CRC is an acronym for Cyclic Redundancy Check. A CRC is additional information included with some types of data to detect, and sometimes even correct, errors in that data.
CRC’s are mathematically calculated from the data being checked. The actual algorithms are often quite complex to reduce the probability that some types of errors would go undetected, or that multiple errors would cancel each other out.
As an extremely over-simplified example, using something similar to a CRC called a “checksum”, consider the following sequence of numbers:
47, 42, 101, 1995, 2, 2013
The sum of those numbers is 4200. (Hence the term “check sum”.)
When written to some media that might suffer damage or degradation, the checksum would be included:
47, 42, 101, 1995, 2, 2013 (4200)
If something happens to damage the data so the numbers are inadvertently changed, then the checksum calculated would not match the checksum that was included, and thus the data would be known to be unreliable.
A CRC differs from a checksum most notably in complexity. A simple sum, for example, would not detect if the numbers were presented out of order, as they would still add up to the same value. A CRC would detect that and more.
CRCs are most notably used on computer hard drives. Each sector of information written to the disk includes a CRC, so that if there is a problem reading the data, it can be detected and presented as a CRC error to the application or user. Checksums and CRCs are used in many more venues, however, including even credit card numbers, where the last digit is actually calculated from the preceding digits to ensure correctness.