Chapter 13 and 14 – Multiple Access, Local Area Networks

Spring 2006Computer Networks2 A Link and Types of Links  A link is a communication channel established through some kind of physical medium (guided or unguided)  Type of links with concern to line configuration  Point-to-point link  Consists of a single sender on one end and a single receiver on the other end.  Broadcast link - Multipoint  Multiple sending and receiving nodes all connected to the same shared broadcast channel  When one node transmits, the channel broadcast the frame and each of the other nodes receives a copy

Spring 2006Computer Networks3 Multiple Access Protocols and LANs  In case of a broadcast link, there must be rules for sharing the channel. Otherwise no-one can “hear” no- one.  Media access protocols are protocols that resolve the multiple access problem  Broadcast links and media access protocols are used in computer networks that are local (span a single or several buildings)  LAN (Local Area Network)

Spring 2006Computer Networks4 Examples for Multiple Access Channels Sharing a room (Conversation of many persons) Sharing the wire Sharing satellite channel

Spring 2006Computer Networks5 Media Access Control Protocols  Three types of sharing a single channel  Channel partitioning or using some kind of multiplexing techniques  Multiple access by allowing contention  Collisions are allowed  Recover from collision  Ethernet is an example  Taking turns in accessing the channel  The access is controlled by a circulating token  Token ring is an example

Spring 2006Computer Networks6 Evolution of Contention Protocols  Developed in 1970 to be used on radio LAN on Hawaiian islands. The access to the channel is random  Improvement to Aloha: Start transmission only at fixed time slots  Carrier Sense Multiple Access: Start transmission only if no transmission is ongoing  CD=Collision Detection: Stop ongoing transmission if collision is detected Aloha Slotted Aloha CSMA CSMA/CD

Spring 2006Computer Networks7 CSMA Operation  Carrier sensing  the ability to detect idle/busy channel quickly  The logic followed: There is no point of transmitting if the channel is busy (another station is transmitting)  The chances for collisions are reduced  The possibility of collision still exists because of the propagation time  The signal sent by a single device is sent immediately by the closest devices. For the others to sense it, some time is needed

Spring 2006Computer Networks8 Persistent vs. Non-persistent CSMA

Spring 2006Computer Networks9 Normal Operation and Collisions A B D Data C Address mismatch packet discarded Address mismatch packet discarded Address match packet processed Send data to node D Transmitted packet seen by all stations on the LAN (broadcast medium) The packet carries the address of the destination A B C D Collision Data transmission for A Data transmission for C

Spring 2006Computer Networks10 CSMA/CD  Sense for carrier.  If carrier present, wait until carrier ends.  Send packet and sense for collision.  If no collision detected, consider packet delivered.  Otherwise, abort immediately, perform “exponential back off” and send packet again.  CSMA/CD is used in traditional Ethernet LAN

Spring 2006Computer Networks11 Exponential Back-off  When a sender detects a collision, it sends a “jam signal”.  Jam signal is necessary to make sure that all nodes are aware of the collision  Length of the jam signal 48 bits  When collision is detected, the sender resends the signal after a random time  The random time is picked from an interval of 0 to 2 N x maximum propagation time  N is the number of attempted retransmission  Length of the interval increases with every retransmission

Spring 2006Computer Networks12 Local Area Networks (LANs)  A computer network in a limited geographical area, a single building or several close to each other buildings  LANs are privately owned and built by the companies  Generally less expensive than WAN for comparable speed  LAN technologies use multiple access channels  Ethernet is the most common LAN technology

Spring 2006Computer Networks13 Traditional Ethernet  Work started back in 1973 by Bob Metcalfe and David Boggs from Xerox Palo Alto Research Center, as an improvement of the ALOHA  Experimental Ethernet implemented in  Cooperative effort between Digital, Intel, and Xerox produced Ethernet Version 1.0 in  Ethernet was adopted with modifications by the standards committees IEEE and ANSI 8802/3.  Structure of Ethernet frame (Length)

Spring 2006Computer Networks14 Structure of Ethernet Frame  Preamble:  7 bytes with pattern followed by one byte with pattern  Used to synchronize receiver, sender clock rates  Addresses: 6 bytes, the frame is received by all adapters on a LAN and dropped if address does not match  Type: 2 bytes, is actually a length field in  CRC: 4 bytes, checked at receiver, if error is detected, the frame is simply dropped  Data payload: maximum 1500 bytes, minimum 46 bytes. If data is less than 46 bytes, pad with zeros to 46 bytes

Spring 2006Computer Networks15 Annimation for Better Understanding  The following link will lead you to several annimations that explain important issues in the area of networking.  Play annimation 6.1 and 6.2 to understand how Ethernet works.

Spring 2006Computer Networks16 Network Interface Card (NIC)  Each device on Ethernet network has its own interface card (NIC) to connect to the network  The NIC is usually plugged into the device and has a 6 bytes (48 bits) physical address  The physical address is normally written in hexadecimal notation  C-4D-1B (example address) NIC for a desktop NIC for a laptop

Spring 2006Computer Networks17 Ethernet Addressing  Each station recognizes three classes of addresses.  Own address  Broadcast address (all 1's)  Optionally, one or more multicast addresses  Major reason for broadcast is address discovery. Brodcast Ethernet address is all 1s, or in hexadecimal  FF : FF : FF : FF : FF :FF  Multicast addresses are used for specialized link  layer functions.  Ethernet addresses are unique  First three bytes assigned to manufacturer by IEEE, the other three bytes assigned by the manufacturer

Spring 2006Computer Networks18 Classic 10Mbps Ethernet  Four different implementation at the physical layer for the baseband 10Mbps Ethernet  Thick Ethernet (10base5) – obsolete  Thick coaxial cable (0.5” diameter)  500meter max length, bus physical topology  Thin Ethernet (10base a) - obsolete  RG58 coaxial cable  185 meter max length, bus physical topology  Twisted Pair Ethernet (10baseT 802.3i)  4 pair UTP (unshielded twisted pair) cable  100 meter max length, star physical topology  Fiber-link Ethernet (10Base-FL)  Fiber cable connected to external transceiver  Star topology is used

Spring 2006Computer Networks19 Physical Layer of the Ethernet  PLS (Physical Layer Signaling) encodes and decodes data  Ethernet uses Manchester encoding  AUI (Attachment Unit Interface) – interface between PLS and medium dependent interface  MAU (Medium Attachment Unit) or transceiver  MDI (Medium Dependent Interface) is a piece of hardware connecting the transceiver to the medium

Spring 2006Computer Networks20 10Base5 (Thick Ethernet) transceiver Resistor terminator Ethernet medium (cable) Drop cable

Spring 2006Computer Networks21 10Base2 (Thin Ethernet) Terminator BNC T connector BNC connector To next equipment or terminator Thin coaxial cable

Spring 2006Computer Networks22 10BaseT (Twisted Pair Ethernet) High-Speed Backplane or Interconnection fabric hub switch  The central device can be a hub or a switch

Spring 2006Computer Networks23 Hub Concept  Separate transmit and receive pair of wires.  The hub retransmits the signal received on any input pair onto all output pairs.  Essentially the hub emulates a broadcast channel with collisions detected by receiving nodes.

Spring 2006Computer Networks24 Ethernet Evolution  Introducing bridges  Unlike a hub, a bridge is capable of filtering frames  Each port of the bridge is connected to a single segment of LAN  Capable of learning which the stations are connected to which ports  Separates collision domains and therefore increases bandwidth  Introducing switches  Similar function as bridges  Contain bigger number of ports  A single device can be attached to a port

Spring 2006Computer Networks25 Bridged vs. Switched Ethernet Bridge ABCD E F Switch

Spring 2006Computer Networks26 Fast Ethernet  Go from 10mbit/s to 100mbit/s  3 competing standards:  100Base-TX  100Base-T4  100VG-Anylan  100Base-T4 and 100VG-Anylan are the losers (were not very well accepted).  100Base TX is the winner. It is almost a standard everywhere.

Spring 2006Computer Networks27 100Base - TX  100 Mbps over 2 pairs of wire (just like 10base-T)  Requires Category 5 UTP wiring or STP  De facto standard today  Very small price difference with 10Mbps-only equipment  Has clearly won over 100baseT4 and 100VG- Anylan by now

Spring 2006Computer Networks28 100Base-FX  Fast Ethernet with fiber optic cables  Uses two optical fibers, one for transmission and one for reception

Spring 2006Computer Networks29 Gigabit Ethernet  Provides speeds of 1000 Mbps (i.e., one billion bits per second capacity) for half-duplex and full-duplex operation.  Uses Ethernet frame format and MAC technology  CSMA/CD access method  Backward compatible with 10Base-T,100Base-T and 100BaseTX  Can be shared (hub) or switched

Spring 2006Computer Networks30 Gigabit Ethernet Implementations  Fiber  1000 Base – SX  Short wavelengths, two fiber-optic cables  1000 Base – LX  Long wavelengths, two fiber-optic cables  Copper  1000 Base – CX  Uses shielded twisted pair copper jumpers  1000 Base – TX  Uses category 5 twisted pair copper cable

Spring 2006Computer Networks Base - T  Four pairs of Category 5 UTP  IEEE 802.3ab ratified in June  Category 5, 6 and 7 copper up to 100 meters  Uses encoding scheme 4D-PAM5  Five level of pulse amplitude modulation are used  Complicated technique