Data link layer – Ethernet
•
•
•
•
Responsible for error free data flow from one network to another.
Organizes data into frames to be sent to layer 1 for transmission
Adds cyclic redundancy check (CRC) to data
Broken down into 2 layers (logical link control)(media access control)
LLC(logical link control)
•
•
•
Establishes and maintains links between communicating devices
IEEE 802.2 standard
All devices that can establish a LLC session are said to be on the same network segment or local area network(LAN)
Media access control (MAC)
•
•
Provides for shared access to network adapter and talks directly to network adapter
Common standards
• 801.3 CSMA.CD (Ethernet)
• 802.4 token bus
• 802.5 token ring
• 802.12 Demand Priority
MAC address is the address of the network adapter at the data link layer
•
Basic media access control
•
•
Contention
Contention based media access methods allow any device to talk at will. Each device determines on its own if it is OK to use the media. If two devices use the media a the same time a collision occurs.
Example: Wired Ethernet
Controlled
Controlled access is governed by the principle that devices can not talk until it is their turn.
Some device or protocol governs who's turn it is to talk.
Polling: Controlling device checks each device in turn to see if it wants to communicate.
Hub polling (token passing). A data structure called a token is passed around the network, the device with the token has the media
Example: Token ring, Wireless Ethernet in infrastructure mode.
•
•
•
•
•
•
Error Control – Causes and Prevention
Error Control – Detection
Errors are typically detected with a mathematical process. Common methods are
Parity Checking
Checksum
Cyclic Redundancy Check
Parity
Simple and easy to process
Each block of data sent has a parity bit appended.
Parity may be odd or even
When parity is odd the sum of all bits including the parity bit must be odd
When parity is even the sum of all bits including the parity bit must be even
Sender adds all the bit values then appends the appropriate bit of parity
Receiver adds all the bits excluding the parity bit then calculates the parity bit.
If the calculated parity bit matches the received parity bit the data is assumed to be correct If the calculated parity bit is different that the received parity bit the data is not correct.Can only catch errors reliable if there is only a single bit in error. Somewhere around 50% effective.
Check Sum
A checksum is similar to parity in that the data sent is summed, however with checksum each byte is summed rather than each bit. After the data is send the calculated sum is appended to the data.
The receiving device sums the data to calculate a checksum value and compares the calculated value to the received value.
If the calculated checksum is the same as the received checksum the data is assumed to be correct. If the values differ the data was received in error.
Checksum is more reliable than parity but still error prone.
Checksums are on the order of 95% effective
Cyclic Redundancy Check (CRC)
CRC uses the same model as Parity and Checksum of the sender calculating a check value then sending that value, the receiver does the same calculation and compares answers.
CRC uses a more sophisticated equation for the computation. (P / G = Q + R / G)
Catches over 99% of errors.
Error Control – Correction
Errors that are detected must be corrected. There are two primary ways to correct errors.
Retransmission
When an error is detected it is simply retransmitted.
This is the most common error correction method for digital transmission.
Forward Correction
Method by where redundant information is transmitted for all messages. When errors occur the redundant information is used to reconstruct