Byte Stuffing

Sometimes when a frame is read, the start-of-head or end-of-tail may appear in the frame data itself.

The bytes similar to start-of-head and end-of-tail appearing within the frame can confuse the receiver for the actual frame headers.

The solution for this problem is called byte-stuffing.

To distinguish between the data being sent and the control information such as frame delimiters, the data that is being sent is modified. This technique is called data-stuffing.

Types

There are 2 types of data-stuffing: 1. byte-stuffing 2. bit-stuffing

Byte Stuffing

Each one of the reserved bytes (which indicates the #frame headers) are replaced by 2 unreserved bytes, one of them acts as a prefix.

Note: soh stands for start-of-head, eot stands for end-of-tail.

Original Data	Modified Data
soh	esc x
eot	esc y
esc	esc z

Bit Stuffing

This refers to bit oriented hardware, as opposed to byte-stuffing which is byte oriented hardware.

Transmission Error

These may occur due to power surges or interference due to which the bytes or bits get corrupted.

Error Detection and Correction

frame includes additional information which is added by the sender for error detection and correction.
The incorrect data is either rejected or it is corrected first and then accepted.

Parity Checking

This is one of the error detection schemes.

Even Parity

The number of 1s in the data should be even.

Odd Parity

The number of 1s in the data should be odd.

Parity Bit

This is that extra bit that is transmitted along with the data so we can use even-parity or odd-parity.

Limitations of Parity Checking

The error detection offered by it is very limited.
Imagine the original data:
101101.
Due to noise, it gets corrupted to:
101011.

The total number of bits before or after the corruption are still even.

Alternative Error Detection Schemes

There are other techniques as well but they differ from each other in following aspects: 1. The size of additional information (transmission overhead). 2. Computational complexity of the algorithms (computational overhead). 3. The number of bit errors which can be detected.

Checksum

In this technique, the data is treated as integers and its arithmetic sum is computed.
Then on the receiving end, same thing is done and if the results match then there is no errors.

Limitations

It also has limitations.

See how the corrupted data is not detected because the checksum turns out to be same for both.

Cyclic Redundancy Check

The message data is treated as co-efficient of a polynomial.
Then this set of co-efficient is divided by a known polynomial and the remainder of this division is sent to the receiver for checking.

Hardware Components Used in CRC

There are 2 hardware components it uses. 1. Shift register 2. XOR gate