Collisions are unavoidable whenever members of a very large set (such as all possible person names, or all possible computer files) are mapped to a relatively short bit string. This is merely an instance of the pigeonhole principle.
The impact of collisions depends on the application. When hash functions and fingerprints are used to identify similar data, such as homologousDNA sequences or similar audio files, the functions are designed so as to maximize the probability of collision between distinct but similar data. Checksums, on the other hand, are designed to minimize the probability of collisions between similar inputs, without regard for collisions between very different inputs.
^ abJered Floyd (2008-07-18). "What do Hash Collisions Really Mean?". http://permabit.wordpress.com/: Permabits and Petabytes. Retrieved 2011-03-24. For the long explanation on cryptographic hashes and hash collisions, I wrote a column a bit back for SNW Online, “What you need to know about cryptographic hashes and enterprise storage”. The short version is that deduplicating systems that use cryptographic hashes use those hashes to generate shorter “fingerprints” to uniquely identify each piece of data, and determine if that data already exists in the system. The trouble is, by a mathematical rule called the “pigeonhole principle”, you can’t uniquely map any possible files or file chunk to a shorter fingerprint. Statistically, there are multiple possible files that have the same hash.