Byte stuffing

To transmit non-character data when using a character orientated data transmission protocol some sort of coding has to be used to avoid confusing the receiver.
This called data transparency and is achieved by using a method called byte stuffing. The character DLE is inserted into the message.

Suppose we are transmitting the following characters and binary digits:
Start and end of message and block characters have been omitted for clarity.

DATA1001110

This would be sent as ASCII as follows:

1000100	1000001	1010100	1010100	0010000	1001110	0010000
D	A	T	A	DLE	1001110	DLE

Suppose 0000100 inadvertently turned up in the binary data? The binary data transmission would be prematurely ended and the rest of the transmission would be corrupted.

0010000	1000110	0000100	0010000	1000100	0010000
DLE	1000110	0000100	!(DLE)	1000100	DLE

One way of tackling this is to stuff another DLE into the binary stream.

0000100	1000110	0000100	0010000	0010000	1000100	0000100
DLE	1000110	0000100	DLE	!(DLE)	1000100	DLE

The receiving device will read the first DLE and then the second. Once this occurs the receiving machine will ignore the first DLE and treat the second one as binary data.

Bit stuffing

A similar problem occurs when bit orientated data transmission protocols are used. Bit orientated protocols look for patterns of data for control purposes.
Suppose that a frame delimiter pattern is 01111110 when we are sending some data. Here the flag is inadvertently repeated in the middle of the frame.

0111110	00100100	11100010	10110011	11110101	00010101	00100101	01111110
			flag	occurs

To overcome this problem the transmitting device stuffs a '0' bit into the data stream after each sequence of five binary '1's.

0111110	00100100	11100010	10110011	111010101	10001010	10010010	10111111	0

The receiver does the opposite, everytime it recognises a sequence of five '1's there will be a '0' that follows. This is removed and the original data stream is reconstructed.