• Easy to install and configure.
• Individual machines do not depend of a dedicated server.
• Users are able to control tehir own shared resources.
• Inexpensive network to purchase and operate.
• You don’t need other equipment other than an operating system, NIC and cables.
• You don’t need a dedicated employee to act as a network aministrator.
• Suited for network smaller than 10 users.

• You can apply security only to one resource at the time.
• Users might have to remember as many passwords as there are shared resources.
• You must perform individual backups on each machine to protect all shared data.
• When someone accesses shared resources, the machine where the resource resides suffers a performance hit.
• There is no centralized organizational scheme to locate or control access to data.

• They provide centralized user accounts, security, and access controls, which simplifies network administration.
• More powerful equipment means more efficient access to network resources as well.
• User only have to remember a single password for network login, which allows them to access all resources that they have permission to access.
• Server-based networks are scalable.

• A server failure can render a network unusable; at best, it results in loss of networks resources.
• Such networks require an expert stafff to manage the complex, special-purpose server software, which adds to the overall cost.
• Cost also increase due to the requirements of dedicated hardware and specialized software.

Modèle hierarchique car chaque couche offre certains services aux couches supérieures.

La couche physique assure la transmission des données sous forme de signaux électriques sur un circuit de communication.

-L'inité d'information manipulée est le bit

La couche liaison est responsable de la transmission sans erreur de blocs d'information sur des liaison de données. La couche de liaison est subdivisé en la sous-couche Logicla Link Control et la sous-couche Media Access Control.

-Forme les trames

-Résous les problèmes de trames endommagées, perdues et dupliquées.

-Exerce le contrôle de flux qui consiste à contraindre l'émetteur à ne pas envoyer plus de données que le récepteur ne peut en accepter.

-L'unité d'information manipulé est le trame.

Protocole de ce niveau: BISYNC, HDLC, LLC, SDLC, LAPB.

La couche réseau est responsable de l'acheminement des paquets de données qui transistent à l'intérieur du réseau.

-Routage des paquets

-Controle de congestion pour éviter des perte de paquets par engorgement de certains chemins.

La couche transport gère le transport des informations de bout en bout au travers du réseau.

-responsable de l'émission, découpage des données en plus petit paquet et à la reception réassemble les paquets

-Services supplémentaires de traitement d'erreurs

-Controle de flux de bout en bout.

-C'est la première couche de haut niveau qui assure la transition entre les couches de traitement 5, 6, 7 et celles de transmission 1,2,3.

-Protocole TCP/IP et ITP (internet transport protocol)

La couche session est responsable de la mise en place et du controle du dialogue entre des utilisateurs sur différentes machines.

-gestion du dialogue

-synchronisation

mapping d'adresse

La couche présentation est responsable de la représentation des données échangées par les applications.

-encogdage des données

-compression de données

-chiffrememnt (encryptage)

-EX: ASCII, EBCDIC,DES (Data Encryption Standard)

La couche d'application fournit les protocoles permettant aux utilisateurs d'accéder au réseau

-Compréhension et exécution des commandes

-Normalisation des bases de données et des commandes de système d'exploitation.

-Terminal virtuel, X400 pour la messagerie électronique.

Network protocols provide the following services : Addressing and routing information, error checking, requesting retrnasmissions, and establishing rules for communicating in a particular netwroking environment. These service are called link services.

DDP (Delivery Datagram Protocol) : Apple’s data transport protocol that is used in AppleTalk.

IP (Internet Protocol) : Part of the TCP/IP protocol suite that provides addressing and routing information.

IPX (Internetwork Packet Exchange) : Novell’s netware protocol used for packet routing and forwarding.

NetBEUI : It provides transport services for NetBIOS.

Responsible for ensuring reliable data delivery between computers.

ATP (AppleTalk Transaction Protocol) and NBP (Name Binding Protocol) : AppleTalk’s sessions and data transport protocols.

NetBIOS/NetBEUI : NetBIOS establishes and manages communications between computers; NetBEUI provides data transport services for that communication.

SPX (Sequenced Packet Exchange) : Novell’s connectio-oriented protocol that is used to guarantee data delivery.

TCP (Transmission Control Protocol) : The portion of the TCP/IP protocol suite that is responsible for reliable delivery of data.

Responsible for application-to-application services.

AFP (AppleTalk File Protocol) : Apple’s remote file management protocol.

FTP (File Transfer Protocol) : Another member of the TCP/IP protocol suite that is used to provide file transfer services.

NCP (Netware Core Protocol) : Novell’s client shells and redirectors.

NFS (Network File System) : A client/server file system protocol primarily used to share directories with UNIX systems.

SMB (Server Message Block) : A protocol that sits above the NetBEUI and NetBIOS that defines and formats commands for information passing between networked computers.

SMTP (Simple Mail Transfer Protocol) : Member of the TCP/IP protocol responsible for transfering mail.

SNMP (Simple Network Management Protocol) : Member of the TCP/IP protocol that is used to manage and monitor network devices.

TCP/IP is the most widely used protocol suite in networking today. This is due in part to the vast growth of the global Internet. TCP/IP is able to span wide areas and is very flexible. In addition, it provides cross-platform support, routing capabilities, as well as support for the Simple Network Management Protocol (SNMP), the Dynamic Host Configuration Protocol (DHCP), the Windows Internet Name Service (WINS), the Domain Name Service (DNS), and a host of other useful protocols. However, TCP/IP's rich set of features are provided at the expense of additional overhead, which may make it too cumbersome for some networks or applications.

It should come as no surprise that the AppleTalk protocol is used for communication with Macintosh computers. By enabling AppleTalk, you allow Mac clients to store and access files located on a Windows NT Server, print to Windows NT printers, and vice versa. An item of note: You must first install the Windows NT Services For Macintosh before you, can install AppleTalk. Also, Mac support is only available from an NTFS partition.

The Advanced Program-to-Program Communication (APPQ protocol, developed by IBM, is a peer-to-peer protocol used in IBM's Systems Network Architecture (SNA) for use on AS/400-series computers.

X.25 is a set of wide-area protocols that are used in packet- switching networks. It was created to connect remote terminals to mainframes. Although many other wide-area communications types are available in the United States, X.25 is still widely used in Europe.

High-level Data Link Control (HDLQ is a flexible, bit-oriented data link protocol that is based on IBM's Synchronous Data Link Control (SDLQ. It has been standardized by the ISO. HDLC can support half- or fullduplex transmission, circuit- or packet-switched networks, peer-to-peer or client/server networks, and transmission over cable or wireless media.

The Xerox Network System (XNS) was created by Xerox for use in Ethernet networks. XNS is the basis for Novell's IPX/SPX, but it is seldom found in today's networks.

UTP and STP have the same 100 m cable length restriction.

Network Device Network Specification (NDIS) as its implementation of the device interface concept, specifically to reside between the NIC driver (Media Access Control) and the protocol stack in the Data Link layer of the OSI reference model. NDIS allows for the binding of multiple protocols to a single NIC, or binds multiple protocols to multiple NIC.

Novell and Apple developped Open Datalink Interface (ODI) which is Novell’s implementation of NDIS.

Transceiver : On a NIC, it is the physical interface between itself and the network.

• Unité physique qui amplifie le signal électrique seulement.
• Opère au niveau de la couche physique

• 3 types de hub : Actif, passif et intelligent.
• Opère au niveau de la couche physique.

Bridges read the target destionation’s MAC address from each incoming data packet, the examine the bridging table to determine what to do with the packet.

The bridge functions bassically as a repeater.

Translates different mediaq-access methods, allowing the translation bridge to link various network types. (Ethernet and Token-ring network).

Learns over time where to direct packets it receives. It does this continually building bridging tables, adding new entries when they become necessary.

• Le pont effectue des transformations mineures dur les trames
• Servent à la connexion des réseaux locaux IEEE 802.X
• Agit au niveau de la couche liaison de données

Allows for communications between dissimilar systems on the network. Can also translates protocols.

Effectue des transformations mineures sur les trames pour rendre les formats de trames compatible. IEEE 802 et X25

2 modes de connexion : connexion (Établissement de circuits entre la source et la destination), et sans connexion (Par datagramme)

Agit au niveau de la couche réseau et supérieure.

Functions at the Network layer and can link 2 or more network segments (or subnet). It uses the network address information found in the Network layer area of the data packet. After obtaining this address information, the router uses the routing table of network address to determine where to forward the packet. There are 2 types of routing devices:

Static routers: Use routing tables that a network administrator must create and update manually.

Dynamic routers: Build and update their own routing tables.

It's a bridge and a router. When a brouter receives a packet, it checks to see if the packet was sent in either a routable or a non-routable protocol. If it is a routable protocol packet, the brouter will perform a routing function, sending the packet to its destination outside the local segment, if necessary.

In contrast, if the packet contains a non-rotable protocol, the brouter performs a bridging function, using MAC address to find the proper recipient on the local segment.

A brouter operates at the Data Link and Network layers.

Allows for communications between dissimilar systems on the network. It also translates protocols. The gateway operates at the application layer.

• Star bus
• Star ring
• Hybrid mesh

5. Token-ring networks use a token-passing method to provide equal access to the network for all computers. Computers cannot transmit data unless they have a token. Token ring networks use a larger data frame than Ethernet networks. This allows token-ring networks to transfer large data blocks more efficiently than Ethernet netwoks.

Used by Apple MacIntosh computers. The cabling is called LocalTalk. LocaTalk uses a network media-access method called Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) which means that the computer broadcast a warning before it transmits to the network.

Attached Resource Computer Network (ARCNet) uses a token-passing method in a logical ring. The token is passed to the next-highest numerical station number. ARCNet is no longer a popoluar networking method because you manually configure the ARCNet cards, and ARCNet speeds are a were 2.5Mbps. It uses RG-62 (93 ohms) cabling and it can be wired as a star, bus or a star-bus and it uses a logical-ring media-access method.

Fiber Distributed Data Interface (FDDI) uses fiber cable and token-passing media-access mecanism to create a fast and reliable network. Spped is up to 100Mbps for 500 nodes over a distance of 100 KM (62 miles). It can be implemented on a dual ring configuration to ensure redundancy: The primary and secondary ring.

• Workstations have to be within 100 m from the concentrator.
• 1024 stations are allowed on a segment without bridging

• 5 trunk segments can be connected, with 4 repeaters or concentrators, with no more than 3 populated segments (Coaxial cable)
• Segment length up to 185m
• 30 workstations per trunk
• Up to 1025 stations per network
• Maximum trunk length up to 925m
• The minimum cable length between stations is 20 inches.

• 2 fiber optic links to transmit and receive data
• Maximum length of 2K

• Max number of WS on type 1 or fiber at 16Mpbs: 260
• Max number of WS on type 3 at 4Mbps: 72
• Distance between MSAU on type 1: 100
• Distance between MSAU on type 2: 45
• Max number of MSAU (Multi-Station Access Unit): 33
• Max distance of the ring: 4km with fiber.

• Max bus segment length on coaxial cable: 300m

Twisted-pairs: 120m

• Max number of WS per coaxial segment: 8
• Max number of WS per network: 255
• Max distance between WS and hub: 200m
• Max distance from passive hub and active hub: 30m

Active hub and active hub: 600m

• Max distance between WS: 6000m
• Max WS per passive hub: 4
• Max distance from WS and passive hub: 30m
• Passive hub cannot be connected to passive hub.

A true multitasking OS is able to support as many simultaneous processes as there are CPUs. However, when a computer only has one CPU, multitasking can be simulated through a technique called time slicing.

Time slicing involves dividing CPU computing cycles between multiple tasks. You do this by giving each task a certain amount of process cycles, then halting that task to make the next task active. This process repeats until each task is finished.

There are 2 types of multitasking:

The operating system controls which processes are allowed access to the CPU and for how long. Ounce the assigned time slice expires, the current process halted and the next process is given its computing time.

The operating system cannot stop a process; once CPU control is given to a process, it retains control until the process is complete. During this time, no other process is allowed to access the CPU.

There are actually 2 types of redirectors in use on any network: The client redirector and the server redirector. Both redirectos operate at the presentation layer. When a client makes a request for a network application or service, the redirector intercepts that request and examines it to determine if the resource is local or remote. If the resource is local, the redirector forwards the request to the CPU for immediate processing. If the request is for the network, the redirector forwards the request across the network to the appropriate server.

A designator is a piece of software that manages the assignment of drive letters to both local and remote network resources or shared drives, which aids in network resource interaction.

The number of bytes read and written to the server.

Queue commands

The number of commands that are awaiting execution is one measure of how busy the server is.

Tells you something about your network's physical topology.

High rates of failed logons, failed access to objects and failed changes to security settings may indicate a security risk on the network.

You can tell a bit about server activity by observing the rate at which connections to the server are made and how those connections are broken, whether by a normal logoff, by an error or by a server timeout. More RAM in the server may levitate the problem or you may need to update hardware.

Occur when data is removed from a program's working set and is moved to another area in the physical disk.

Occur when the data has gone unused for so long or there is such a shortage of physical memory that program data is actually stored on the hard disk.

Leased WAN links generally use one of the following technologies to make the necessary connections:

• Analog connections: These use conventional telephone lines, with voice-signaling (modem) technologies.
• Digital connections: These use digital-grade telephone lines, with digital technologies all the way.
• Switched connections: These use multiple sets of links between sender and receiver to move data.

Switching (as in switched connections) refers to finding a path for data transmission across a number of potential links between sender and receiver. On the other hand, analog and digital connections require a fixed connection to exist, at least for the duration of each communication session. Switching methods include both circuit switching and packet switching.

When you call someone, the phone company maps out and reserves a single communications line between you and the person you are calling. However, if you call the same party numerous times in a day, you probably won't be connected through the same set of lines.

Circuit-switched networks can be inexpensive but they are generally slow and not exceptionally efficient for transmitting data, especially in large amounts, or when delivery time is a serious concern (for voice or video traffic, for instance).

The Public Switched Telephone Network (PSTN) can provide data communications as well as voice communications.

Analog transmission consists of sending streams of continuously modulated data, rather than two signals (one for zeros and another for ones), as with digital transmissions.

There are two predominant types of analog lines used for data communications:

• Dial-up lines are voice-grade lines for which you use a modem to connect to another modem that is attached to another computer. These lines are usually limited to a 56 Kbps rate, and frequently you will have to contend with noise on the line.
• Dedicated/Leased Lines: A leased line is faster and more reliable than a dialup line, but is also more expensive because it's always available whether it's in use or not. Line conditioning (a service that reduces delay and noise on the line, allowing for better transmissions) can make the leased lines even more reliable, and is often part of a leased-line package, or you can obtain it from the service provider at an additional cost.

DDS lines use a point-to-point synchronous method that can transmit at 2.4,4.8, 9.6, and 56 Kbps. high- speed digital lines offer nearly error-free transmission. Such lines are available in numerous forms, including full and fractional T1, T3, and Switched 56.

Data sent via packet switching is broken down into small pieces of information called packets. Each packet consists of a piece of the data to be transmitted and certain header information that contains the destination address. Packets are sent one at a time, and rely on special network protocols to find a path between sender and receiver, and to deliver them to their proper destinations. It's highly likely that not all packets will travel the same route from source to destination, nor will they arrive in the same order as they were sent. If a packet gets lost or damaged during transmission, it is a relatively easy task to ship out a replacement. But, it is pretty certain that all packets will ultimately arrive at their proper destinations, and be reassembled into whatever original form the sent data may have taken.

The actual PDN connection may pass through a computer with an X.25 interface, or through a standalone device called a packet assembler/disassembler (PAD).

Because of its error checking and retransmission of erroneous or lost data packets, X.25 is one of the slowest of the advanced WAN technologies, but it is also one of the most broadly available and affordable options. It remains pervasive outside Europe and the United States.

The Integrated Services Digital Network (ISDN) primary goal is to integrate voice and data services by replacing analog telephone lines with digital equivalents that are suited for both voice and all kinds of digital traffic, including data, Video, and other digital data streams.

Although ISDN is available in many locations, it is used considerably less than PSTN lines. This is due in part to the higher costs of ISDN connections, but also because ISDN does not always offer a sufficient boost to bandwidth compared to PSTN lines (which can support data throughput as high as 115 Kbps including compression). ISDN offers nominal bandwidth of 64 Kbps per channel, and most ISDN lines offer nominal bandwidth of 128 Kbps, because channels are often used in pairs (with compression, some vendors claim throughput as high as 400-plus Kbps across two ISDN bearer channels).

You may purchase ISDN in two different forms:

• The Basic Rate Interface (BRI) targets home users or small businesses. A BRI consists of two bearer channels, also called B-Channels (64 Kbps), and a data channel, also called a D-Channel (16 Kbps), for a total of 144 Kbps of bandwidth. Each of the B-Channels may be used to transmit and receive either voice or data. The D-Channel is used for call setup and control (and sometimes for X.25 or fax communications). With the proper equipment and software, both B-Channels may be combined into a single virtual channel with a data transmission speed of 128 Kbps.
• The Primary Rate Interface (PRI) targets service providers and large organizations. A PRI uses a full T-1 line with 23 B-Channels and a single 64-Kbps D-Channel. Users with access to a PRI can use B-Channels individually, in pairs (typically for incoming BRI connections), or aggregated in any combinations up to all 23 channels together.

ISDN is a dial-up technology that furnishes voice and data at speeds up to 128 Kbps. Although it took a long time to get to the marketplace, it now enjoys widespread use as a source for Internet connectivity.

Frame relay uses variable-length packets in a packet-switching environment. It establishes a logical path that's called a Permanent Virtual Circuit (PVC) between end-points. PVCs take fixed paths, so a PVC is the equivalent of a dedicated line in a packet-switched network. The path is fixed, so network nodes don't have to waste time calculating routes. Frame relay connections operate at speeds between 56 Kbps and 1.544 Mbps because they use PVCs, and there is no built-in error checking. Frame relay services are gaining popularity: They are much faster than other networking systems at performing basic packet-switching operations, and customers can specify exactly what amount of bandwidth they

want to pay for.

Frame relay costs less than a dedicated line or an ATM connection and provides data transmission rates of up to 1.544 Mbps over conventional or fiber optic media.

T1 is a point-to-point transmission technology that consists of 24 64-Kbps channels for a total transmission capability of 1.544 Mbps. Each of the channels may be used as a separate voice or data communications channel, or channels may be combined to provide higher transmission rates.

A faster commercial T-carrier line is called a T3. It is the equivalent of 28 T1 lines and handles a data rate of 44.736 Mbps. A T3 is the highest capacity leased-line service available from most communications carriers, and is designed to transport large amounts of data at high speeds between two points.

T1 lines are the most common high-speed connectivity in use today. They can transmit up to 1.544 Mbps. One nice feature is that a user can lease a fraction of the TI line as needed.T3 lines are 28 times the size of aT1 (roughly 45 Mbps), and can be consumed in fractions as well.

a Switched 56 line is nothing more than a circuit- switched version of a standard 56-Kbps DDS leased line. As customers pay only for connection time, resulting costs are usually significantly lower than those of a dedicated line.

Switched 56 is merely a circuit-switched version of a standard 56Kbps line. This is good for customers because they only have to pay for what they use.

ATM is an advanced packet- switching technology that transmits data over LANs or WANs in fixed-length 53-byte chunks, called cells, at speeds of up to 622 Mbps. ATM can accommodate voice, data, fax, realtime video, CD-quality audio, imaging, and multimegabit data transmission.

Unlike frame relay, which uses variable-sized packets, ATM cells have a fixed length of 53 bytes. Of these 53 bytes, 48 bytes contain data and 5 contain header information. Because data packets of uniform length are much easier to transport than random-sized packets, ATM can use network equipment to switch, route, and move cells much more quickly than the same equipment could handle randomly sized frames.

ATM is a packet-switched technology that transmits data in fixed length, 53-byte cells. Theoretically capable of speeds of 1.2 (OC24) and even 2A Gbps (OC-48), it usually transmits in the 155 Mbps (OC-3) to 622 (OC-1 2) Mbps range.

FDDI is not like a regular Token Ring network because more than one computer at a time can transmit a token so that multiple tokens can circulate on the ring at any one time. The token-passing system is used in a dual-ring setting. Traffic in the FDDI network consists of two similar data streams moving in opposite directions around two counter-rotating rings. FDDI's dual-ring architecture increases its reliability.

Synchronous Optical Network. SONET is a fiber optic WAN technology used to deliver voice, data, and video at speeds in multiples of 51.84 Mbps. SONET'S main goals were to create a method by which all carriers could interconnect, and to unify differing standards used in Europe, the United States, and Asia especially Japan.

SONET is a fiber Optic WAN technology used to deliver voice, data, and video at speed up to 622 Mbps, and beyond.

Switched Multimegabit Data Services. Like ATM, SMDS uses a fixed-length cell of 53 bytes for data transmission. Like ATM and frane relay, it provides no error checking, leaving that up to devices at the connection points.