The 8b/10b data encoding used by the PCI express bus is 20% less efficient than using some other forms of data encoding. So why is it used?

Question

What will be an ideal response?

Answer 1

Encoding digital data is not as simple as you might think. You might think that data could be transmitted directly
as data (i.e., 110101 is transmitted as 110101). However, when transmitting at very high speeds over transmission
paths that might be long, there are many practical factors to be considered related to the way in which electrical
signals propagate along a bus or other transmission path.
When signals are encoded they can be described by certain parameters such as efficiency; for example, if one
data bit is transmitted as one bit, the efficiency is 100%. If it takes two bits to transmit one data bit, the efficiency
is 50%. An important aspect of data channels is their bandwidth; that is, their ability to handle both slowly and
rapidly changing signals. In practice, this means that is it more difficult to transmit certain bit patterns. Finally,
since a clock signal has to be derived from the received data, there must not be a long period between signal
transitions in the data signal.
The 8b/10b encoding mechanism uses 10 bits to transmit 8 data bits; that is, it maps an 8?bit data symbol onto a
10?bit transmitted signal. 8b/10b has an additional advantage. It encodes data in units of 8?bits, which is largely
the standard unit of data in the digital world.
So, 8b/10b encoding is used because it is easy to encode and decode, handles 8?bit characters, has good clock
extraction properties, and good transmission properties.