1. Section 3 The OSI Data Link Layer CSIS 479R Fall 1999 “Network +” George D. Hickman, CNI, CNE

2. Objectives Identify the basic purpose of the OSI Data Link layer Identify the characteristics of the two logical topologies Identify the characteristics of the three media access methods Describe how addresses are defined and managed at the Data Link layer

3. Objectives (con’t) Describe the transmission synchronization techniques used at the data link layer Describe the connection services implemented at the Data Link layer Describe the IEEE 802.x standards Describe the 802.3 standard and Ethernet

4. Objectives (con’t) Describe the 802.3u Fast Ethernet standard Describe the 802.5 and Token Ring standards Describe the Fiber Distributed Data Interface (FDDI) standard Describe commonly used wide area networking protocols

5. Data Link Layer Media Access Control Sublayer controls how transmitters share single media Logical Link Control Sublayer establishes and maintains device to device link Organize Physical layer’s bits into frames Detect and sometimes correct errors Control Data Flow

6. Data Link Layer Identify Computers on the network Data Link layer header contains: Source and destination addresses Frame length information Indication of upper layer protocols involved

7. Data Link Devices Network Connectivity Devices Bridges Switches NICs

8. Data Link--MAC Logical Topology Process Bus and Ring methods Media Access Process Contention, Token Passing, & Polling methods Addressing Process Physical device method

9. Data Link -- LLC Transmission Synchronization Process Asynchronous, Synchronous, Isochronous methods Connection Services Process LLC-level flow control Error control

10. Logical Topology The actual signal path As opposed to the Physical Topology, which is the physical layout of wires Logical and Physical paths do not have to be the same Token Ring Example Physical star Logical Ring

11. Media Access Control Contention Devices transmit when ever they want Causes collisions Carrier Sense Newer contention scheme Listens to media, transmits if no signal detected CSMA Carrier Sense, Multiple Access

12. Carrier Sense CSMA Collision detection and retransmission is the responsibility of a “higher layer” protocol Waiting and overhead of going up and down OSI models make CSMA less effective CSMA/CD Adds collision detection at or below the DL layer by sensing cable before and after transmitting After collisions, wait random time and retransmit Good for bursty traffic

13. Token Passing The Token (a small frame) is passed to give media access control. Only devices with the token may transmit Devices know where they get token from and where they pass it to IEEE 802.5 Token Ring Standard Token passing access control, physical or logical ring topology FDDI Good for time sensitive (voice, video) traffic or heavily populated networks

14. Polling Systems One device (the controller, primary, or master) is media access administrator Queries other devices (secondaries) in a predefined order to see if they need to transmit Ideal for networking time-sensitive devices like automation equipment

15. Addresses–Defined and Managed Data Link layer is concerned with the physical device address or MAC address Most DL layer implementations place the source and destination addresses in the frame header The frame is sent to every device on network, which reads header and reads or ignores the data as needed Bridges use these addresses to let frames “through” or not Switches use these addresses to know which port to send data frames to

16. Transmission Synchronization Techniques—DL layer The physical layer synchronization was bits The Data Link layer synchronization is the coordination of frame transmission

17. Asynchronous Each device has own clock, not synchronized with the other Start and Stop bits used Good for random interval transmissions Parity bit can be added to detect some errors Even parity Parity bit is set to give an even number of 1 bits/byte Odd parity Parity bit is set to give an odd number of 1 bits/byte May not detect multiple bit errors

18. Synchronous Devices responsible for a framing clock Can be separate channel Or use SYN or SYNC characters for “start” A CRC value can be put near end for error checking Both devices must use same algorithm to compute CRC

19. Isochronous A clock signal is sent out to all network devices to create time slots Other devices may fill “slots” with data Clock signal is not provided with every frame (like asynchronous) or at start of a string of data (like synchronous)

20. DL layer implementations of Connection Services Unacknowledged connectionless services Send and receive frames with no flow, error, or packet sequence control Connection-oriented services Flow, error, and packet sequence control provided by use of acknowledgments Acknowledged connectionless services Flow and error control provided by acknowledgements between point to point transmissions

21. LLC level Flow Control Guaranteed Rate Flow control A rate is agreed upon before transmission and is maintained as long as the transmission lasts Window flow control Static An acceptable window or buffer size is determined. That number of frames is maximum sent without an acknowledgement Dynamic Window or buffer size can be adjusted. A choke packet is sent by receiver when the buffer exceeds a specified level, slowing down the sender. Transmission is slowly increased until another choke packet is sent.

22. Error Control How lost or scrambled frames are handled Sending device receives NAK (or nothing) Checksums do not match Packet size is off (too small) Can be caused by Noise, Interference, or Distortion Can be caused by a buffer overflow

23. IEEE 802.x standards 1980 IEEE defined LAN standards for Physical and Data Link layers 802.1 Allows 802 compliant device to speak with another 802 device on another LAN or WAN 802.2 Defines LLC sublayer of Data Link layer

24. IEEE 802.x standards 802.3 Physical layer specifications Baseband/broadband Media type Topologies Data rate Three part naming convention Speed (megabits per second) BASE or BROAD Special designator or effective distance 10BASE2

25. IEEE 802.x standards 802.4 Factory and Industrial automation needs Physical bus topology Token Passing media access Baseband or Broadband media 75 Ohm CATV-type cable or optical fiber 802.5 Based on IBM Token Ring Token Passing media access 1, 4, or 16 Mbps No specific transmission media or physical topology mandated (IBM Token Ring mandates both.)

26. IEEE 802.x standards 802.6 Distributed Queue Dual Bus (DQDB) 802.7 Standards for broadband communications 802.8 Fiber Optic standards 802.9 Isochronous Ethernet (voice/data)

27. IEEE 802.x standards 802.10 Used with encryption key information 802.11 Used with wireless LAN implementations 802.12 100 Mbps physical star, contention based (100VG-AnyLAN)

28. Ethernet Combination of 802.2 and 802.3 Designed as simple, low access-overhead LAN architecture See diagram on page 3-36 Can use Thick or thin Co-axe BUS Topology Can use Twisted Pair, or fiber optic cable, using either switches or hubs STAR Topology

29. 10BASE5 Thick Coaxial cabling NICs use external transceiver 50 Ohm terminator both ends, 1 grounded 500 Meter maximum segment length 100 devices per segment maximum (incl repeater) 3 populated segments maximum 2.5 M between taps / 5 M (max) tap to node

30. 10BASE2 Thin Coaxial NICs use internal transciever 50 Ohm terminator both ends, 1 grounded 185 M maximum segment length 30 devices per segment max (incl repeaters) 3 populated segments maximum.5 M minimum between T connectors

31. 10BASE-T Twisted Pair 100 M maximum segment length 1,024 maximum workstations (theoretical) 4 repeaters maximum between communicating devices

32. 5-4-3 Rule Coaxial 5 cable segments maximum 4 repeaters maximum 3 segments populated UTP 5 cable segments maximum 4 hubs maximum

33. IEEE 802.3u (Fast Ethernet) Physical and Logical Topologies Physical hierarchical star / Logical Star Media Independent Interface (MII) 100BASE-TX, 2 pair Cat 5 UTP 100M/segment 100BASE-T4, 4 pair Cat 3+ UTP 100M/segment 100BASE-FX, 2 strand FO, 412-10,000M/segment Auto Negotiation (AUTONEG) 10 or 100 Mbps auto negotiated Media Access Control (MAC) CSMA/CD

34. IEEE 802.5 Token Ring Specifications for Physical Layer and MAC sublayer of Data Link Physical Star Logical Ring Topology 802.5 does not specify cabling type Cable lengths differ with media used (page 3-46) 3 cable segments per series 33 MSAUs maximum 802.5 specifies 250 nodes maximum IBM STP specifies 260 nodes IBM UTP specifies 72 nodes All network devices must run at same speed (4 or 16 Mbps) unless connected by a bridge

35. 802.5 Token Ring MAC Token Passing Special packet allowing device to transmit Device receives packet, transmits a frame. When frame returns to sender, sender puts a new token out on ring. Early token release-creates/releases new token immediately after sending data frame Active monitor performs maintenance on ring

36. Beaconing Allows some automatic error recovery Upon ring break, stations send beacon frames until they receive a beacon frame from an “upstream neighbor” Soon only one station (after break) is beaconing MSAU attempts to reconfigure ring around the break

37. FDDI Standard Fiber Distributed Data Interface Physical Layer and MAC sublayer+SMT Assumes 802.2 (LLC sublayer specification) Fills need for secure high bandwidth Backbone implementation (connects LANs) Computer room networks (mainframes, minis) High Data Rate LANS (CAD/video needs)

38. FDDI and 802.5 Both use Token passing MAC Both physical star logical ring Both (can) use fiber-optic media FDDI has higher maximum data rate

39. FDDI Network Design 2 counter rotating rings Primary caries data, secondary management Secondary becomes primary if P. media fails 1000 workstations max, 200 km total cable 500 / 100 incase of media failure Multi-mode fiber optic cable/62.5 micrometer Repeater required every 2 km or less

40. FDDI Network Design (con’t) Class A Workstations Connect to both rings. Higher fault tolerance Class B Workstations Connects to primary ring. Can’t reconfigure FDDI ALWAYS releases new token at end of transmitted frames, so there may be multiple frames on network at once. See figure 3-27 on page 3-53

41. WAN Protocols Dial up connections SLIP (Serial Line Internet Protocol) Physical layer protocol Not a strict standard, may not work PPP (Point-to-Point Protocol) Physical and Data Link layer protocol Allows Dynamic IP addressing Support multiple protocols on same link Password login Error control

42. WAN Protocols (con’t) PPTP (Point-to-Point Tunneling Protocol PPP extension Encapsulates other protocols for IP transmission Corporations use Internet to connect their LANs Can read multi-protocol packets X.25 Attaching computer to a packet-switched network Physical, Data Link, Network Layers SprintNet, Tymnet, GTE

43. WAN Protocols (con’t) Frame Relay Designed for high speed bursts on digital network No error checking while transmitting, so faster Error checking at receiving point Ethernet, X.25, Token Ring common 56 kpbs to 1.544 Mbps common (56K, T-1, T-3) Physical and Data Link layers Purchased by CIR (Committed Information Rates), the minimum guaranteed capacity of virtual circuit

44. WAN Protocols (con’t) ISDN B-ISDN Integrated Services Digital Network Channel A 4 KHz analog channel Channel B 64 Kbps digital channel Channel C 8 or 16 Kbps digital Interconnects X.25, PPP, frame relay Physical, Data Link, Network Layers Popular with SOHO

45. WAN Protocols (con’t) ATM Asynchronous Transfer Mode B-ISDN and Cell Relay (Cell is 53 byte block) Considered a LAN and WAN protocol Primarily Data Link and Network Layer Designed to be independent of Physical Layer for faster processing speeds 155 Mbps and 622 Mbps specified. 10Gbps expected FDDI or others specify the Physical layer