Elimination of errors during transmission is controlled by hardware and software. There are two categories of network errors when it comes to data; Corrupt data, data that has been changed during the transmission and Lost data, data that does not reach its final destination. When you design a network it should be constructed to prevent, detect and correct both the corrupt as well as the lost data errors. Network errors are inherent part of data transmissions and can occur with varying frequencies. The object of a good network is to minimize the frequency and severity of the errors so as to not affect the network and users on the network. By placing cabling away from machines and fluorescent lighting you eliminate certain impulse noise interference. By creating a strong signal you can eliminate white noise that could become a problem. If you use cabling that is shielded there would be less chance for cross talk. Increasing the guard bands on transmissions will also help eliminate the chance for crosstalk. By keeping your equipment tuned up and running at the optimal capacity you can help eliminate transmission errors that could create data loss or corruption. Having repeaters or amplifiers in places where transmissions need to travel some distance will eliminate the weakening of the signal that could result in network data errors.
Detecting errors through a parity check, Checksum, Cyclical Redundancy check will help to identify when an error in the data has occurred. Having the ability to send acknowledgments when data is received or request retransmission when issues are detected is also a considering factor when designing the network cable plan. Using error correction redundancies to prevent retransmission will eliminate some network traffic that could slow the network.
The result is to get the highest volume of accurate data transmitted over the network. This will result in the most cost effective and