Download as pdf
Download as pdf
You are on page 1of 33
‘> responsible for delivering data from one network Interface to another ‘> Responsible for node-to-node data delivery What is a “network interface”? > In the context of the OSI and TCP/IP stack, "network interface” is the interface to the network layer What is a node? > "Node" above refers to a device (intermediate or end) at the network layer ‘> Data-link Layer Revisited > Local Area Network > Ethernet ‘> Switching and Bridging wih Data tekiger or Data-link layer link Data-link layer link example > pret pot er*datatnkleyer cou inks 2105s a at jer tecrology i" i car see ua Ss co eae gw _B ser carom cng tora anton? & : & Tar sna" oma ar “ 4 eS & & = — Functions of Data-link and Phy layers Data-link Sublayers ‘> The data-tink layer works closely with the phy layer > Data Link Layer functions 1 Framing and addressing ‘= Error control for reliable data transmission ‘Flow control 1 Media Access control ' link activation and deactivation (Diane) > Physical Layer Raw transmission (eg. Modulation demodulation) 1 Physical interface specifications Local Area Networks | This refers to assigning an identifier to datavtink-layer endpoints > Examples = MAC address 1» ATM endpoint ID (VCT [virtual channel ID], VPI [virtual path 10]) ‘> ALAN Is a group of computers or devices sharing a ‘common muiti-node communications system > Basic Topologies > Basic Building blocks 1+ End nodes 1» Network adapters 1 Intermediate nodes, + media we @ LAN: Structure — WiFi Extensions a Nes 8 by the wireless ae. wees OG ‘> LANs are usually implemented using a Hierarchical structure, for example LAN: Structured Cabling LAN: Structured Cabling > Ethernet networks are usually implemented using > Horizontal Cabling structured cabling "One telecommunications closet 1 One or more cross-connect 1 Network Outlets + workstations LAN: Structure Cabling LAN: Structured Cabling ‘> Maximum cable lengths for horizontal cabling Common Protocols and Standards ‘TEE 802.1 (Bridging) ‘> TEEE 802.2 (Logical Link Control sublayer) ‘> TEEE 802.3x (Ethernet) ‘> IEEE 802.5 (Token Ring) ‘> IEEE 802.6 (MAN) ‘> IEEE 802.11 (Win) ‘> TEE 802.15 (Wireless PAN) ‘> SLIP and PPP FDO! asignment HDLC ‘oad bey about each of Ethernet Technologies > Ethernet 1» Refers tothe family of LAN implementations derivative from the orginal Ethernet invented by Xerox ‘ Sthemet and IEEE 902.3 1100 Mbps or Fast Etheret ‘+ 1000 Mbps or Gigabit Ethernet +10 Ggabit Etheret sete ‘= The most common LAN implementation because of ‘+ Feabiity ‘+ Relative simply LAN Addressing ‘>To distinguish between network nodes at the data- link-layer level, MAC Addresses are used ‘> Examples: (oom Lafomdohoafn| NIC MAC ost pote AA1F:3C:F@:DC:01 broadcast address FEFEFEFR AREF GIST Basic Ethernet LAN Implementations ‘> 10Base-5 or 10base-2 Implementations using coaxial media Tegoey Techno) Basic Ethernet LAN Implementations (2) > 10-base-5/2 use Bus type topology [Cesscy Tecincoay | [Legacy Tecnaogy but topology stated line Basic Ethernet LAN Implementations (3) > 10Base-T Implementation using twisted-pair media [egacy Tecialony bt star topony ‘ilu Basic Ethernet LAN Implementations (5) ‘> Combination twisted-pair and coaxial media Legacy sep tay the coax buss reac bya star Ethernet/IEEE 802.3x: Physical Layer Ethernet/IEEE 802.3x: Layers 1 Mapping of various layer 1 options = eects ‘ena favs ayer = inplementaenire ference Mal mt ES =o [=I = ‘arous 802 physical ayer spetications ‘rame Format Ethernet/IEEE 802.3: Frame Format + Preamble * atemiting pattern of ones and zeros ts receiving tations that a frame scoing (Ethernet or IEEE 802.3) “m ‘ The Ethemet rane ncdes an adtana bye hats the “Suton of the Start of Frame fel specie nthe IEEE + Start-of-Frame (SOF) The IEE 802.3 dente ye eds th wo consecutive bis, he sve to sycranize te fame ecpton partons of al talons on the LAN, SOF exp speciad = ‘ Inetheret| ‘Destination and Source Addresses sccm | see The st 3 bytes ofthe adresses are spect bythe IEEE ‘ne vendor dependent bass > The at? byte are epctiod by the Ether oF IEEE 902.3 von Ethernet/IEEE 802.3: Frame Format Ethernet/ IEEE 802.3: Frame Format Mh ac aes ahs nt (se e 1 Broadcast >> The destination address can be unicast, multicast (group), re rtronic (a ms) Unicast Es MA6C:2:69 = Mulneast\ Seideraertoun|_them merce) * eghenee 82:88 -BS--Re >| \\ Gee >|- eee >| \ }+ (aa) >|< aoe >| \ +--+ seme >| Dioguished by bt 0 ‘ofthe MAC 1* octet ‘Type (Ethernet) The type species the upper potca to rece the ata afte Ethernet processing completed + Length (IEEE 802.3) The length ndeates the numberof bytes of data that oto ths ld ‘Data (Ethernet) ° data cortained inthe frame sent to an upper yer rotoca wich ented inthe ype al » athough Eheret Version 2 docs nt spect any padding (nconrastto IEEE 802’), terme expects a eas 36 bytes of daa Ethernet/IEEE 802.3: Operation and Setup 1» Host connectivity is established using media standard BaBR 1+ The operation depends on the medium/devices used ‘+ Hal-duplex ony: coaval cables (1Obase-2/1Obase-5) ‘ Fullkduplex UTP with the use of switches 1 Shared media requires a media contention method *coMAjoo Ethernet/IEEE 802.3: Frame Format ‘Data (IEEE 802.3) > If dat i he fe slice tl he rae tos minimum 64 Dee see, pating bytes reinserted to ensure teat a 6p frame ‘Frame Check Sequence (FCS) his sequence cotans a byte yc redundancy check (CRC) va, whe created by the sending device an Fecaleultd by the recaing devi to check for daraged frame, The CRC plynomi is A307 HAM EAHA Ethernet/IEEE 802.3: Basic Operation '* Operation through the layers h Ethernet/IEEE 802. Basic Operation 1 Early Ethemet is designed for broadcast-type between stations with each receiving all frames transmitted «All stations have simultaneous access te the medium but {or proper operation only one must be transmitting at any given time otherwise collision occurs 1 Uses Carrier Sense Multiple Access with Collision Detection (CSMA/CD) to resolve multiple access problem 1 Stations drop frames ifthe received data ‘1s not an integral number of bytes ‘+ not destined forthe receiving station ‘has errors based on the FCS CRC value Persistence strategy > Persistence strategy is the approach used for ‘media access control > Non-persistent CSMA 1. If the medium is idle, transmit; otherwise, go to step 2 2. Ifthe medium is busy, wait an amount of time drawn from a probablity distribution (the retransmission delay) and repeat step 1 1+ The use of random delay reduces the probability of colisions {= The medium wil generally remain idle following the ‘end of a Varsmission even If there are one or more stations waiting to transmit. So the capactty is wasted. Ethernet/IEEE 802.3: Basic Operation Persistence strategy persistent protocol ‘= To avoid idle channel time. 1. Ifthe medium is idle, transmit; otherwise, goto step 2 Continue to listen until the Is sensed idle; then transmit immediately. * Used by IEEE 802.3 Persistence strategy Ethernet/IEEE 802.3: Collisions > p-persistent CSMA > Ethemet Collisions 1 A compromise that attempts to reduce colisions, like ‘A condition that results from concurrent transmissions ron-persistent, and reduce ile time, lke t-persistent from multiple data terminal equipment (OTE) sources within a single colision domain 4. Ifthe mecium is idl, transmit with probability p, and 1 This is bad for network performance delay one time unit vith probability (1-p). The time unit is typically equal to the maximum propagation delay ‘> This issue is addressed by the use of switched 2. Ifthe medium is busy, continue to listen uni the Ethemnet channel s idle and repeat step 1 3, If transmission is delayed one time unit, repeat step 1 Ethernet/IEEE 802.3: Collisions Ethernet/IEEE 802.3: Collisions > Collision Domain + Colision domains are prevalent in the following setup "A single, half duplex mode CSMA/CD network ‘If two oF more Media Access Control (MAC) sublayers are thin the same collsion domain and both transmit atthe ‘same te, aeoision will occur ‘+ MC sublayers separated bya repeater are in the same calision domain. — ‘MAC sublayers separated by a bridge are within diferent oun callsion domains. ay os Ethernet/IEEE 802.3: Collisions onbined igor ‘alien dma Ethernet/IEEE 802.3: Broadcast domain ‘> What is a broadcast domain? ‘+ A section of @ network where nodes can send broadcast messages to each other > Which belongs to the same broadcast domain? "= Any computer or node connected tothe ‘Same hub or switch ‘Same hierarchy of hubs or sitches > Which is larger, a broadcast or colsion domain? Ethernet/IEEE 802.3: Collisions =" Ethernet/IEEE 802.3: 5-4-3 Rule 1» The 5-4-3 rule specifies that a maximum of five Ethernet segments can be joined through the use of a ‘maximum of four repeaters in a shared media system ‘The ute of too many repeaters to extend Ethemet networks result in excessive regeneration delays ‘Developed ven 10base5 and 1Obase2 were used Ethernet/IEEE 802.3: 5-4-3 Rule ‘= The "3" pat ofthe rule specifies the maximum number of populated segments out of the maximum of Five ‘ Unpopuiated segments are more suitable for use in ‘extending Ethernet network than populated ones because ofthe possibilty ofcolisions inte later 1» The 5-4-3 rule limits the span of the network but in some cases can be exceeded (refer to 10BaseF8) “this tims the network span to within 2500 meters when sing 108a5e-5 9 The wee of pt ers nreates the span beyond this Ethernet/IEEE 802.3: g aan ae pant g = > a> ta ten a tee p = i coon ace * ‘comes propo * *” Ethernet/IEEE 802.3: Timing g oe y ee ke = 5 cola ce He cmp gga x ok Ethernet/IEEE 802.3: ‘> Timing Requirements for half-duplex Ethernet 1+ The slot time is set so that a station recognizes a calision before the end of transmitting a minimum-sized frame gee gf OBR, Ethernet/IEEE 802.3: Timing a g unt tt —a a cee jn ea Terri Ethernet/IEEE 802.3: Timing > cable delay vs Ethemet slot time 0.9 si2us 4 1 Cat3 cable delaynn = 100 /(0.68(3e8)) = 0.49 us 1 Cats cable delaysayn = 2500 / (0.68(3e8)) = 12.25 us Thick coax delaynig = 2500 / 0.77 f 3€8 = 10.8 us > Brercise ' Determine the typical propagation delays in a CatS and CatSe UTP cables Ethernet/IEEE 802.3: Timing ‘> Typical propagation speeds Neato pagation speed Tack Coax "770 (231/000 kn/zee) Thin Come 1650 (195,000 km/sec) Teistea gair 1396 (177,000 im/ace) Fiber 66 (198,000 ka/aee) AUT Cable "65e (195,000 kn/ee ‘> From datasheets (ex. Cat3 UTP) ewe CCharactorisc Impedance : 1008180. / 2 Nominal Velouiyof Propagation : 68%. ‘Maxum Propagation Delay : 538 ns (00! @ 100 MHz Ethernet/IEEE 802.3: Basic Operation > inter-frame Gap ‘= a quiet time in between frames to allow clock circuitry In repeaters, hubs, and stations to resynchronize 1 This is 9.6 us, 0.96 us and 0.096 us for standard (10- mbps), fast: and gigabit Ethemet, respectively ‘> Binary Exponential Back-off Algorithm ‘After colison, time is cvided into discrete slots whose ‘uration is equal to the worst ease round-trip propagation time ” side *100ne a 51.2 us for 30 bps yr086 Ethernet/IEEE 802.3: Basic Operation + Ate fst colision cling stations wat 0 or 1 slots {(Gecided randomly before retrying ‘+ TFthey code again, they watfor either 0,1, 20r 3 slots and so on > The masini possible st to vat ned at 1023 ater ‘itch communeaton faire wl be decir > Capture Effect 1 Leads to a scenario where the back-off algorithm becomes unfair 10Base-5 (Thick Ethernet) 1» 10Base'5 in newer networks are used as backbone for 10Base-T hubs 10Base-5 (Thick Ethernet) 1» This Is the standerd that more closely resembles the original Ethernet standard for 10 mbps transmission 1 Makes use of a thick S0-ohm coaxial cable 1 The transceiver, tap and housing forms the medium ‘sccase unit (MAU) 10Base-2 (Thin Ethernet) 1 Provides a cheaper and more flexible alternative to 10Base-5 for communication up to 10 mbps 1 Uses a thinner coaxial cable ‘= Maximum segment length is 185 meters ‘+ Minimum segment length is 0.5 meter "= Supports only up to 30 stations per segment 1» Makes use of a T BNC-type connector 1 Transceiver electronics is integrated with NIC goeoa- ' Provides for network communications at 10 mbps over, 4100-ohm UTP cables 1» 10Base-T networks typically use a physical star and. logical bus architecture 1 Cable runs to/from hubs and stations is limited to 100m 10Base-T 1+ 10Base-T uses a UTP cable with two twisted pairs ‘One pair for transmiting and one par for receving 1» The cable endpoints use an 8-wire RJ-45 connector although only four are used f= A 10Baset-T network can be buat using NICs, hubs and UTP cabling = Within a collision domain, ‘only half-duplex operation is. possible 10Base-T oem o— oO a. o_O oO Oo oO mm Oo Ethernet/IEEE 802.3 Pubes i inthe want 100 Mbps Ethernet 1» IEEE 802.3u provides three Fast Ethemet specifications + 1008a56-7%, 1008a5e-FX and 100Base-T# + IEEE 602.12 defines another standard based on demand Priority mechanism + Supported either token ing or Ethernet « Refered to as 100VG-AnyLAN ‘Not 2 commercial success 1 Interoperable with 10Base-x 10Base-F and FOIRL ++ To extend the physical range of Ethemet-based ‘communication, fiber optic cables can be used ‘+ nial standard used was Fiberoptc repeater lnk (FOIRL) followed by the 10Base-F standard = FOIRL “Limited to 1000km between repeaters and subject to the 53 rue ‘+ Standards apply to repeater links only ‘The dstance lit for FOIRL relates t the timing constraints ofthe IEE 802.3 standard 100 Mbps Ethernet: Layering Layering scheme wrath in Pipl a dent siaet || sie | 100 Mbps Ethernet: Layering 100 Mbps Ethernet: standards ah aoe” Se ee \ [eect > aes. | ater ‘aie aap SUR TT einige eS = Rateno tp To = =) Pa) |) er Dae Deas RTE == seen =| = a oe co oe =| = ie S| Saar eae oy aa 100 Mbps Ethernet: auto-negotiat > Auto-negotiation ' Allows automatic detection and configuration of speeds for networks with multi-speed components 1+ The scheme used for 108ase/100Base interoperability is based on NWay by National Semiconductor ‘s Nay provides an autosensing scheme for Ethernet ‘rcuts to detect both cable type and transmission speed 1 Relies on the modification ofthe link integrity test pulses of 10Base-T “The lnk integrity pulses are replaced by fas lnk pulses ap) > FLD consists ofa series of ck and data puss foming the {eb ink code word (<0) 1000 Mbps Ethernet ‘> Gigabit Ethernet (IEEE 802.32) ‘> Standards: 1000BaseT and 1000BaseX 1000 Mbps Ethernet 10GB and Beyond > Standards and transmission distance > 1068 Ethernet is available and higher speeds are stil in the offing abe Tae a Toons 7 > 1068 Ethernet is implemented through several ‘coma rom standards using seamsasex tom eos a = 1 Optical Fiber cables {eOAECE Slat talmcedegpconee STP) 38 Assignment: Read to do or answer the following: fomnaseT Urey 8 00 + dentiy the 1068 Ethernet Standards a + What dstances are possible with ether typeof cables? Power over Ethernet (POE) ‘> Allows power to be sent along data through Ethernet cables mae Power over Ethernet (POE) PoE [=] STANDARDS IEEE Extension Type Power Budget gerDevie IEEE 20230 Type saw Power over Ethernet (POE) PoE[-]LIMITATIONS > Identify specific products for the above ‘> Identify other applications of Ethernet besides the usual computer networking setup Other application areas of Ethernet + Industral Ethemet > Automotive Ethernet Ethernet Bridges ‘> Extend the span of the network by linking together limited-tength segments in traditional Ethernet tata fo, Bridges and Switches ‘> What are bridges? ‘> Why use a bridge? ‘> What is @ transparent bridge? ‘> How transparent bridges operate > How bridges leam forwarding tables > What are bridge loops? ‘> How bridge loops are addressed by STA >What are switches? ‘> What is switched Ethemet? > What are the advantageous of switched Ethemet Types of bridges / Why Bridge? ‘> Types of bridges ‘= Transparent bridges (IEEE 802.1) 1 Source-Route bridges 1 Source-Route transparent bridges > Why Bridge? = Relabilty 1 Performance 1 Securty ' Geography Transparent Bridges - Operation > Bridges accept frames from all ports and forwards them to the appropriate destination port > Bridges know which port to forward to by maintaining a table of host accessibility = If destination is not registered inthe table, flooding occurs > Bridges usually work with MAC addresses to identify the hosts > The network number is associated with a particular bridge port Bridge Loops a Spanning tree agri © Sample network += Determine the table for each bridge Spanning Tree Algorithm 1 IEEE 802.14 1» Each bridge is assigned a unique id 8 bytes 1 Bridge priority + base MAC address 1» Each port is also assigned an id '¢ Bridge prionty + port MAC address 1» Uses the property of spanning tees ‘For any connected graph consisting of nodes and edges ‘connecting pais of nodes, a spanning tree of edges ‘maintains the connectivity ofthe graph but cortalrs no loops: Spanning Tree Algorithm 1» Each bridge port is assigned a path cost ‘Based onthe speed ofthe lnk Spanning Tree Algorithm pcr Pe count mn ges Spanning Tree Algorithm > Operation 1 Selection of root bridge (one with the lowest id) 1+ Root port on each bridge is determined '* Ths portis where the rot canbe reached withthe onest cost (oot path cost) 1» A designated bridge is assigned for each LAN, ‘The bridge vith the lowest root path cost * aftwo or moe bridges hae the same rp the one with tones i slated ‘+The designated port is that which connects the LAN to its designated brdge ‘+ Only the designated bridge ic allowed to forward frames Spanning Tree Algorithm 1+ The resulting lop free bridged network ‘= The spanning tree operation i intited when abridge is powered on and when bridges reconfigure themselves > Basic function "= A multi-port bridge, accepts frames at one port and forwards the frame in the destination port "= Can support simultaneous communication sessions = A switched ethemet can have an extended network diameter ‘> Techniques ' Crose-point or cutthrough ‘Reads the incoming frame upto the destination address and after lookup, begin fonvardng the infermaton “No eror detection can forward defective frames «Least latency Switches: techniques 1 Storeandforward + Store fll incoming frame (i. requires buffs) 1 Does CRC checking 1 Does not forvard defective frames + Greater latency 1» Hybrid or adaptive ‘ Supports both cut-through and store-and-forward ‘operation + Switches automatically between the two modes depending onthe currence of erroneous frames techniques Legend: = Pct ates naira @ -Ponceetraton atten aned @ ~ Satoh dernier pot 1 Cut-through switch ‘Switches: methods fracas ———— Ko = Switches: methods Switches: high-speed ports > segment-based + Some switches have One port is mapped to many nodessations mooie pos = = Ter ued t eset i = these prs: rates 2 Some cao ‘Switched Ethernet | Bus or hub-based star Ethernet networks operate in ‘> Switch instead of a hub in star configuration half-duplex mode and form larger collision domains ‘Smultaneou : i One colson Separate ‘aman one Semets tegiealnebork raison segment compar omans Virtual LAN (VLAN) 1 Allows logical groupings > Rationale of nodes it separate ‘+ Minimization of broadcast domains subnet 1 Scalability ‘Works even the nodes. are = Seasky ‘geographically 1 Network manageability Separated ‘Traffic management «= Partitioning into IP subnets ‘= Each Vian has its own broadcast domain and subnet address VLAN Membership 1 Static + Por-baed = Dynamic MAC based Wireless LANs The wireless approach > Section Goals 1 Wireless networking involves the use of wireless media '= Put WLAN technology into perspective instead of wired media ‘= Understand the importance of Wi 1+ The two common wireless media are 1 Understand the implementation of Wifi ‘Radio frequency waves 1 Awareness of Win Standards ‘Infrared (hich technical speaking also falls under EM) 1 Understand the isues with Wi > Advantages panels een ‘= Completes the access technology portfolio. "Adds to lt of wired access technologies = Goes where cable and fiber cannot + Involves reduced time to revenue 1 Provides broadband access extensions The wireless approach > Limitations 1 Regulated 1 Prone to interference Security risks ‘= Distance limitation > Wireless LAN Standards f= In the US: HomeRF, Bluetooth and IEEE 802.11 1 In Europe, two adeitional: HyperLAN, HyperLANZ IEEE 802.11x >A family of wireless LAN standards which serve as the basis for the WiFi certification ‘> Defines wireless LAN physical layer and MAC sublayer specifications, > Competes against the HyperLAN standards in Europe but is gaining wide populzrity over the latter Wireless LAN Standards (Historical) ‘= comparison Tester i oer aa [as —— IEEE 802.11x secesrons IEEE 802.11x BSS withow AP Indepondont 855 BSS whan AP Standard 855 IEEE 802.11x Server or Gateway IEEE 802.11x IEEE 802.11x a) > specoum usage . S29 ome a kamen IEEE 802.11x IEEE 802.11x > Reading Assignment «= Difference between the physical layer implementation of IEEE 802.119, 802.11n and 802.11ac = What is Wi Direct? ee ‘= How do you implement an adhoc WiFi Network? ‘The more users, the leser the banduidth per user and ‘greater latency > Common Characteristics 1» Half-duplex operation ‘Only one device can transmit at any given time ‘> Sample Wif Device Specs 1 Capacty/range relations 1 See a WiFi equipment manufacturer website sconnchy play decrentes wth tance ‘= Different transmission rates, same channel “2 Ths i detrimental to higher rate users: IEEE 802.11x IEEE 802.11x 1 Transmisonadrovedgenens Te en ton probleme aoe re tin '¢ Required since stations don’t listen while transmitting ‘each others communication range Toga poten ecm tr ae + Protea overhead Dheteeetaueh aes ris man ek fede ee rote SO Seana hoor = Affected by the following factors ‘etree, mare enya oy Tress apart BY ‘Eonar cen = Used Caner Sense Miltpe Access wth Colson Avoidance «csmaca) 1 no wtf calon bd exponen ame » second Methoe| 1 Acero the poly at ot tars ea tae "Vester NA) wht seg ston St asd on fH ‘tres ag annie IEEE 802.11x 1» RTC/CTS with NAV mechanism IEEE 802.11x SSD ‘Aide Adair Ad - IEEE 802.11x ‘Point Coodination Function (PCF) * Acentra autor (the base station) cork are trae > Fragmentation 1» Data is fragmented to allow shorter retransmission of corrupted data > Frame Format IEEE 802.11x Case 4 What is WDS? > Wireless Distribution System 1 ents ange of WF 1 can paterm nding or repeating Trcion|

You might also like