Error Detection and Correction (Information Technology)

In some situations, the address of the error will be known because the physical layer received an analog signal that was far from the expected value for a 1 or 0 and declared that bit to be lost. This situation is known as erasure channel.

Redundant code is not an error-correcting code.

Hardware implementation of the link layer is NICs (Network Interface Cards).

The code rated will be:
m/(m+r)
i.e. the fraction of code word that carries information that is not redundant

In a linear code, the r redundant bits (check bits) are computed as a linear function of the m data bits.

In a systematic code, the m data bits are sent directly with r redundant bits, rather than being encoded themselves before they are sent.

Convolutional codes are widely used in the deployed networks such as GSM mobile phone system.

To find out the number of bits difference, XOR can be used with two crosswords and counting the number of 1 bits in the result.
For example:
1001100
1010100

0011000 (After XORing the above two)
Here, two bits are different.

Typical LAN links provide bit error rates of 10^-10.

The check bits are at position 1,2,4,8 and so on. Hence , 0001 is a set of check bits.

Reed Solomon codes are widely used in DSL, satellite communication, DVDs, CDs.

Low Density Parity check is the error correcting codes not an error detecting codes.

CRC (Cyclic Redundancy Check) or polynomial codes are based upon taking bit strings as representations of polynomial with 0 and 1 as coefficients only.

There are two strategies for dealing with errors:

1. Error correcting codes.
2. Error detecting codes.

Data link layer is responsible for error detection and correction.