Business Data Communications Chapter Three Data Link Layer Fundamentals
Primary Learning Objectives Understand the function of the data link layer Distinguish Logical Link Control from Media Access Control Describe the two types of flow control Explain line discipline Define the components of error control Recognize two methods of delineating data in a bit stream Identify devices and components associated with the data link layer
The Data Link Layer – Its Function Sits above the physical and below the Network Layers Formats data bits into frames Has two components: Logical Link Control – 802.2 Media Access Control – 802.3 for Ethernet Is responsible for: Line discipline Flow control Error control
Components of the Data Link Layer
Logical Link Control (LLC) Designated by the IEEE as 802.2 and sits above the Media Access Control Provides three types of frame delivery service using protocol data units: Type 1 – connectionless without acknowledgement, the most used delivery service Type 2 – connection-oriented with acknowledgement Type 3 – connectionless with acknowledgement
Logical Link Control (LLC) LLC supports three types of frames: I – Information is connection-oriented S – Supervisory manages the Information frames U – Unnumbered used by connectionless services and terminates connection-oriented services Only Type 2 delivery service uses all three types of frames
Logical Link Control (LLC)
Media Access Control (MAC) Has various IEEE designations, with the most common being 802.3 for Ethernet Determines how devices share a common circuit Falls into one of two categories: Contention (802.3, Ethernet, for example) Controlled access (802.5, Token Ring, for example) FDDI, another form of controlled access, is an ANSI/ITU-T standard
Media Access Control – 802.3
Media Access Control – 802.5
Media Access Control (MAC) Standard 802.3 networks using hubs and bridges can suffer from significant collision impairment under high traffic: Modern 802.3 networks using switching technology have greatly eased this problem Switches do not change the underlying 802.3 architecture
Media Access Control (MAC) Format of a MAC Protocol Data Unit (PDU)
Flow Control Prevents a sender from overwhelming a receiver with traffic: A sender and receiver each have a memory area in which they can store frames This memory is sometimes referred to as a buffer A sender can overwhelm, or overflow, a receiver’s memory buffer without proper flow control If an overflow occurs, data would likely be lost
Flow Control Two common forms of flow control are: Stop-and-wait Each single frame sent requires receipt of one acknowledgement Sliding windows The sending of multiple frames requires a single acknowledgement returned
Flow Control Stop-and-wait: Sliding windows: Most efficient for messages containing a few large frames that traverse short links Requires one acknowledgement for each frame sent Sliding windows: Most efficient for messages containing many small frames that traverse long links Allows for one acknowledgement for multiple frames
Stop-and-Wait Flow Control
Sliding Windows Flow Control
Line Discipline Can be viewed as a “polite” means of controlling a conversation between communicating devices Associated with two types of network environments: Point-to-point between communicating devices using half- or full-duplex circuits Multipoint with communicating devices going through a central controlling device The central control device is often a mainframe with connected terminals
Error Control No system is perfect; errors should be expected Errors can result when data is lost, corrupted, or damaged, making error control critical Error control has two components: Error Detection Error Correction The two components are equally important
Error Detection Common error-detection methods include: Parity checking 50% probability of detection Longitudinal redundancy checking 98% probability of detection Checksum checking 99.6% probability of detection Cyclical redundancy checking 99.9%+ probability of detection
Parity Checking An extra parity bit is added to the byte Assuming even parity: 10000010 – data sent 10000110 - data received Error detected on receiver side (single bit) 10011010 – data received No error detected on receiver side (multiple bit) Simple parity detects only single bit errors
Longitudinal Redundancy Checking -- LRC Longitudinal literally means “lengthwise” The sender, for each byte in the message, calculates a parity value, creating an additional block check character or BCC As with parity checking, the parity value is odd or even The BCC is added to the end of the message block The receiver performs the same lengthwise LRC computation If the receiver’s calculated BCC does not equal the sender’s calculated BCC, the receiver assumes a transmission error
Longitudinal Redundancy Checking 01000010 01011001 01010100 01000101 – Before BCC
Longitudinal Redundancy Checking 01000010 01011001 01010100 01000101 00001010 – After BCC The BCC added to the end of the data block.
Checksum Checking – CC The message sender: The message receiver: Evaluates each binary byte in the message to its decimal value Totals the decimal values of all bytes Divides the total by 255, creating a remainder Using the remainder for the CC, adds the CC to the end of the message block The message receiver: Performs the same byte-by-byte calculation and creates his own CC Compares his calculated CC to the sender’s Assumes a transmission error if the two CC values differ
Checksum Checking – CC 308 / 255 = 1.21 CC = 21 66 89 84 69 308 Char Col 1 Col 2 Col 3 Col 4 Col 5 Col 6 Col 7 Col 8 B 1 Y T E TOTAL Decimal Equivalent 66 89 84 69 308 308 / 255 = 1.21 CC = 21
Error Control The most common error correction technique is to simply retransmit the data in error Easy, but requires time for the retransmission A second error correction technique is called forward error correction: The core message is sent along with redundant data bits The redundant data bits can, if necessary, be used by the receiving device to correct errors on site without retransmission However, forward error correction results in inefficient use of a circuit if too many redundant data bits are sent and not used
Data Delineation A transmitted bit stream contains not only the core message but control information as well Control information could include: Source address Destination address Length of message field Error control data Other “non-core” information Data delineation differentiates between core and other data
Data Delineation Two key methods providing data delineation are: Asynchronous data link protocols Synchronous data link protocols Asynchronous protocols: Are used mostly by mainframes and their connected terminals Provide byte-by-byte delineation Synchronous protocols: Are used in LANs, BNs, MANs, and WANs Provide delineation for groups of bytes
Data Delineation Popular asynchronous protocols include: XModem YModem ZModem Kermit Asynchronous protocols require that every data byte have a start and stop bit before and after it Generally less efficient than synchronous protocols
Asynchronous Transmission
Data Delineation Synchronous protocols are either bit- or byte-oriented Bit-oriented protocols are more flexible: They do not require a predetermined “byte” character format, such as EBCDIC or ASCII They are more complicated Bit stuffing may be required HDLC is a formalized bit-based protocol Byte-oriented protocols: Are based upon known “byte” based data Ethernet is a very common byte-based protocol
Simple Synchronous Transmissions
Data Link Layer Devices and Components A Network Interface Card is: A component rather than device Essential to connect a device to a network Bridges: Link segments of the same logical network Filter traffic, and so can improve network performance Switches: Offer more functionality than hubs and bridges Provide point-to-point connections to devices plugged into them Have transformed how standard Ethernet is configured
A 3Com Network Interface Card (NIC) NICs have a physical address NIC addresses must be unique NIC addresses can be bypassed, or overridden, by software, but care must be taken when doing this to avoid address duplication
A Linksys Wireless Bridge Bridges filter network traffic
A Bridge Filtering Traffic
A standard Cisco Switch Switches have mostly replaced hubs in modern Ethernet networks
In Summary The data link layer: Is stacked above the physical and below the network layers Formats data bits into units called frames Is composed of two stacks, the logical link and the media access controls Performs error control Has devices such as bridges and switches with which it is particularly associated