring is maintained.

HE ecause ,DDI implements a logical ring in a physical star< you can $uild hierarchical networks. ,DDI operates o'er a single6mode and multi6mode fi$er optic ca$le as well a STP 0shielded twisted pair1 and +TP 0unshielded twisted pair1 copper ca$le.

9((I ;peration and Access

ethod

,DDI uses a token6passing access method. A token frame is passed around the network from station to station= if a station needs to transmit< it acquires the token. The station then transmits the frame and remo'es it from the network after it makes a full loop. A regulation mechanism is used to pre'ent one station from holding the token for too long. The ,DDI frame si/e is 4<A55 $ytes.

9((I now has three trans-ission -odes1


As4nchrono+s ring -ode: This is token6$ased. Any station can access the network $y acquiring the token. In this mode traffic is not prioriti/ed. &4nchrono+s token?passing ring -ode: Allows prioriti/ation. ,DDI cards with synchronous capa$ilities gi'e network managers the a$ility to set aside part of the $andwidth for time6sensiti'e traffic. Asynchronous workstations then contend for the rest. Synchronous capa$ilities are added 'ia software upgrades. The a$o'e two modes are a'aila$le in the original ,DDI standard. The third mode< circuit6$ased can pro'ide dedicated circuits that can $e prioriti/ed for 'oice and other real time traffic. This mode is a'aila$le in the new ,DDI6II standard< which requires new adapter cards.

HF

9((I?II:
Is designed for networks that need to transport real6time traffic. It is ,DDI modified to support synchronous data such as 'oice circuits and ISDN 0Integrated Ser'ices Digital Network1 traffic. ,DDI6II requires all nodes on the ,DDI6II network to use ,DDI6II= otherwise the network re'erts to ,DDI. ,DDI uses multiple&ing technologies to di'ide the $andwidth into 4B dedicated circuits that can pro'ide on6time deli'ery for prioriti/ed traffic. The circuits operate at from B.4??)$its"sec each to a ma&imum of HH.5E8 )$its"sec. The reason for this 'ariation is that the $andwidth is allocated to whate'er station that has the highest priority for it. (ach of these channels can $e su$di'ided further to produce a total of HB B?9$it"sec circuits. These channels can support asynchronous or synchronous traffic. 2egular time slots in the ring are allotted for the transmission of data. Prioriti/ed stations use the num$er of slots they need to deli'er their data on time. If the slots are unused< they are reallocated immediately to other stations that can use them. ,DDI6II has not $ecome a widespread networking technology $ecause it is incompati$le with the e&isting ,DDI design. Another reason is that the 4556)$it"sec (thernet and AT) 0Asynchronous Transfer )ode1 ha'e pro'ided $etter solutions in most cases.

6+estions
4. #hat is (thernetI #hat are its limitationsI 8. (&plain the functions of 45)$ps - 455)$ps (thernetsI ;. #hat is an A2!N(T *ANI (&plain. ?. (&plain in $rief I ) Token ringI ,i$re Distri$uted Data Interface 0,DDI1 TechniquesI

HH

!hapter B INT(:2AT(D S(2>I!(S DI:ITA* N(T#.29

Introduction Types of ISDN ,unctions of ISDN ISDN Standards ,uture Applications of ISDN +ser Network Interfaces

455

Introd+ction to I&(N
The telephone ser'ice has $een de'eloped o'er the last 455 years. Initially its sole aim was to pro'ide simple one to one 'oice communications $etween su$scri$ers< $ut we ha'e seen that technology has influenced the telephone network in two ways. ,irstly impro'ements in technology such as the introduction of digital switching< computer control and common channel signaling ha'e meant that the network can offer its users far more facilities than simple one to one 'oice calls. Secondly the introduction of new technology in other $usiness areas has resulted in a situation in which the P.TS 0Plain .ld Telephone Ser'ice1 are carrying a wide 'ariety of data communications traffic. Although it is true to say that the ma7or use of the network is still for 'oice communications< a growing percentage of the traffic is accounted for $y digital traffic< i.e. data communications and facsimile. The limitations of Traditional !ommunications Networks< which used analogue switching and transmission are caused $y the following factorsG 4. The old network is noisy< resulting in $it errors. 8. !all setup times are long< the call set up time may e&ceed the holding time. ;. Transmission is limited to specific $andwidth pathway. ?. 2outing of calls is not fi&ed< and thus 'ariations in transmission performance due to effects such as group delay are e&perienced on different calls $etween any two gi'en locations. .'er recent years the communications infrastructure has e'ol'ed in such a way that ser'ices are pro'ided on dedicated networks< each with its own su$scri$er access and interface requirements. The ser'ices pro'ided can $e categori/ed intoG

454 4. Point to Point Digital *eased lines 8. !ircuit Switch Telegraph 0Tele&1 ;. Packet Switch data network The cost of $uilding and maintaining dedicated networks is so large that it can only $e contemplated if the demand for the ser'ice is large enough to generate sufficient re'enue to make it economic. These high costs therefore prohi$it the introduction of new speciali/ed communication ser'ices.

458

;,ICA,I;N 9;3 I&(N


,hree factors are responsi2le for the de'elop-ent towards I&(N:
4. Sociological or societal needs 8. (conomic necessities ;. Technological de'elopment

&ociological or societal needs:


The rapid de'elopments in 'arious facets of the society call for increasing and comple& communication facilities. A $iotechnologist today would like to e&amine a $lood e&ample remotely< simultaneously compare the analytical results of other samples stored in a centrali/ed data$ase< consult his assistant who is presently in a la$oratory some distance away< and report the finding as the findings as the in'estigation progresses< to his superior who is in another $uilding. To meet such a demand< we need to electronically transmit the microscopic image of the $lood sample and reproduce the same graphically on the computer screen of the $iotechnologist< at a rate fast enough to faithfully reproduce the mo'ements of li'ing cell< etc. As another e&ample< a senior e&ecuti'e of a company< who often has to take important decisions at home or late in the e'ening or while on a holiday would like to gi'e instant effect to his decisions. This may call for access to different computer systems connected in the form of a network< processing facilities< all in the place where he is at present. In effect< the society is demanding a telecommunication system that can support uni'ersal access to a host of ser'ices. In such a system< it should $e possi$le for a user to the network anywhere in the world the equipment of his choice to o$tain a particular ser'ice. The user will $e allotted a permanent identification num$er or code< like the income ta& permanent account num$er or the social security num$er< which would $e 'alid for his lifetime.

45; Traditionally< network pro'iders ha'e put up separate and independent networks to support different ser'ices. Tele& network data network< telephone network and !AT> networks are e&amples of such a de'elopment. Independent networks call for separate administration< maintenance staff< and $uilding for housing switching systems. The independent and duplicate infrastructural facilities lead to high capital cost< low maintenance efficiency and high maintenance cost. In addition< the network facilities are ne'er fully utili/ed as the ser'ices are independently supported on different networks. Searching for new solutions is of no a'ail unless technology de'elopments make possi$le such solutions. In fact< it is the technology factor that $rought a$out the independent network solutions earlier. The end equipment@s for different ser'ices were analog in nature and had different electrical< electronic< signal and communication characteristics. It was necessary to design different communication characteristics. It was necessary to design different networks to suit each of these de'ices. The desire of the network pro'iders to use a common network infrastructure can fructify only if there are uniform for all the ser'ices. Today< the digital technology has matured to a le'el where all the a$o'e 6 mentioned domain. functions of a telecommunications network can $e reali/ed in the digital

$'ol+tion
Integrated Digital Ser'ices Network 0ISDN1 has $een perhaps the most important de'elopment to emerge in the field of !omputer !ommunications in the 4HF5@s and it will pro$a$ly continue to dominate the de'elopments in the 4HH5@s too. +nlike many other de'elopments< ISDN is a well concei'ed and planned area of de'elopment in the field of communications. !!ITT has $een pioneering and guiding the efforts towards the de'elopment of ISDN. !!ITT was quick to recogni/e the feasi$ility of digital telecommunication networks and set up a study group called Special Study :roup D in 4HBF to look at a 'ariety of issues related to the use of digital technology in telephone networks. This study

45? group is the forerunner of today@s Study :roup M>III set up in 4HEB< and has the responsi$ility for all ISDN related acti'ities within the !!ITT. Integrated Ser'ices Digital Network An integrated digital network in which the same digital switches and digital paths are used to esta$lish different ser'ices< for e&ample< telephony and data. In 4HF5 the first set of ISDN standards emerged which laid down the conceptual principles on which ISDN should $e $ased. 4. ISDN will $e $ased on and will e'ol'e from the telephony IDN $y progressi'ely incorporating additional functions and network features including those of any other dedicated networks. 8. New Ser'ices introduced into the ISDN should $e so arranged and should $e compati$le with B? 9$ps switched digital connections. ;. The transition from the e&isting networks to a comprehensi'e ISDN may require a period of time e&tending o'er one or two decades. ?. During the transition period arrangements must $e made for the networking of ser'ices on ISDNs and ser'ices on other ser'ices. A. The ISDN will contain intelligence for the purpose of pro'iding ser'ice features< maintenance and network management at functions. This intelligence may not $e sufficient for some new ser'ices and may ha'e to $e supplemented $y either additional intelligence in the customer terminals. B. The layered functional set of protocols appears desira$le for the 'arious access arrangements to ISDN.

45A The analogue systems are $eing replaced $y new digital networks which ha'e $een de'eloped to cater for all forms of digital communications. CThe merging of technologies coupled with increasing demands for the efficient collection< processing and dissemination of information is leading to the de'elopment of integrated systems that transmit and process all types of information. The ultimate goal of this e'olution in communications is called the Integrated Ser'ices Digital Network 0ISDN1. The standard mo'ement was started $y the International Telephone and Telegraph !onsultati'e !ommittee 0!!ITT1. The ISDN will e'entually $e a worldwide pu$lic telecommunications network which will deli'er a wide 'ariety of ser'ices. The ISDN will $e defined $y the standardi/ation of user interfaces< and will $e implemented as a network of digital switches and transmission paths which support a $road range of traffic types and pro'ide 'alue added processing ser'ices.

eaning & (efinition


In early 4HE8< two definitions were formulated $y !!ITT which descri$ed the de'elopment of an analog into a digital telephone network and its further e'olution in to an ISDN. CAn Integrated Digital Network 0ISDN1 is a network in which connections esta$lished $y digital switching are used for transmission of digital signals.D CAn Integrated Digital Network 0ISDN1 is an integrated Digital Network 0IDN1 in which the same digital switches and digital switches and digital paths are used to esta$lish connections for different ser'ices.D CIntegrated Ser'icesD refers to ISDN@s a$ility to deli'er two simultaneous connections< in any com$ination of data< 'oice< 'ideo and fa&< o'er a single line. )ultiple de'ices can $e attached to the line and send as needed.

45B CDigitalD in ISDN refers to its purely digital transmission< as opposed to the analog transmission of plain old telephone ser'ice. ISDN transmits data digitally< resulting in a 'ery clear transmission quality. There is none of the static and noice of analog transmissions that can slow transmission speed. CNetworkD refers to the fact that ISDN is not simply a point6to6point solution like a leased line. ISDN networks e&tend from the local telephone e&change to the remote user and include all of the telecommunications and switching equipment in $etween. #hen you ha'e ISDN< you can make connection throughout the world to other ISDN equipment. If your ISDN equipment includes analog capa$ilities< you can also connect analog. #hile ISDN accommodates telephones and fa& machines< its most popular ad'antage is in computer applications. Jou can plug an ISDN adapter into a phone 7ack< like you would an analog modem and get a much faster connection with no line noise. An ISDN is a network< in general e'ol'ing from telephony ISDN< which pro'ides end6to6end digital connecti'ity to support a wide range of ser'ices< including 'oice and non6'oice ser'ices< to which users ha'e access $y a limited set of standard multipurpose user network interfaces.

,he ke4 points of the a2o'e definition ha'e to 2e noted1


4. The ISDN is an infrastructure to support a wide 'ariety of ser'ices and is not a network designed for any specific ser'ice. 8. The end6to6end digital connecti'ity implies that the digiti/ation process $egins right at the user premises. ;. It should $e possi$le to support e'ery concei'a$le ser'ice on ISDN< for any such ser'ice is either a 'oice or non6'oice ser'ice.

45E ?. A small set of carefully chosen interfaces should ena$le the support of all possi$le ser'ices. The users of ISDN should not $e $urdened with too many speciali/ed interfaces< $ut at the same time< an e&pensi'e uni'ersal interface should $e a'oided.

Types of ISDN Ser'icesG


4. Basic 3ate Interface !B3I": It pro'ides two single B? 9$ps channels per line. #hen the two channels are $ounded in a single connection< you get a speed of 48F 9$ps< which is a$out four times the actual top speed of the fastest analog

45F modems. Telecommuters< for e&ample< $enefit immensely from ISDN< whether you access the corporate *AN in the e'enings or maintain a full6time< remote home office= ISDN is the ne&t $est solution.

8. /ri-ar4 3ate Interface !/3I": is intended for users with greater capacity requirements. Typically the channel structure is 8; one B? 9$ps D channel for a total of 4HF? 9$ps. channels plus one B? 9$ps channels plus D channel for a total of 4A;B k$ps. In (urope< P2I< consists of ;5

&er'ices offered 24 I&(N:


ISND en'isaged $eing an intelligent network. In the future< concepts of artificial intelligence and e&pert systems will $e applied to network functions. In particular< network maintenance and network are the potential areas for the application of AI concepts and e&pert systems.

&+pple-entar4 &er'ices:
Supplementary ser'ices call for additional functionalities $oth in the lower layers and in the upper layers< depending on whether they supplement a $asic $earer ser'ice or a $asic tele6ser'ice.

45H

Broad2and I&(N:
Is defined as a network capa$le of supporting data rates greater than the primary rate supported $y ISDN. In the conte&t of ISDN< the original ISDN concept is often ISDN is to support 'ideo and termed narrow6$and ISDN 0NISND1 The main aim of image ser'ices. ISDN ser'ices are $roadly classified asG 4. Interacti'e Ser'ices 8. Distri$ution Ser'ices

Interacti'e &er'ices -a4 2e classified as:


!on'ersational Ser'ices )essaging Ser'ices 2etrie'al Ser'ices

(istri2+tion &er'ices are classified as


roadcast Ser'ices !yclic Ser'ices

a" Con'ersational &er'ices: It Supports end6to6end information transfer on


real time< $i6directional $asis. There is a wide range of applications that may $e supported using con'ersational ser'ices< the most important one $eing the 'ideo telephony or 'ideophone. .ther applications include 'ideo conferencing and 'ideo sur'eillance. A num$er of data oriented con'ersational applications may also $e supported. These include distri$uted data$ases< program downloading< inter6process communication and large 'olume high speed data e&change as encountered in !AD"!A) or graphics $ased applications.

2"

essaging &er'ices: It .ffers store and forward communication.

Analogous to M.?55 messaging ser'ices on NISDN< 'oice mail< 'ideo mail and document mail containing te&ts< graphics etc. may $ecome the important messaging ser'ices on ISDN.

445

c" 3etrie'al &er'ices: In

ISDN offer the capa$ility to retrie'e sound ISDN retrie'al ser'ices are an enhancement

passages< high resolution images< graphics< short 'ideo scenes< animated pictures etc. from centrali/ed or distri$uted data$ases. of 'ideote&t ser'ices in NISDN.

d" Broadcast distri2+tion ser'ices: It pro'ides support for $roadcasting


'ideo< facsimile and graphical images to su$scri$ers. ,or e&ample such applications include tele'ision $roadcasting o'er the network and electronic newspaper distri$ution.

d" C4clical distri2+tion ser'ices: It offers some control to the user in the
presentation of information on the screen. The cyclic distri$ution ser'ices are an enhancement of the con'entional telete&t ser'ices.

New &er'ices:
ISDN will support a 'ariety of ser'ices including the e&isting 'oice and data ser'ices and a host of new ser'ices. Short list of some of the important new ser'ices areG 4. >ideote& 8. (lectronic mail ;. Digital facsimile ?. Tele& A. Data$ase access B. (lectronic fund transfer E. Image and graphics e&change F. Document Storage and transfer H. Automatic alarm ser'ices e.g. smoke< fire< police and medical. 45. Audio and >ideo conferencing.

444

A few of the ser'ices are descri2ed in the following areas:


4. Cideote<: Is a generic term for systems that pro'ide easy to use< low cost computer $ased ser'ices 'ia communication facilities. Three forms of 'ideote&t that e&ists areG o Ciew (ata: is fully interacti'e 'ideote&t< this means that requests for information or ser'ice from a user and performs to send< recei'e and act $y a centrali/ed computer. o ,elete<: It is $roadcast or pseudo6interacti'e 'ideote&t ser'ice. Telete& users may select the information to $e seen< the pace at which the information is to $e displayed< and often< the sequence of display. Telete& is one way communication system and there is no real interaction $etween the user and the computer. o ;pen channel telete<: is totally interacti'e and is a one6way 'ideote&t. #ith this form of 'ideote&t< the user recei'es pre6selected information in a predetermined order. 8. $lectronic -ail: (lectronic mail is popularly known as the e6mail and may $e defined as the communication of te&tual messages 'ia electronic means. (lectronic mail communication is from user6to6user means. (lectronic mail is a store and forward 0S-,1 ser'ice. It is a computer $ased messaging system. It permits communication $etween two parties without the parties actually $eing present simultaneously. Pri'acy is also ensured as only the intended recipient can open it. (mail also reduces the consumption of paper in the office. ;. 9acsi-ile: Documents are e&changed through the facsimile systems and it is emerging as a ma7or application of telecommunication systems. It is capa$le of transmitting and recei'ing printed matter which may include graphics< drawings< and pictures< hand written te&t< etc.

448 ?. ,elete<: It is an upgrade to the con'entional tele& ser'ice. The terminal6to6 terminal communication ser'ice of tele& will $e turned into office6to6office document computers. A. (ata2ase access: A user can $y suita$le search query< o$tain all the information generated in a particular topic. There are o'er A555 data$ases in different parts of the world< co'ering a wide 'ariety of su$7ects< which include social sciences< science and technology< engineering and industry. These data$ases may $e accessed online using the telephone network< modem and a personal computer. transmission system $y telete&. Telete& en'isages direct communication $etween electronic typewriters< word processors and personal

:ow (oes I&(N 9+nction0


SignalingG ISDN uses a common channel signaling scheme< the signaling is done o'er the D channel which acts as the common signaling channel for the and % channels which carry the user information. D channel may also $e used for carrying some user information< if there is spare capacity. In such cases also< the required signaling is done on the D channel. Signaling in ISDN falls into two distinct categoriesG 4. +ser le'el Signaling 8. Network le'el Signaling All user generated signaling and the signaling features that are open to the user are treated as user le'el signaling and are defined as part of the layer ; user network interface standards. The signaling facilities employed $y the network to support user le'el signaling and to implement network control functions< not directly related to the user are treated as network le'el signaling.

44;

8ser le'el signaling in I&(N per-its a +ser to:


4. (sta$lish< control and terminate circuit switched connections in 8. !arry out user6to6user signaling and ;. (sta$lish< control and terminate packet switched connections in channels. or D channel<

8ser le'el signaling is of two t4pes:


4. )essage $ased signaling 8. Stimulus signaling )essage $ased signaling is employed when the user end equipment is an intelligent terminal. In ISDN parlance< as intelligent terminal is known as functional terminal. It pro'ides a user6friendly interface for signaling and performs the functions of forming< sending< recei'ing and replying messages. The process of esta$lishing< controlling and terminating a call is achie'ed $y e&changing messages $etween the network and terminal. The messages may $e placed under four groupsG 4. !all (sta$lishment )essages 8. !all !ontrol )essages ;. !all Discount )essages ?. )iscellaneous )essages !all esta$lishment group includes set6up< call proceeding< alert< connect and connect acknowledge messages. Alert signal corresponds to ring $ack signal and is used when a non6automatic answering terminal is used at the recei'ing end. If the auto6 answering facility is a'aila$le< the terminal responds with connect signal directly and the alert signal is skipped

!all control group includes suspend and resume messages and also user6to6user messages.

44? !all disconnect group includes disconnect release and release complete messages. The primary function of the miscellaneous messages is to negotiate network facilities to support additional ser'ice. All user le'el messages ha'e a common message format. There fields are mandatory for all messagesG 4. Protocol discriminator 8. !all reference ;. )essage type As the D channel may carry computer and telemetry data etc. in addition to signaling messages< it is necessary to ha'e a mechanism for differentiating packets and their associated protocols. The protocol discriminator field is pro'ided for this purpose. At present< only two message protocols are supportedG the ISDN signaling messages protocol and the le'el ;6packet protocol. The field has ; su$6fieldsG length su$6field< flag and the reference 'alue. The call reference field gi'es reference to the < % or D channel information transfer acti'ity to which a signaling packet pertains. Depending on the ser'ice and the channel used< the length of the call reference 'alue may 'ary. Stimulus signaling is used when the user and equipments are dum$ de'ices with no intelligence< like digital telephone. As the de'ices do not ha'e functional capa$ilities< stimulus6signaling messages are generated as a direct result of actions $y the terminal user. These signals 7ust indicate e'ents like handset off6hook or depression of a specific push $utton< which are all due to manual action $y the user.

44A

Network Le'el &ignaling:


Network *e'el signaling in ISDN is concerned with inter6office signaling. Signaling features accessi$le $y the user to o$tain enhanced ser'ices< from the network and other network related signaling. .ne of the main aims has $een to e'ol'e fle&i$le design for the signaling system to accommodate new ser'ices and connection types that may come a$out in the future to $e supported in the future. A$out ?5 network le'el messages ha'e $een standardi/ed so far and these messages may $e placed under H $road categoriesG

44B 4. ,orward address 8. :eneral Setup ;. ackward Setup ?. !all super'ision A. !ircuit super'ision B. !ircuit group super'ision E. In6call modification F. (nd6to6end H. +ser6to6user )essages $elonging to 4 to ? categories a$o'e are used to support the call setup process initiated $y the user and start the accounting and charging functions. !ircuit and circuit group super'ision messages permit $locking and de6$locking of circuit and circuit groups respecti'ely. .ther functions include connection release< temporary suspension and su$sequent resumption of circuits.

I&(N 3eco--endation: It descri$es a reference model for user6network


interfaces to the ISDN. The definitions of equipment in the 2eference model areG

44E

Network ,er-ination 1!N,1": The main function of this equipment is the


physical and electrical termination of the transmission line $etween the local e&change and the customer@s premises. .ther functions of the NT4 include maintenance and performance monitoring $y pro'iding digital loop $ack facilities< and the a$ility to feed D! power from the transmission line to other equipment in the installation.

Network ,er-ination 2 !N,2": This may $e a PA M< a local area network


0*AN1 or a terminal controller. The functions associated with an NT8 include protocol handling< multiple&ing< switching< concentration and other maintenance functions.

,er-inal $G+ip-ent !,$": A T( is a user equipment< typically a telephone


or data terminal< the functions of which include physical and procedural interfaces and maintenance< as well as the general communications function of the de'ice.

,er-inal $G+ip-ent 1 !,$1": A T(4 is a T( as defined a$o'e< and will $e


a digital telephone< data terminal< facsimile terminal or other workstation that complies with the ISDN user6network interface recommendations. :enerally it co'ers more modern equipments which ha'e $een specially de'eloped for ISDN operation.

,er-inal $G+ip-ent 2!,$2": A T(8 is a T( as defined a$o'e< $ut does


not conform to ISDN user6network interface recommendations. :enerally these will $e older types of equipment such as data terminals conforming to > or M interface specifications and group ; facsimile machines.

;ther t4pes of ter-inal eG+ip-entH pertaining to certain -an+fact+res1


New ISDN applications $eing introduced all the time. Typically new applications include *AN $ridges and ISDN P! cards to permit P! to P! $ulk file transfer. +sers of *ANs< operating on geographically dispersed sites are now a$le to transfer data $etween

44F each other. The transactions $eing transparent to the users who do not require knowledge of the location or address of the user to whom they wish to communicate.

,rans-ission Channels: There are ; types of fundamental channels in ISDN


around which the entire information transmission is organi/ed. 4. asic information !hannel

8. Signaling !hannel ;. %igh speed channel

Cideo Conferencing: (arlier< in order to achie'e accepta$le quality for a


'ideoconference< a leased digital link operating at 8)$it"s was required. Today such quality can $e achie'ed with digital circuits operating at rates as low as 48F 9$its"s. Techniques such as 'ideo and speech compression are used to produce a highly comple& digital signal in which the $andwidth occupied $y the 'ideo and audio signals are constantly changing. The added ad'antage of such an ISDN $ased >! system is that

44H there is no longer a requirement for a costly permanent leased circuit $etween sites< 'ideo conferences can simply $e dialed up when required and are charged on a pay as you go $asis.

$ncr4pted speech: The e'olution of ISDN has $rought with it the digital
telephone. It is now a relati'ely simple matter to produce a secure speech link $etween two users< $y introducing some form of encryption de'ice $etween the digital telephone and the channel o'er which it is to $e connected. D channel signaling messages are not encrypted< as they would then $e unreada$le $y the local e&change.

I&(N &tandards:
Standardi/ation is an essential process in the introduction of any ma7or and comple& international ser'ice. The capa$ility of pro'iding true international connecti'ity and interpreta$ility $etween networks is critically dependent on the a'aila$ility of standards and the strict adherence to them. The importance of standards has $een well recogni/ed in the conte&t of ISDN from the 'ery early stages. !!ITT has $een playing a leading role and acting as a coordinating $ody $y issuing ISDN related recommendations and there$y guiding the introduction of ISDN internationally.

485

The first definition of ISDN appeared in !!ITT@s recommendations issued in 4HE8. Su$sequent studies led to the emergence of the first ISDN standard in 4HF5. A !!ITT ser'ice is said to $e completely standardi/ed only whenG 4. (nd6to6end compati$ility is guaranteed. 8. Terminals to pro'ide the ser'ice is standardi/ed ;. Procedures for o$taining the ser'ice are specified ?. Ser'ice su$scri$ers are listed in the international directories. A. Testing and maintenance procedures are standardi/ed and B. !harging and accounting rules are spelt out.

8ser Le'el Interface


!omprehensi'e user network interface definitions are key to ensuring worldwide ISDN compati$ility. (&ample of an interface standard that ser'es us so well and yet goes almost unnoticed is the electrical power user interface. #e can purchase an electrical appliance almost anywhere in the world and plug it in our house socket. In ISDN< user network interfaces ha'e $een gi'en careful consideration to a'oid potential inconsistencies that may arise. ISDN caters to a 'ariety of ser'ices such as 'oice< data telemetry and image. In such a situation like this< one encounters conflicting requirements. .n one hand< a num$er of custom designed interfaces may ideally suit each ser'ice $ut would lead to a proliferation of interfaces. .n the other hand< one single multi6purpose interface may turn out to $e o'erkill for most of the ser'ices. 9eeping such factors in mind< two information rate access interfaces ha'e $een standardi/ed for ISDN.

484 4. asic rate access

8. Primary rate access

N+-2ering & Addressing


In telephone and data networks< the end Tequipments are more often single units than multiple de'ices units like PA M or *AN. %istorically< a telephone< a computer< or a terminal has $een the pre6dominant end equipment. The num$ering systems for these networks ha'e also e'ol'ed to identify single equipment end6points. In ISDN< multiple de'ices at the end points are more of a norm than single units< in 'iew of the multiple ser'ice en'ironments. It then $ecomes necessary to identify specific end equipment. ,or e&ample< computer of facsimile to render the ser'ice. Identifying this specific equipment is a two6le'el process= first the end6point is identified as in the case of telephone or data networks and then the equipment at the end6point. The component of the ISDN addresses which is used to identify the end6point. The component of the ISDN address which is used to identify the end6point is known as the ISDN num$er.

I&(N ser'ices are placed +nder two 2road categories:


4. earer Ser'ices 8. Tele Ser'ices oth the earer and Tele ser'ice functionalities may $e enhanced $y adding to

the $asic ser'ice< the functionalities of what are known as supplementary ser'ices. Supplementary ser'ices cannot stand6alone and are always offered in con7unction with either a earer ser'ice or a Tele Ser'ice.

488

3uestions
4. #hat is an Integrated Ser'ices digital networkI #hat are the types of ISDNI (&plain its functions 8. (&plain in $rief the ISDN StandardsI ;. (&plain the !oncepts of +ser Network InterfacesI ?. ring out the future applications of ISDNI

48;

!hapter E ST.2A:( A2(A N(T#.29S

Introduction enefits )anagea$ility .pen Standard Platforms Ad'anced Application !apa$ilities Ad'anced Storage )anagement

48?

&,;3A=$ A3$A N$,W;3@&


This technology is e'olutionary< and the demand for its applications is surging. Storage area networking promises to re'olutioni/e modern day network computing. ,rom a client network perspecti'e< the SAN en'ironment complements the ongoing ad'ancements in *AN and #AN technologies $y e&tending the $enefits of impro'ed performance and capa$ilities all the way from the client and $ack$one through to ser'ers and storage. 2apid growth in data intensi'e applications continues to fuel the demand for raw data storage capacity. Applications such as data warehousing< data mining< on6line transaction processing< )ultimedia< Internet and Intranet $rowsing ha'e led to a near dou$ling of the total storage capacity $eing shipped glo$ally on an annual $asis. And analyst predictions that the num$er of network connections for ser'er6storage su$systems will e&ceed the num$er of client connections are further fuelling the demand for network storage.

LI I,A,I;N& L;;

;C$3 &83=$ ;9 (A,A:

#ith the rise of client net working< data6centric computing applications and electronic communication applications< 'irtually all network6stored data has $ecome mission6critical in nature. This increasing reliance on the access to enterprise data is challenging the limitations of traditional ser'er6storage solutions. As a result< the ongoing need to add more storage< ser'e more users and $ackup more data has $ecome a monumental task.

48A

%a'ing endured for nearly two decades< the parallel Small !omputer System Interface 0S!SI1 $us that has facilitated ser'er6storage connecti'ity for *ocal Area Network 0*AN1 ser'ers is imposing ser'e limitation on network storage.

I-pending li-itations of e<isting network ser'er connecti'it4:


andwidth to ser'ice clients and maintain data a'aila$ility. Scala$ility for long term< rapid growth. ,le&i$ility to pro'ide optimum $alance of ser'er and storage capacity. )anagea$ility for ease of installation and maintaina$ility.

,he sol+tion: &torage Area Networking1


The Storage Area Network 0SAN1 is an emerging data communication platform< which interconnects ser'ers and storage at giga $aud speeds. y com$ining *AN networking models with the core $uilding $locks for ser'er performance and mass

48B storage capacity< SAN eliminates the $andwidth $ottlenecks and scala$ility limitations imposed $y pre'ious S!SI $us $ased architectures. In addition to the fundamental connecti'ity $enefits of SAN< the new capa$ilities< facilitated $y its networking approach< enhance its 'alue as a long6term infrastructure. These capa$ilities< which include compute clustering< topological fle&i$ility< fault tolerance< high a'aila$ility< and remote management< further ele'ate SAN@s a$ility to address the growing challenges of data6intensi'e< mission6critical applications. ,rom a client network perspecti'e the SAN en'ironment complements the ongoing ad'ancements in *AN and #AN technologies $y e&tending the $enefits of impro'ed performance and capa$ilities all the way from the client and $ack$one through to ser'ers and storage.

Benefits of the storage area network en'iron-ent:


%igh $andwidth. )odular scala$ility %igh a'aila$ility and fault tolerance. )anagea$ility. Total cost of ownership.

9i2re Channel: the open &AN sol+tion1


.'er the past year< ,i$re !hannel6Ar$itrat ed *oop 0,!6A*1 has emerged as the high6speed< serial technology of choice for ser'er6storage connecti'ity. )ost organi/ations prefer this solution $ecause of the widely endorsed open standards. This $road acceptance is attri$uted not only to ,!6A*@s high $andwidth and high scala$ility $ut also to its unique a$ility to support multiple protocols< such as S!SI and IP< o'er a single physical connection. This ena$les the SAN infrastructure to ser'e as $oth a ser'er6 interconnect and as a direct interface to storage de'ices and storage arrays.

48E

:igh Bandwidth
,!6A* pro'ides a 8.A to 456fold increase in effecti'e data $andwidth o'er the traditional parallel S!SI storages interface. Additionally< it offers future e&panda$ility. #hile the current ,,6A* standard for $andwidth is 4 giga $aud< planned enhancements to 8 and ? giga $aud gi'e ,!6A* a solid platform to address longer6term $and6width requirements.

&AN 's ,C; 2enefits &an 2enefits


!onnects the e&isting *ANs ,ully managed en'ironment Integrated fault tolerance Independently scala$le ser'ers and storage

,C; 2enefits
.ptimi/ing the e&isting in'estments )inimi/ed support cost )inimi/ed down time !omplements Network !omputer 0N!1 paradigm %ighly efficient scaling of resources

&er'er and storage scala2ilit4:


The modular scala$ility of ,!6A* is a key to ena$ling an infrastructure for long6 term growth and managea$ility. Traditional parallel S!SI $us connections ha'e $een limited to a total of E or 4A storage de'ices. AS $us $andwidth is pushed further and further this limit is compressed to e'er fewer de'ices per $us. In contrast< ,!6A* supports up to 48B nodes per loop with a typical configuration consisting of a com$ination of ser'ers and multi6disk arrays per node.

48F

Scala$ility in terms of capacity management and capacity $alancing is an area of significant differentiation $etween ,!6A* and S!SI. *argely dictated $y the limits on physical ca$le length< parallel S!SI storage connecti'ity requires close pro&imity to its host system< typically a ser'er. This translates to a single< integrated ser'er6storage enclosure that contains $oth ser'er processing power and one or two S!SI $uses of limited scala$ility.

Inter6dependent capacity scaling with integrated ser'er6storage model


+nder this single ser'er6storage enclosure model< the scaling of ser'er capacity and storage capacity $ecomes infle&i$le and inefficient. Single enclosures typically hold only ?645 dri'es. In order to scale the storage capacity $eyond this limit< additional ser'er6storage enclosures< including the cost of the ser'er processor $oard and peripherals is required. #ith a di'erse com$ination of data6intensi'e applications and ser'er processing6intensi'e applications running concurrently in the enterprise< the need for more fle&i$le and efficient scaling is needed. #ith less stringent ca$le length limitations< ,!6A* ena$les the networking of separate ser'er and storage enclosures within the SAN en'ironment.

48H

This capa$ility pro'ides a more fle&i$le and cost6effecti'e path for the independent scaling of ser'er performance and storage capacity< where either may $e e&panded independently to achie'e an optimum $alance.

)odular connecti'ity
In addition to superior fle&i$ility in scaling ser'er processing capacity and data storage capacity< the networking approach of ,!6A* introduces aspects of interconnect scala$ility that ha'e not $een possi$le with pre'ious architectures. Through the use of modular networking de'ices such as hu$s< switches< $ridges and routers< ad'anced SAN topologies can $e created to scale o'erall $andwidth< enhance a'aila$ility< ena$le ad'anced SAN application capa$ility< and ena$le ad'anced SAN application capa$ilities in storage management and load $alancing.

Ad'anced storage management


&torage -anage-ent challenge *ength of time required to $ackup data Ina$ility to $ackup< mirror or restore remotely *ack of alternati'es to local $ackup and mirroring. +se of *AN connections for ser'er $ackup consumes client network capacity &AN sol+tion andwidth and protocol efficiency accelerate $ackup !a$le length up to 45 km support remote operation. Ideal platform for distri$uted hierarchical storage management. Separation of ser'er6storage connections from *AN connections reduces *AN traffic.

4;5

:igh a'aila2ilit4 and fa+lt tolerance1


)any ,!6A* de'ices pro'ide features that ease the general deployment of fault6 tolerant SANs. .ne e&ample of these on $oard capa$ilities is the feature of dual porting< which has $ecome standard on ,!6A* disk dri'es< to facilitate dual loop configurations.

4;4 These dual loops pro'ide a redundant path to each storage de'ice in the array in case one of the loops is down or is $usy.

(+al loop arra4 config+ration1


The implementation of 2edundant Array of Independent Disks 02AID1 configuration is storage arrays and has $ecome a standard approach for fault tolerance and is fully supported $y the SAN en'ironment. In fact< to e'en further em$race the 2AID approach< ,!6A* disk dri'es pro'ide internal e&clusi'e6or 0M.21 logic< which effecti'ely pro'ides *e'el A 2AID capa$ilities from within the disk dri'e itself.

anagea2ilit4
>isi$ility down to the node and de'ice le'el is essential to case the efforts of installation< deployment and maintenance of any network. y em$racing a network management approach< SAN connecti'ity de'ices< such as hu$s and switches< ha'e integrated highly e'ol'ed management capa$ilities modeled after pro'en *AN and #AN management techniques. A fully managed SAN platform can offer monitoring and $ypass control of indi'idual nodes< loops< enclosures< storage de'ices< and connecti'ity de'ices.

;pen standards platfor-s for &AN -anage-ent1


S!SI command set. S!SI (nclosure Ser'ices 0S(S1 S!SI Self )onitoring Analysis and 2eporting Technology 0S.).A.2.T1 SA,6T(0S!SI Accessed ,ault6Tolerant (nclosures1 Simple Network )anagement Protocol 0SN)P1 #e$6 ased (nterprise )anagement 0# ()1. y em$racing the $est6practice network management standards esta$lished $y *AN and #AN platforms< information regarding SAN topology< status and alerts can $e easily $e accessed $y system administrators. It can also simplify remote system reco'ery and restoration in the e'ent of a failure. Traffic monitoring capa$ilities can also $e

4;8 em$edded into the SAN management system to facilitate sophisticated< cost6effecti'e load $alancing and capacity planning.

$ase of integration1
#ith ad'ance capa$ilities of networked ser'ers and storage< the a$ility to integrate SAN solutions into an e&isting network pro'ides tremendous 'alue in ease6of integration. Since the SAN en'ironment e&ists $ehind the ser'er< e&isting ser'er6*AN connections can easily $e le'eraged to facilitate a gateway $etween *AN and SAN< and allow the utili/ation of legacy ser'ers. The $road ca$ling options supported $y ,!6A* also ease the introduction of SAN s into e&isting campus networks. SAN connection distances up to 45km can $e achie'ed without the need to pull new ca$le. As a key $uilding $lock of SAN deployment< SAN connecti'ity de'ices offer dynamically configura$le< hot plugging capa$ilities. !om$ined with a graphical management interface< these features simplify trou$leshooting and accelerate installation.

,otal Cost of ;wnership


.ffering an infrastructure for cost6effecti'e< long6term growth< fault tolerance and managea$ility< the SAN en'ironment pro'ides Total !ost .f .wnership 0T!.1 ad'antages< which ha'e ne'er $efore $een possi$le with ser'ers <storage or ser'er storage connecti'ity.

Ad'anced application capa2ilities


y introducing the network like features of e&tended connection distance< IP support< and use of hu$s< $ridges switches and routers for comple& topologies < the SAN infrastructure ena$les a $road range of new capa$ilities includeG ad'anced storage management and ser'er storage clustering.

4;;

Ad'anced storage -anage-ent /ro2le-: Increasing amounts of network6storage data ha'e $ecome cum$ersome< if not
impossi$le< to maintain in a timely< secure< fault6tolerant and restora$le manner.

&ol+tion: The high $andwidth and topological fle&i$ility offered $y the SAN
en'ironment accelerates the data $ackup process and facilities new< inno'ati'e platforms for remote $ackup< mirroring and hierarchical storage. Perhaps the $iggest challenge facing storage management is the need to pro'ide efficient< secure< high a'aila$ility access to critical data. To efficiently o'ercome these challenges< a num$er of fundamental issues must $e addressedG The $andwidth and connecti'ity limitations imposed $y ser'er6to6storage parallel S!SI connections and ser'er6*AN connections offer little to address these formida$le tasks. Through its $andwidth< e&tended connecti'ity and transport efficiency< the SAN en'ironment uniquely offers a $road range of solutions for storage management< including remote $ackup< mirroring< reco'ery and distri$uted hierarchical storage management using a $road range of online and near6line storage de'ices.

4;? #ith the increasing comple&ity of networked computing systems and glo$al enterprise solutions it is refreshing when a single technology yields $oth unmatched performance and e&ceptional Total !ost of .wnership $enefits. In the case of ,i$re !hannel Ar$itrated loop and the rapidly de'eloping Storage Area Network< an e'olutionary open technology promises to re'olutioni/e the network centric< data6 intensi'e computing era through a new< inno'ati'e market space.

6+estions
4. #hat is Storage Area NetworkI #hat are its $enefitsI 8. #hat do you mean $y .pen Standard PlatformsI ;. ring out the ad'anced application capa$ilities of SANI ?. #hat is ad'anced Storage )anagementI

4;A

!hapter F !+22(NT T2(NDS IN !.)P+T(2 N(T#.29 *+( T..T% T(!%N.*.:J Introduction luetooth %istory

System !hallenges luetooth Security

The asic Structure luetooth for (m$edded Internet

The Need for luetooth

Introduction to Bluetooth

4;B luetooth is the radio technology that allows de'ices within ;5 feet of each other to communicate without wires. The luetooth technology eliminates the need for numerous and incon'enient ca$le attachments for connecting computers< mo$ile phones< mo$ile computers and handheld de'ices. All the things that can $e connected $y ca$le now can $e connected without it using luetooth technology. luetooth is not a new wireless *AN= it is something much simpler< more powerful and is a ca$le replacement. It is a Yradio $lockY that ena$les de'ices to talk to each other. It replaces the ca$les that traditionally 7oin pieces or equipment together. It makes them accessi$le $ehind walls and has capa$ility of connecting multiple units luetooth is the name gi'en to a new technology standard using short6range radio links< intended to replace the ca$le0s1 connecting porta$le and"or fi&ed electronic de'ices. The standard defines a uniform structure for a wide range of de'ices to communicate with each other< with minimal user effort. Its key features are ro$ustness< low comple&ity< low power and low cost. The technology also offers wireless access to *ANs< PSTN< the mo$ile phone network and the internet for a host of home appliances and porta$le handheld interfaces 0,ig. 41.

4;E

9ig+re 1G

Wireless connecti'it4 o'er Bl+etooth1


luetooth ena$led electronic de'ices connect and communicate wirelessly 'ia short6range< ad hoc networks called piconets. (ach unit can simultaneously communicate with up to se'en other units per piconet. )oreo'er< each unit can simultaneously $elong to se'eral Pico nets. These piconets are esta$lished dynamically and automatically as luetooth de'ices enter and lea'e the radio pro&imity. luetooth is further fueled $y the demand for mo$ile and wireless access to *ANs< internet o'er mo$ile and other e&isting networks< where the $ack$one is wired $ut the interface is free to mo'e. This not only makes the network easier to use $ut also e&tends its reach. The ad'antages and rapid proliferation of *ANs suggest that setting up personal area networks< that is< connections among de'ices in the pro&imity of the user< will ha'e many $eneficial uses. luetooth could also $e used in home networking

4;F applications. #ith increasing num$ers of homes ha'ing multiple P!s< the need for networks that are simple to install and maintain< is growing. There is also the commercial need to pro'ide Zinformation pushZ capa$ilities< which is important for handhelds and other such mo$ile de'ices and this has $een partially incorporated in luetooth. luetoothYs main strength is its a$ility to simultaneously handle $oth data and 'oice transmissions< allowing such inno'ati'e solutions as a mo$ile hands6free headset for 'oice calls< print to fa& capa$ility< and automatically synchroni/ing PDA< laptop< and cell phone address $ook applications.

Bluetooth History
luetooth was in'ented in 4HH? $y *. ). (ricsson of Sweden. The standard is named after %arald laatand Z luetoothZ II< king of Denmark H?56HF4A.D. A runic stone has $een erected in his capitol city Oelling 0Outland1 that depicts the chi'alry of %arald and the ZrunesZ sayG 4. %arald christeni/ed the Danes. 8. %arald controlled Denmark and Norway. ;. %arald thinks note$ooks and cellular phones should seamlessly communicate. The luetooth Special Interest :roup 0SI:1 was founded $y (ricsson< I )< Intel< Nokia and Toshi$a in ,e$ruary 4HHF< to de'elop an open specification for short6range wireless connecti'ity. The group is now promoted $y ;!.)< )icrosoft< *ucent and )otorola also. )ore than 4H55 companies ha'e 7oined the SI:. The following section descri$es some of the requirements from the luetooth system and in essence< suggests the functionalities planned for it.

System Challenges
The can $e used. luetooth system is now recogni/ed more than 7ust a ca$le replacement technology. >arious inno'ati'e usage models ha'e opened up new areas where luetooth

4;H

The most important requirement from the wireless link is that there should $e a uni'ersal framework that offers means to access information across a di'erse set of de'ices

In the practical scenario all de'ices are not e&pected to $e capa$le of all functionalities and users too may e&pect their familiar de'ices to perform their $asic functions in the usual way. So luetooth must offer the facility for colla$oration $etween de'ices< in the pro&imity of one another< where e'ery de'ice pro'ides its inherent function $ased on its form.

The standard must ena$le the de'ices to esta$lish ad hoc connections. Also< introduced is the unconscious connecti'ityZ paradigm< where de'ices can connect to those in pro&imity almost without any user command or interaction.

Support for $oth data and 'oice is e&pected as these are two most important kinds of information $eing transmitted o'er networks today. The standard should $e a$le to incorporate new usage models without requiring any registration of the new ser'ice with a central authority. The communications should offer similar protection as in ca$les. There should not $e any compromises on security in switching o'er to wireless. The implementations of the standard should $e simple< small and power efficient for easy mo$ile usage. It is necessary for the rapid deployment of the system and for the luetooth

$enefits to actually reach the users that a large num$er of de'ices $e ena$led with the luetooth standard. The de'ices to $e ena$led comprise a highly no uniform set and no single company can ha'e the e&pertise to manufacture all these.

Connection $sta2lish-ent in Bl+etooth


4. InG+ir4: The de'ice on reaching a new en'ironment would automatically initiated an inquiry to find out what access points are within its range. This will result in the following e'entsG a. All near$y access points respond with their addresses. $. The de'ice picks one out of the responding de'ices.

4?5 8. /aging: The de'ice will in'oke a $ase$and procedure called paging. This result in synchroni/ation of the de'ice with the access point< in terms of its clock offset and phase in the frequency hop. ;. Link esta2lish-ent: The *)P will now esta$lish a link with the access point. As the application in this case is email< an A!* link will $e used. >arious setup steps will $e carried out as descri$ed $elow. ?. &er'ice (isco'er4: The *)P will use the SDP0Ser'ice Disco'ery Protocol1 to disco'er what ser'ices are a'aila$le from the access point< in particular whether email access or access to the rele'ant host is possi$le from this access point or not. A. L2CA/ channel: #ith information o$tained from SDP< the de'ice will create an *8!AP channel to the access point. This may $e directly used $y the application or another protocol like 2,!.)). B. 39C; channel: Depending on the need of the email application an 2,!.)) or other channel will $e created o'er the *8!AP channel. E. &ec+rit4: If the access point restricts its access to a particular set of users or otherwise offers secure mode communications to people ha'ing some prior registration with it< then at this stage< the access point will send a security request for ZpairingZ. This will $e successful if the user knows the correct PIN code to access the ser'ice. F. ///: Assuming that a PPP link is used o'er serial modem as in dial up networking< the same application will now $e a$le to run PPP o'er 2,!.)). H. Network /rotocols: The network protocols like T!P"IP< IPM< and AppleTalk can now send and recei'e data o'er the link.

Bl+etooth &ec+rit4
The luetooth system is intended to $e used as a uniform interface to all of a personYs information sources and will thus $e e&pected to transfer sensiti'e personal data. Security of the data is thus understanda$ly an important issue. ,urther< luetooth de'ices are e&pected to $e omnipresent and at some places the access to these de'ices $y pu$lic users may ha'e to $e restricted. This calls for authentication procedures to $e pro'ided.

4?4 As the channel used is wireless and the packets $eing transmitted are a'aila$le to all mem$ers of a piconet< the security initiali/ation communications should not send any information that can allow an unauthori/ed de'ice to know the secret authentication keys. The application may itself encrypt its data for added security. That can add to the safety of the data< $ut the most of the authentication is $ased on the link le'el security procedures.

The basic structure


The procedures for security use four 'aluesG the de'ice address 0which is pu$lic1< a pri'ate authentication key 048F $its1< pri'ate encryption key 0F648F $its< configura$le1 and a random num$er. As the keys ha'e to $e secret< they cannot $e o$tained $y inquiry. The e&change procedures will $e descri$ed $elow. The security procedure requires a secret PIN to $e known to the user. 4. An initiali/ation key is generated using the PIN< the length of the PIN< a random num$er and the de'ice address. The dependence on the de'ice address makes it more difficult for a fraudulent de'ice to try a large num$er of PINs as each has now to $e tried with different de'ice addresses. 8. An authentication procedure is carried out using the challenge response scheme. The 'erifier unit sends a random num$er generated $y a specific process for the authentication. This random num$er is such that a claimant de'ice which has the correct initiali/ation key. ;. The claimant may also carry out 'erification on the 'erifier using a similar procedure as a$o'e. ?. (ach luetooth unit has a unit key< installed in its non 'olatile memory. The de'ice now uses the initiali/ation key to encrypt this unit key and sends it to the other de'ice. A. The second de'ice may add its own unit key to the unit key of the first de'ice and generate a com$ination link key if $oth the de'ices are capa$le of handling this.

4?8 An encryption key is now generated from the link key< a random num$er and another num$er o$tained from a fi&ed procedure. oth the de'ices can generate this encryption key as all the required information is known to $oth de'ices.

Bluetooth for Embedded Internet


HF[ of the computing de'ices 0microprocessors and microcontrollers1 sold today are em$edded products and only the remaining small fraction consists of general purpose microprocessors used in P!s or workstations. Not 7ust electronic equipment like 'ideo players< music systems or telephones $ut e'en mundane consumer goods like washing machines< dishwashers< o'ens and toasters now ha'e an em$edded processor sitting $ehind the control panel. This re'olution has come a$out due to the e'er increasing num$er of transistors $eing packed into a smaller and smaller area of silicon ena$ling high computational powers to $e pro'ided at 'ery low cost. !om$ine this with the increasing proliferation of wired and wireless networking which has completely transformed the way information flows around us. The con'ergence of the a$o'e two technologies is leading to what is called the Zem$edded internetZ6 the immense new 'alue that is emerging $y connecting these computational components. The internet will not $e restricted to $eing a newtork of P!s and the like< $ut will now include all intelligent de'ices located in the human en'ironment.

:ow far is it0


!heap microcontrollers today are 7ust capa$le of supporting an em$edded operating system< the T!P"IP stack and run a #e$ ser'er $ased on the omnipresent %TTP. (&amples of thin ser'ers a$ound a nota$le one $eing the ,airchild A!(4454)TF processor supported fingernail si/ed we$ ser'er. #hat is now needed is for this de'ices to support ser'er side programming or client ser'er computational models that can ena$le these de'ices to process e&changed data. Themost o$'ious mechanisms for this are the cgi interface and Oa'a applets. Thin ser'ers can almost run cgi and (m$edded Oa'a is

4?; $eing de'eloped too. OINI is another ser'ice that is $eing de'eloped for similar applications.

!omputation power alone is not sufficient to create real world utility. The processing has to $e on physical data and the output has to $e used $y physical de'ices. This requires sensors to pick up information and actuators to $ring a$out the desired changes. >arious efforts are on this direction. )()S or )icro electromechanical systems are a hot area of research and are soon e&pected to pro'ide us usa$le and cost effecti'e sensors and actuators which can $e deployed rapidly for the purpose. The :PS 0:lo$al Positioning System1 can now $e accessed from e&tremely small de'ices< like those $uilt into watches or PDAs. Passi'e or $attery6less electronic tags ha'e $een successfully used and are already in commercial use< for instance at music stores to pre'ent theft.

,he Need for Bl+etooth1


#ireless is important for the em$edded de'ices to $ecome really u$iquitous. This throws up certain issues like low power consumption< connections without user interaction< a$ility to route data on an ad hoc $asis and their related addressing issues. Data security and access control can not also $e neglected #ith its e&tensi'e support for integration with e&isting protocols and APIs< luetooth seems to $e the ma7or contender among other such wireless solutions for the physical layer connecti'ity. The em$edded internet $ased on re'olutionali/e our li'ing and work en'ironments today. luetooth seems all set to

4??

WI3$L$&& A//LICA,I;N /3;,;C;L


#AP stands for #ireless Application Protocol. The popularity of digital wireless user 6 agents has $een staggering growth in recent years with a massi'e glo$al increase in the use of mo$ile phones. The addition of further capa$ilities mans that the mo$ile phone is no longer merely a telephone $ut a communication de'ice capa$le of running applications and communicating with other de'ices and applications o'er a #ireless Network. #AP is the de'elopment of esta$lished internet protocols and concepts intended to standardi/e the way in which pages< mo$ile phones< and personal digital assistants access information and ser'ices.

Li-itations of WA/
There are some limitations to #AP de'ices and the main aspects $eingG6 4. Small display monitor 8. *imited processing power and memory ;. *imited $attery life and power ?. *imited data input and users interaction capa$ilities A. *imited $andwidth and connection speeds. B. ,requent unsta$le connections

WA/ 9or+-:
A forum was formed in 4HHE $y the leading mo$ile phone manufactures like (ricsson< )otorola< Nokia and Phone.com and is called the #AP forum. #ithin two years< more than 455 companies 7oined the group to define the standards for pro'iding internet content and ser'ices to wireless de'ices. #AP is actually not a single protocol= rather< is a collection of protocols and standards that make up a complete lightweight protocol stack along with special markup and scripting languages< which together define a complete solution

4?A

(e'ices +sed in WA/


Some of the #AP de'ices are hand6held6digital6wireless de'ices such as mo$ile phones< pagers< two6way radios< smart phones< and communicators6from low6end to high6 end. The ase Station Switching !enter is the control element for the $ase transmitter TS. Thus in a dense metropolitan area< S! switching site. stations< $ut need not $e co6located with the

se'eral antenna sites may $e used< $ut they require only one small

#hene'er the mo$ile handset is switching on and at regular inter'als thereafter< it uses the control channel to register itYs presence to the nearest mo$ile switching center. The mo$ile switching centers are the main controlling elements of the networks. (ach control has a gi'en geographic area o'er which a num$er of TS are spread. The information is held $y the home )S!s in a data$ase called the %ome *ocation 2egister or %*2. The local )S!s duplicates some of this information in a temporary 'isitor location 2egister or >*2< until the caller lea'es the )S!s are. The telephone networks are circuit switched networks.

;& co-pati2le with WA/


#AP is designed work with almost all wireless networks and application en'ironments. It can $e $uilt on any operating System including #indows< .S"H< Oa'a .S< etc. It pro'ides ser'ice interopera$ility e'en $etween different de'ice families.

WA/ Browsers
It runs on the #AP de'ice and displays the contents it recei'es. It also pro'ides the front6end< through which the user can na'igate the #AP application. The $rowser may $e $uilt into the phone or mo$ile de'ice< or into the SI) card< the de'ice contains. Some of the #AP $rowsers currently a'aila$le are gi'en in the $o&

:ow the WA/ protocol Works0


A simple #AP application consists of files< located on a we$ ser'er< written in #ireless )arkup *anguage 0#)*1 and possi$ly script files written in #)* script and graphics files in #)* itmap format. The #AP follows the steps mentioned $elowG6

4?B 4. The +ser presses a phone key that has a +2* assigned to it. 8. The phone sends a +2* request to a #AP gateway using the #AP protocol ;. The gateway creates a con'entional %TTP request for the specified +2* and sends it to the #( ser'er. ?. The %TTP request is processed $y the ser'er. The +2* may refer to a static #AP file or may use a !:I script to create the #AP content. The ser'er will fetch the file and add an %TTP header to it< or if the +2* specifies a script application< the ser'er will run the script. A. The we$ ser'er returns the #)* content with the added %TTP header. B. The #AP gateway 'erifies the %TTP header< and the #)* content< then encodes them into $inary form. The gateway then creates a #AP response containing #)* and sends it to the phone. E. The Phone recei'es the #AP response and processes the #)* to display the appropriate content.

WA/ &+--ar4:
4. #AP does for wireless de'ices that %TTP does for we$ $rowsers 6 it allows them to $ecome clients in an Internet6$ased client"ser'er world. 8. #AP is a protocol< a data transport mechanism. In many ways it is similar to %TTP and #AP was also $uilt on top of esta$lished standards< such as IP< +2*s< and M)*. ;. #AP is not a single protocol= rather< it is a collection of protocols and standards that make up a complete lightweight protocol stack along with special markup and scripting languages which together define a complete solution.

4?E ?. #AP forum is the industry association comprising of hundreds of mem$ers that ha'e de'eloped the de facto world standard for wireless information and telephony ser'ices on digital mo$ile phones and other wireless terminals. A. #AP de'icesG %andheld digital wireless de'ices such as mo$ile phones< pagers< two6way radios< smart phones< and communicators 6 from low " end to high " end. B. The #AP $rowsers run on the #AP de'ice and display the contents it recei'es. It also pro'ides the front6end< through which the user can na'igate the #AP application. The $rowser may $e $uilt into the phone of mo$ile de'ice< or into the SI) card.

Introd+ction to W L
In Oune 4HHE< phone.com originally known as unwired planet along with Nokia< )otorola and (ricsson launched the #AP forum 6 a nonprofit organi/ation dedicated to the de'elopment and proliferation of a single standard protocol for wireless application. +sing phone.comYs %D)* 0%andheld de'ice markup language1 as the $asis for its own standard markup language< the ,orum created and distri$uted #)* 6 a language different form< $ut in many respects similar to %D)*. #)* is a markup language used for descri$ing the structure of documents to $e deli'ered to wireless de'ices. #)* is to wireless $rowsers as %T)* is to a $rowser on a desktop computer. #)* was created to address the display $andwidth and memory limitations of mo$ile and wireless de'ices such as cellular phones and wireless handheld computers. Since #)* uses an M)* 'oca$ulary< it could $e useful to understand some $asic principles of M)* 0(&tensi$le )arkup *anguage1< a tag6$ased system used for defining< 'alidating and sharing document formats. Although they are 'ery similar< #)* differs from M)* in the following waysG 4. #)*Ys white6space handling rules are not as ela$orate as M)*Ys. 8. #)* relies on well6formed e&pressions. ;. #)* has a $uilt6in method for handling international characters.

4?F

&ec+rit4 in WA/
Security in #AP has $een implemented in such a way to pro'ide ma&imum $enefits with little or no hassles. Security on the internet is pro'ided at a num$er of le'els through the in'ol'ement of 'arious protocols< the most common of which is the Transport *ayer Security Protocol T*S formerly known as secure socket *ayer 0SS*1. #AP implements most of its security in wireless transport layer security protocol< $ased on T*S with su$tle differences. #T*S is capa$le of running o'er #ireless Data6gram protocol or +ser Data6gram Protocol.

WA/ & Internet:


*et us 'iew how #AP differs from internet. In this model< connection is esta$lished with the #AP gateway through the network operator rather than through the ISP. The phone call is routed through the network operator@s modem to a 2emote Access Ser'er 02AS1. There is also a le'el encryption. The 2AS ser'er also performs authentication and routes the data to a #AP gateway. This is not the feature in the regular process of internet communication< The #AP gateway then con'erts the #)* script to and from the $inary format that is transmitted o'er the air and passes on the data to the we$ ser'er using %TTP protocol. The #AP forum defined a new protocol #T*S that is $ased on T*S and pro'ides a similar le'el of security. #AP utili/es a security certificate in order to present the pu$lic"pri'ate key pair generated once for the client to the #AP gateway and secure the #T*S layer for authentication. *imitations of #ireless de'ices are the display of mo$ile phones is 'ery small and na'igation poses a pro$lem. !urrently most of the mo$ile de'ices are phones and the only input facility a'aila$le is the keypad. #AP de'ices are $asically mo$ile phones and they ha'e limited Processing Power and 2A). #AP de'ices ha'e 'ery little $andwidth as compared to that of a P!. Pro'iding users with graphics when they are using is more difficult and the deck si/e is small.

4?H

I-portant Aspects of WA/:


4. #AP is a collection of languages and tools and an infrastructure for implementing ser'ices for mo$ile phones. 8. #AP introduces a gateway $etween the phones and the ser'ers pro'iding content to the phones. ;. The #AP gateway talks to the phone using the #AP protocol stack< and translates the requests it recei'es to normal %TTP. ?. Authentication is the process of making sure that another party is actually who they claim to $e. A. (ncryption is the process of encoding information in to a different format that cannot $e easily understood and only the intended recipient understands. B. !ryptographyG The art of keeping messages hidden or sure. E. #AP implements most of its security in #T*S< $ased on T*S with su$tle differences. #T*S is capa$le of running o'er #DP or +DP. F. (a'esdropper attack< impersonation attack< man in the )iddle attack is a few threat models.

Con'erting $<istent We2 sites to WA/:


y con'erting e&isting sites i.e. %T)* content to #)*< %T)* is the most common form of te&t on the we$ and the content con'erters are also known as Ztrans codersZ. Some #AP gateways do this automatically. !on'erters work $y e&tracting te&t from a source page< then re6formatting that te&t in to the target markup language< in this case #AP. The con'erter is performing the con'ersion of formatted data to pure date< so we< as the con'ersion author decide the format we want the output to $e in. The

4A5 intermediate data can $e manipulated without $ack6end and front6end processes affecting that manipulation. #e can either e&tract all the possi$le contents in the page such as title< welcome messages< and links and so on or e&tract specific parts of the page say 7ust the news headlines< or 7ust the stock quotes.

Introd+ction to $?-ail:
(mail is an asynchronous message e&change technology. This simply means that when you send an (6mail message the recipients doesn@t ha'e to $e a'aila$le at that instant to recei'e the mail< $ut may collect the message at his own leisure. ,rom the users point of 'iew e6mail is sent 'ia S)TP< collected from their mail$o& using P.P; or I)AP< and any address $ook information is searched for using *DAP or A!AP.

What WA/ & $?-ail can offer0


The popularity achie'ed $y 'ery limited short messaging technology 0S)S1 indicates the demand for messaging 'ia mo$ile phones certainly e&ists and gi'ing mo$ile phones all the functionality of e6mail definitely seems to $e the ne&t logical step. (6mail is su$stantially more ad'anced technology than S)S< e'en if it is only used for simple S)S like that messages. )essage recipients are not limited in how they recei'e their messages. 2ather than only $eing a$le to access the messages from a single mo$ile phone< the user can use any e6mail client he prefers. #AP de'ices and e6mail capa$ilities seem to $e an ideal technological fit. Since< they allow for useful synergy of personal communication technology< deli'ering the con'enience of porta$ility from mo$ile phones< while allowing instant access to e6mail< pro'iding asynchronous access to written messages.

4A4

/+sh ,echnolog4:
The internet user pulls the content from the network. There is a lot of information that is a'aila$le and needs to $e pushed to the user at a certain predefined inter'al or notify the users when certain important e'ents occur. ,or e.g. Tourist or hotel information can $e pushed to wireless de'ice users in a particular area. The push technologies help us to pro'ide this functionality to a #AP user. This technology is already in e&istence in the mo$ile phone networks using S)S and cell $roadcast mechanism in :S) networks $ut they lack an important feature< interacti'ity.

/+sh 9ra-ework:
Push architecture consists of client ser'er architecture. The ser'er ha'ing the potential of push initiator< #AP client can listen for push requests. The push initiator sends an instruction to a pro&y gateway which $roadcasts the command to wireless networks using the Push6o'er6the6air protocol< which shall $e discussed later. The message is $asically M)* packets. The contents are 'ery $rief< containing a message followed $y a link to a we$6site. The Push Access Protocol 0PAP1 is designed to work on the top of one of the application le'el protocols like %TTP or S)TP on the internet Push6 .'er6the6air 0.TA1 protocol is used on top of the #SP layer of the #AP stack of protocols. The Push6pro&y6gateway is placed $etween the push origin ser'er 0PI1 and the #AP client. It has to implement the entire PAP protocol stack plus PAP and .TA.

4A8

9+t+re in WA/
4. It pro'ides the user with permanent connecti'ity< remo'ing one of the ma7or frustrations of :S)< namely dropped connections and the incon'enience and delay of ha'ing to dial up repeatedly to perform a #AP $ased transaction or interaction o'er :S)< or indeed any other circuit6switched network. 8. :S) is a circuit6switched technology. .n the other hand< packet switched technologies6such as :P2S and ;:6allow users of mo$ile de'ices to esta$lish a connection with their $arrier< which is then maintained indefinitely. ;. (D:( stands for (nhanced Data 2ates for :lo$al ('olution. (D:( is a further enhancement of :S) $ased technology and may e'entually offer data transmission rates that match those of ;: networks. ?. luetooth offers ine&pensi'e< easy to $uild and use< low power consumption< wireless communication o'er short distance $y means of small radio chips. luetooth< like a num$er of other key technologies such as 'oice recognition< impro'ed displays and key $oards< will make the user e&perience more con'enient and rewarding for wireless de'ices. A. The (P.!;8 operating system designed and $uilt for mo$ile computing is no dou$t one of the $est platform contenders for wireless client de'ices in terms of its capa$ilities and architecture. B. In future we are likely to ha'e a porta$le de'ice< which can $e called a #ireless information de'ice 0#ID1< which is going to $e far smarter than anything currently a'aila$le. +nified messaging< com$ining 'oice< e6mail< 'ideo6mail< fa& and any other messaging ser'ice imagina$le will $ecome a reality. There will $e many slips and stum$les along the way for many of these things to $e reali/ed. %owe'er< we can see the foundation technologies< ideas and ser'ices all around us.

4A;

#ireless *ANs
T. +N#I2( AN (NT(2P2IS(
#ireless *ANs 0#*ANs1 pro'ide fle&i$le connecti'ity as an e&tension< or an alternati'e to a wired *AN within a $uilding or a campus. #*ANs are usually used to connect handled terminals and note$ook computers to e&change real6time data with enterprise applications on the corporate $ack$one. These networks are growing popular in 'ertical markets for applications related to health6care< consulting and sales< retail< manufacturing< and education and research. The #ireless *AN 0#*AN1 market is likely to grow to a$out +SX ; $illion $y 8558< according to !ahner@s In6Stat :roup. #*ANs augment wired *ANs< making it possi$le to access shared information within the campus without needing to physically connect to the network. There is no need to e&tend the e&isting ca$ling or to configure additional nodes. And the enhanced mo$ility pro'ides producti'ity and ser'ice opportunities that are otherwise not possi$le.

A//LICA,I;N&:
.ften #*ANs pro'ide the last few meters of connecti'ity to the corporate $ack$one within a campus. Take a look at some of the applications a'aila$ility of information has greatly enhanced their efficiency. Trade shows and product demonstrations make great use of #*ANs for pro'iding temporary connecti'ity. #*ANs are 'ery effecti'e in rapidly changing connecti'ity scenarios $ecause they make mo'es< additions and changes the network much easier.

4A? #arehouse workers roaming around the warehouse e&change information with the central data$ase o'er #*ANs< #*ANs are also $eing used as $ack6ups for wired *ANs in mission6critical applications. Teams meeting in corporate conference rooms make quicker decisions with immediate access to real time information o'er #*ANs. #*ANs are of great help to the ser'ice industry< such as restaurants< car rentals< and so on6$ecause the a'aila$ility of real time information is 'ery 'ital to the efficiency of this industry.

4AA

AI;3 B$N$9I,&: I-pro'ed prod+cti'it4 with -o2ilit4: Access to real time information anywhere
in the organi/ation makes possi$le higher le'els of ser'ice.

$ase of installation: #*AN installation is so much easier $ecause there is no need to


draw ca$les. The #*AN reach is also much wider than that of wired *ANs.

Lower cost of ownership: Although the initial in'estment in #*ANs may $e more
as compared to wired *ANs< the cost of ownership o'er the entire life cycle< keeping in 'iew the frequent mo'es< is significantly lower.

&cala2ilit4: #*ANs are highly scala$le as they can $e set up in a 'ariety of topologies
to meet specific requirements.

WLAN ,$C:N;L;=I$&G
#*ANs use radio or infrared 0I21 wa'es to communicate information from one point to another. In a typical #*AN configuration< a transmitter"recei'er de'ice< called an access point< interfaces with the wired network using standard ca$ling. The access point $uffers and transmits data from the wireless *AN to the wired networks< A single access point can support a small group of users within a few hundred feet. The antenna of the access de'ice is mounted at a location to pro'ide radio co'erage in the desired area. A num$er of such access points along the wired network augment the reach of the wired network.

4AB At the user end< the handled de'ices ha'e a #*AN adapter< which interfaces with the operating system of the de'ice and the airwa'es 'ia an antenna. Typically< a #*AN can pro'ide throughput to the order of 4644 )$ps.

Wireless LANs can operate on an4 of the following technologies: Narrow2and radio system transmits and recei'es user information on a specific radio frequency. (ach user operates on a different frequency. &pread spectr+- is used $y most wireless *ANs. It is a wide6$and radio frequency technique de'eloped for relia$le< secure mission6critical communication systems. Infrared uses 'ery high frequency< 7ust $elow the 'isi$le light in the electromagnetic spectrum< to carry data. Oust like light< I2 cannot penetrate opaque o$7ects. It is either directed line6of6sight or diffused. Ine&pensi'e line6of6 sight1 systems may pro'ide a 'ery limited range suita$le only for personal area networks. %igh6performance I2 systems may $e impractical for wireless users and may $e used to implement fi&ed su$6networks using line6of6sight. Wh4 sho+ld 4o+ deplo4 WLANs0 #*ANs pro'ide tremendous fle&i$ility< scala$ility and mo$ility. Some reasons why they should $e deployed areG Area of co'erage: ased on the power of the equipment< an entire indoor area can $e co'ered using #*ANs. The range 'aries from 455ft to A55ft. )icro cells created $y using access points can increase.

4AE

3elia2ilit4: Though it may seem that radio interference would downgrade the performance of #*ANs< sturdy designs and the limited distance o'er which a #*AN has to operate ensures ro$ust connections. These connections can often $e more relia$le than wired *ANs. Interopera2ilit4: #*ANs seamlessly integrate with wired *ANs< including (thernet and Token 2ing. Interopera$ility $etween #*ANs is $ecoming easier with standardi/ation. Industry standards like I((( F58.44 make it possi$le for #*ANs from different 'endors to work together. Interference: Since the radio frequencies used $y #*ANs may not $e licensed< there is a possi$ility of #*ANs interfering with some other de'ices like microwa'e o'ens. )ost 'endors of #*ANs design their products to take care of this interference. Costs: !osts include the cost of wireless access points and the wireless *AN adapters. The num$er of access points depends on the si/e of the area that is to $e co'ered. The price of access points ranges from +SX B55 to +SX 4A55 #*AN adapters cost $etween +SX 4A5 and +SX A55. ut #*ANs sa'e on the cost of ca$ling and the cost of implementing changes to the network. &afet4: The output power of #*AN equipment is much less than that of a handheld cellular phone. Since radio wa'es fade 'ery rapidly o'er distance< e&posure to 2, energy

4AF to the people in the 'icinity is 'ery little. No ill effects on health ha'e $een attri$uted to #*ANs. &ec+rit4: #ireless technology has its origin from the military. Security pro'isions are typically $uilt into #*ANs< often making them more secure than most wired networks. !omple& encryption techniques make ea'esdropping e&tremely difficult. Integration with e<isting applications: #hile planning access to the wireless infrastructure< e&isting applications should not $e disrupted or rede'eloped. Towards this< #)* could $e used for faster deployment. ,or more fle&i$le and maintaina$le systems< M)*6$ased architecture is recommended. Depending upon specific $usiness needs< it@s perhaps time to $uild a 'ery scala$le and fle&i$le #*AN solution to suit your corporate requirements.

3uestions
4. #hat is lue tooth TechnologyI ring out the $rief history of lue tooth technologyI 8. #hat are the System !hallenges - Security aspects in lue toothI ;. (&plain how lue tooth is essential for em$edded InternetsI ?. #hat are the needs for lue tooth TechnologyI (&plainI A. #hat is a #API #hat are its limitationsI #hy #AP forumI B. #hat are #AP ProtocolsI :i'e a small note on #)*. E. ring out the concept of Push Technology - Push ,rameworkI #*ANI H. (&plain the Technological aspects of #*AN. 45. #here is the need to deploy #*ANI F. #hat is #ireless *ANI #hat are its limitationsI ring out the ma7or $enefits of

4AH

3eference Books:
4. Data !ommunications - Distri$uted Networks #yless D lack 8. *ocal Area Network Architectures Da'id %utchison ;. (ncyclopedia of Networking Tom Sheldon ?. Integrated Digital Network *S *awton A. Integrated Ser'ices Digital Network Oohn *ane B. ISDN Tutorials ISDN Jahoo Search (ngine