The Data Link Layer Data Link Types Data Link Protocols Access Methods Data Link Sub-layers 30/11/10 14-Datalink

Data Link Types Point-to-Point Point-to-Multipoint Links through Switched Networks Broadcast 30/11/10 14-Datalink

Binary Synchronous Communication (BSC) Developed by IBM, standardised by ISO (Basic Mode) Connection-oriented Half Duplex Point-to-Multipoint or Point-to-Point Asymmetric (Master-Slave polling) Synchronous Character-oriented Byte Stuffed VRC/LRC Error Checking (CRC for transparent mode) Idle RQ (Stop and Wait) Explicit Negative Acknowledgements (NAKs) 30/11/10 14-Datalink

High-level Data Link Control (HDLC) Developed by ISO (based on IBM’s proprietary SDLC protocol) Connection-oriented Full Duplex Synchronous Bit-oriented Bit Stuffed CRC Error Checking Two classes (Six modes) Unbalanced. E.g. Normal Response Mode Balanced. E.g. Asynchronous Balanced Mode 30/11/10 14-Datalink

HDLC Normal Response Mode Point-to-Multipoint (or Point-to-Point) Asymmetric (Master-Slave polling) Continuous RQ (Sliding Window) with Go Back N No Explicit Negative Acknowledgements Used for Mainframe-Terminal networks with Point-to-Multipoint or Point-to-Point Private Circuits 30/11/10 14-Datalink

HDLC Asynchronous Balanced Mode Point-to-Point Symmetric (Equal status for both ends) Continuous RQ (Go Back N or Selective Reject) Positive Acknowledgements - Receiver Ready (RR) Explicit Negative Acknowledgements Reject (REJ) for Go Back N Selective Reject (SREJ) Flow Control - Receiver Not Ready (RNR) Often used as Layer 2 protocol for IP over Point-to-Point Private Circuits 30/11/10 14-Datalink

Link Access Protocol Balanced (LAPB) Subset of HDLC Asynchronous Balanced Mode Only supports Go Back N ARQ Also supports Asynchronous Balance Mode Extended which provides 7 bit sequence numbers (allowing window sizes of 128) for high speed or long delay circuits (E.g. satellite) Used as the layer 2 protocol by X.25 packet switched networks 30/11/10 14-Datalink

HDLC Frame Formats Unnumbered Frames Information Frames Used for Connection Control with no sequence numbers at all Information Frames Used to carry data with sequence numbers, but also contains field for sequence number of acknowledged frames (piggybacking) Supervisory Frames Used for Flow and Error Control with no data field but do have sequence numbers of acknowledged frames 30/11/10 14-Datalink

Point-to-Point Protocol (PPP) Based on HDLC Connection-oriented Full Duplex Point-to-Point Symmetric Synchronous Character-oriented Byte Stuffed CRC-16 or CRC-32 Error Checking No flow control or error recovery. Relies on higher layers 30/11/10 14-Datalink

PPP Header © Tanenbaum, Prentice Hall International 30/11/10 14-Datalink

Point-to-Point Protocol PPP is used for Internet access over a switched network (E.g. Public Switched Telephone Network) As well as supporting a frame protocol similar to HDLC, it supports other sub-layer protocols Link Control Protocol Network Control Protocol Two alternative authentication protocols (PAP and CHAP) 30/11/10 14-Datalink

Link Control Protocol (LCP) Encapsulated in a PPP frame Establishes, maintains, configures, tests and terminates links Determines quality of link, closing if necessary Negotiates options Maximum Receive Unit Authentication Protocol Compression 30/11/10 14-Datalink

Link Control Protocol Phases © Tanenbaum, Prentice Hall International 30/11/10 14-Datalink

Network Control Protocol (NCP) Encapsulated in an PPP frame The NCP is used to configure the Network Layer protocol A specific Network Control Protocol is used dependent on which Network Layer is supported For IP, the IP Control Protocol (IPCP) is used. For IPX, it is IPXCP IPCP will also configure an IP address, if necessary 30/11/10 14-Datalink

Password Authentication Protocol (PAP) PAP is encapsulated in a PPP frame 2 way handshake performed once at link establishment Sends username password repeatedly in plaintext and hence is not secure The ISP must close the link, if it is incorrect 30/11/10 14-Datalink

Challenge Handshake Authentication Protocol (CHAP) 3 way handshake (challenge, response and accept/reject) Password is not transmitted across the network and is hence secure Challenge packet contains a challenge value (usually a few bytes) Response applies a pre-defined function to the challenge value and the user’s password Challenger performs the same calculation and checks the response Challenge can also be repeated at any time 30/11/10 14-Datalink

Switched Links Data Link Protocols, such as HDLC and PPP are often used to carry data over links through switched networks such as: Public Switched Telephone Network (PSTN) Integrated Services Digital Network (ISDN) ADSL Broadband Access Network 30/11/10 14-Datalink

Broadcast Links Data Link Protocols are often used over broadcast links, but the Point-to-Point and Point-to-Multipoint protocols will not support broadcast links where there are multiple transmitters The major problem to overcome with broadcast links such as radio and Ethernet LANs is contention (resulting in collisions) accessing the link. Only one device can be permitted to access a broadcast link at any one time The means of overcoming contention is called the Access Method 30/11/10 14-Datalink

Access Methods There are several strategies for handling contention: Frequency Division Multiple Access (FDMA) Time Division Multiple Access (TDMA) Code Division Multiple Access (CDMA) Carrier Sense Multiple Access (CSMA) Token Passing Spread Spectrum 30/11/10 14-Datalink

Frequency Division Multiple Access (FDMA) FDMA solves the contention problem by dynamically allocating a different frequency to each station that wants to transmit Allocations are requested via a common control channel which has its own access method FDMA was used on 1st Generation (analogue) mobile phone networks A similar method, called Wavelength Division Multiple Access is used on all optical LANs 30/11/10 14-Datalink

Time Division Multiple Access (TDMA) TDMA solves the contention problem by dynamically allocating different timeslots to each station that wants to transmit Allocations are requested via a common control channel which has its own access method TDMA is used on 2nd Generation (GSM) mobile phone networks 30/11/10 14-Datalink

Code Division Multiple Access (CDMA) CDMA is a technique that allows all transmissions to share the same channel by marking each communication with a unique ID (the code) The receiver is able pick out the one communication it wants to receive by detecting the code A good analogy is a conversation between two people in a noisy party where lots of other conversations are going on around them, but a listener can focus in on just one speaker CDMA is used by 3rd Generation mobile phone networks 30/11/10 14-Datalink

Carrier Sense Multiple Access Carrier Sense means listen to make sure that nobody else is transmitting before you start CSMA/CD (with Collision Detection) - used by Ethernet Also keep listening after starting to transmit to make sure that nobody else has transmitted while you were Back off for a random period of time, and retransmit CSMA/CA (with Collision Avoidance) - used on Wireless LANs Send a short “request to send” frame and wait for “clear to send” frame before transmitting data All other stations will see this and will hold off Any collision of CTS/RTS frames will be dealt with by backing off for a random period 30/11/10 14-Datalink

Token Passing A single token is circulated around the network No station can transmit unless it has the token Once the station verifies that its transmission has been received it releases the token for another station to use The network needs to be continually monitored to ensure that the token doesn’t get lost and must be recreated when necessary Token passing networks unlike CSMA networks can guarantee fair access with a fixed maximum delay The IBM Token Ring LAN is an example of a network that uses a token passing access method 30/11/10 14-Datalink

Spread Spectrum Spread spectrum techniques, which share a wide frequency band (much wider than the original signal), are used to support multiple access in Wireless LANs. There are many different techniques Code Division Multiple Access (CDMA) frequency of transmitted signal is made to vary according to a defined pattern (code) Frequency Hopping Spread Spectrum (FHSS) Hops between frequencies according to pseudo-random sequence Direct Sequence Spread Spectrum (DSSS) codes one bit as n bits Orthogonal Frequency Division Multiplexing (OFDM) used in IEEE 802.11a – splits signal across many frequencies which do not interfere with each other High Rate Direct Sequence Spread Spectrum (HR-DSSS) used in IEEE 802.11b – higher speed version of DSSS 30/11/10 14-Datalink

Data Link Sub-layers The IEEE who have standardised LAN protocols have divided the Data Link Layer into two sub-layers Logical Link Control (LLC) Common to all LANs Independent of physical media, topology and access method Media Access Control (MAC) Specific to particular media, topology and access methods 30/11/10 14-Datalink

Logical Link Control Functions Identifies and encapsulates Network Layer Protocols Flow Control Error Control 30/11/10 14-Datalink

Media Access Control (MAC) Functions Access to the physical media Physical (MAC) Addressing Ordered Delivery Managing network topology (including bridging) 30/11/10 14-Datalink

MAC Addresses MAC (or Physical) addresses are burned into Network Interface Cards They uniquely identify the card They are 48 bits (6 bytes) long They are split into two 24 bit parts Organisationally Unique Identifier (OUI) which identifies the manufacturer of the card and is assigned by the IEEE A manufacturer assigned serial number MAC addresses are usually shown as 12 hexadecimal digits, grouped in pairs to represent each byte and separated by colons or hyphens. E.g. 00:06:f5:00:1f:2c 30/11/10 14-Datalink