Networking Guide : The OSI Reference ModelThe International Organization of Standardization (ISO) defined procedures forcomputer communications which was called Open System Interconnection (OSI)Reference Model or OSI Model for short. The OSI Model describes how data flowsfrom one computer to another computer in a network. Picture. The OSI Model When computers A and B communicate, data travels across the 7-layered protocol stack.The OSI Model is defined as a protocol stack that consists of seven logical layers.Each layer has specific functions and handles a unique data format. When twocomputers communicate, data flows from the user-to-network interface (application) atthe sending computer down through the protocol stack to the connecting physicalmedium (i.e. cable, radio, or infrared) and up through the protocol stack to the
network-to-user interface (application) at the receiving computer. When data flowsfrom an upper layer to a lower layer, it is converted to the lower layer data format anda lower layer header is added to it. This process is called encapsulation. Conversely,when data flows from a lower layer to an upper layer, it is converted to the upper layerdata format and the lower layer header is discarded.Networking Guide : The OSI Reference ModelEven though most computer network technologies do not follow strictly to the OSIModel in that they combine several OSI layers functions in one protocol, the OSIModel is still used as a reference and a guideline in network design. Understandingthe OSI Model will help you understand how a network works. The OSI Model protocolstack is explained in the following table: Layer Layer Name Functions Examples Application HTTP, FTP, Telnet, SMTP,Layer 7 application support Layer SNMP, POP3, IMAP4 data format conversion, PresentationLayer 6 data compression, and Layer encryptionLayer 5 Session user identification; SIP
Layer establishing, maintaining, and terminating a session TransportLayer 4 end-to-end transport TCP, UDP, RTP, RTCP Layer Network IP, IPSec, IPX, NetBEUI,Layer 3 addressing, routing Layer AppleTalk, ICMP Ethernet, Wi-Fi, HomePNA, medium access Data Link HomePlug, PPP, PPTP,Layer 2 control, error detection, Layer L2TP, ATM, Frame Relay, retransmission Token Ring, FDDI electrical/optical RF, UTP, STP, coax, fiber Physical signaling, cabling,Layer 1 optic, connectors, Layer connector pin signaling, voltages assignmentDue to its complex functions, the Data Link Layer is divided into two sublayers, thatisMedia Access Control (MAC)sublayer and Logical Link Control (LLC)sublayer. MACsublayer is the lower part, closer to the Physical Layer. MAC sublayer controls accessto the physical medium. LLC sublayer is the upper part that interfaces with theNetwork Layer.
Session Layer, Presentation Layer, and Application Layer are often referred to asUpper Layers. These layers basically handle user connection and data formatting. Inmost network technologies, such as TCP/IP, the differences between the three layersare blurred and their functions are often handled by one protocol.Physical Layer, Data Link Layer, Network Layer, and Transport Layer are referred toas Lower Layers. The lower layers generally concern with how data is transportedacross the network.Networking Guide : Physical Media - Twisted Pair CableA network needs physical medium to connect its nodes together. The physical mediumis where the data actually flows. There are several media types often used innetworking. They are described in the following paragraphs. Picture: Twisted PairTwisted pairis two insulated copper wires that are twisted around each other tominimize interference and noise from other wires. Based on the presence of individualshield and overall (outer) shield, there are three types of twisted pair, i.e. UTP, STP,and ScTP. Individual shield encloses a single twisted pair, while outer shield enclosesall twisted pairs in a cable. A shield is a protective sheath that is made from conductivematerial (metal) and functions to protect the twisted pair from external interference. Aninsulator is made from non-conductive material, such as plastic.
Picture: Unshielded Twisted Pair (UTP)UTP (Unshielded Twisted Pair)is a cable containing several twisted pairs that is onlyinsulated but not shielded. UTP is the most widely used cable in telephone andcomputer networks because it is relatively cheaper than other cables and performswell in normal electrical environment such as inside an office or a house.Networking Guide : Physical Media - Twisted Pair Cable Categories Picture: Shielded Twisted Pair (STP)STP (Shielded Twisted Pair)is a cable containing several twisted pairs that hasindividual shields, an outer shield, and an insulator. STP is more reliable than UTP.However STP is less known because it is used only in situation where there iscomplex cabling such as in factory building.
Picture: Screened Twisted Pair (ScTP)ScTP (Screened Twisted Pair) is similar with STP but each twisted pair has noindividual shield.Twisted pair cable is graded based on the number of twists per inch, its cable structureand traffic carrying capacity into several categories, as follows: Category Type Bandwidth Typical applications Cat 1 UTP < 1 MHz telephone Cat 2 UTP 1 MHz telephone UTP, telephone, 10BaseT, 4 Mbps Token Cat 3 ScTP, 16 MHz Ring STP UTP, Cat 4 ScTP, 20 MHz 16 Mbps Token Ring, 10BaseT STP UTP, Cat 5 100 MHz 10BaseT, 100BaseT ScTP,
STP UTP, Cat 5e ScTP, 350 MHz 100BaseT, 1000BaseT STP UTP, Cat 6 ScTP, 550 MHz 1000BaseT, ATM STP ScTP, Cat 7 600 MHz 10 Gbps network STP Table: Twisted Pair categories and their applicationsNetworking Guide : Physical Media - Coaxial Cable Picture: Coaxial cable (coax) structureCoaxial cable contains a solid or stranded wire in the core that is insulated with adielectric layer, then protected with a solid or braided metallic shield, and covered withan outer insulator. Electromagnetic wave propagation in a coaxial cable is confinedwithin the space between the core and the outer conductors. The structure of a coaxialcable makes it less susceptible to interference, noise, and crosstalk than the twisted
pair cable.Coaxial cable is often classified based on its characteristic impedance. Most coaxialcables have characteristic impedance of 50 or 75 Ohms. Coaxial cables in the marketare usually named with RG prefix which may stand for Radio Grade. Each RG type isrelated with certain characteristic impedance and outer diameter. For example RG-6which has impedance of 75 Ohms is used for connecting cable modem or TV to aCATV network. RG-58 (50 Ohms) is used in earlier Ethernet networks (10Base2).Coaxial cable is terminated with RF (BNC) connectors.Networking Guide : Physical Media - Fiber Optic Picture: Fiber optic structureFiber Optic (optical fiber) is a thin glass or plastic strand in the core which issurrounded by a cladding and a protective coat and is used to carry information inoptical (light) pulses. Because in fiber optic, information is transmitted in optical pulsesinstead of electrical signals, fiber optic is not affected by EMI (electromagneticinterference) and RFI (radio frequency interference). Moreover, fiber optic has verylarge bandwidth which is limited only by the equipment that lights the fiber (i.e.SDH/SONET, ATM, DWDM). But fiber optic is more expensive than twisted pair, coaxand radio.
Picture: single-mode fiber (left) and multimode fiber (right)Fiber optic is often classified into single-mode and multimode. In a single-mode fiber,light travels in one path (mode). In a multimode fiber, light travels in multiple paths(multimode). Single mode fiber can reach longer distance than multimode fiber, so it ismostly used for MAN or WAN. While multimode fiber is suitable for implementing highspeed LAN.Because of its reliability and wider bandwidth, fiber optic is often used in backbonenetworks where cables run in ducts and in broadband networks that deliver bandwidthintensive applications, such as HDTV, video streaming, video conferencing and Videoon Demand.Networking Guide : Network ComponentsRepeaterRepeater receives signal from a transmitter, amplifies it, and retransmits it to areceiver. A repeater is put in a network to extend the network to a longer distance or agreater area. There can be more than one repeater between a transmitter and areceiver, however the number of repeaters is not unlimited, because additionalrepeaters may introduce more interference or noise.
Picture: Repeater A repeater extends the reach of transceivers 1 and 2. Note: Transceiver is transmitter and receiver.Repeater is also known as regenerator. Some vendors name it range expander or line(cable) extender.BridgeWhen you have two or more networks with different layer 2 protocols, such asEthernet, HomePNA, HomePlug, and wireless LAN (Wi-Fi) you can connect thenetworks using a bridge. Bridge is also used to split a network into separate segments.This is intended to filter traffic and create an efficient network. Picture: Bridge A network bridge enables communication between two computers at different networks.A bridge function can be handled by a software application. In Windows XP, when youinstall two or more network adapters, a Network Bridge is automatically created foryou. A Network Bridge icon will appear in the Network Connections folder. However inWindows XP Service Pack 2 (SP2), a Network Bridge will be created only after userconfirmation. This behavior is in line with SP2 main goal, thats to improve computer
and network security.ing Guide : Network ComponentsHubHub is the central connection point in a network. Hub is used in a network that usesstar topology. A sending computer transmits its signal to a hub, the hub thenretransmits the signal to all other computers. A passive hub functions as a relaystation that receives and retransmits signal. An active hub functions as a repeater thatregenerates signal before retransmitting. Picture: HubWhen A sends to C, the Hub receives signal from A and retransmits it to both B and C. Only C then processes the signal.Using a hub, the network bandwidth (capacity) is shared by all available computers,therefore each computer only uses a portion of bandwidth. Thats why hub is mostlyused in small networks where there are only a few connected devices or computers.However, hub is not required if there are only two computers in a network. In thatcase, a direct connection using cable or wireless link can be used to connect bothcomputers.Switch
Like hub, switch works as the central connection point in a network. However when aswitch receives a packet from a sending computer, it examines the destinationaddress (i.e. MAC address of the destination computer) from the packet header andretransmits the packet to the destination computer only. Thats possible because aswitch maintains a table that maps all its ports with all connected devices MACaddresses. Picture: Switch When A sends to C, the Switch receives signal from A and only retransmits it to C. B doesnt receive the signal.Using a switch, the whole bandwidth can be used by each connected computer. Thatswhy most big networks in which a large amount of data must be transferred at anygiven time, use a switch instead of hub. Switch is not always a separate device, it isvery often integrated with router.GatewayGateway functions to connect two completely different networks. It performs protocoltranslation. Although gateway is considered a Layer 7 device in many publications, itactually works across the seven layers of the OSI Model. In Internet Telephony, a
gateway connects the VoIP network to the PSTN. Picture: Gateway VoIP/PSTN Gateway performs protocols and signaling translation, so a VoIP-enabled phone or PC can communicate with a regular phone. Networking Guide : Network Components - Summary The following table explains network components along with their functions and the corresponding layers in the OSI Model. Click each component name for a more detailed explanation. Network Functions OSI Model Component converts a computer message into Network electrical/optical signals for Physical (Layer 1) Adapter transmission across a network. puts a message (baseband signal) on Modem a carrier for efficient transmission; (Modulator Physical (Layer 1) takes the baseband signal from the demodulator) carrier.
Repeater receives signal, amplifies it, then Physical (Layer 1)(Regenerator) retransmits it. connects networks with different Data Link (LayerBridge Layer 2 protocols; divides a network 2) into several segments to filter traffic. connects computers in a network; receives a packet from a sendingHub Physical (Layer 1) computer and transmits it to all other computers. connects computers in a network; receives a packet from a sending Data Link (LayerSwitch computer and transmits it only to its 2) destination. connects computers in a wireless network; connects the wireless Data Link (LayerAccess Point network to wired networks; connects 2) it to the Internet. forwards a packet to its destinationRouter by examining the packet destination Network (Layer 3) network address.Residential connects a home network to the Network (Layer 3)
Gateway Internet; hides all computers in the home network from the Internet. connects two totally different Gateway networks; translates one All layers signaling/protocol into another.Networking Guide : ArchitectureNetwork architecture describes the relation among nodes in a network.A client-servernetwork has a node that functions as a server which providesresources (e.g. programs, disk, printers) for other nodes (client computers) andmanages clients access to the network resources. Corporate networks are typicallyclient-server with one or more servers that store corporate information and employeescomputers as clients.
Picture: Client-Server Network Client computers access programs or information provided by the servers.A peer-to-peer network does not have a server, each node (computer) in the networkcan share its own resources with other nodes and determines other nodes accesslevels to its own resources. Home networks are typically peer-to-peer. A peer-to-peernetwork is also called a workgroup. Picture: Peer-to-Peer Network