Generally cited in fiction as a good but easily breakable security method (see the movie Sneakers).

Technically, it is the series of harmonics within the sound of a person's voice, or rather the ratio and speed of change within the harmonics that identify a voice as belonging to someone.

In society, it is used to identify people, much like sight (or sometimes smell) is used to identify features specific to that person. Most people sound different from one another, though it isn't uncommon to come across astounding similarities (at least in timbre and pitch). Much more rare is finding two people who share not only a simliar tone of voice, but also a similar method of expression: phrases, grammatical errors, accent, etc.

In my experience, I have better luck matching a voice to name than a face to one. Since I use my sense of hearing a lot more carefully than my sight, this isn't surprising. It would be interesting to find a corrolation between auditorially-oriented people and how they relate to other people.