You are asking questions that are already were answered if you cared to RTFM/LMGTFY.

Shortly:

- Unique how? Optimally genetically, but this has practical problems that the field of paleogenetics is trying to work on. Until then must use: classical morphological features, odontology, dermatoglyphics, biochemical characteristics.

- How long enough? Depends on the type of group. There are different levels of human group classification, both above the traditionally understood "races" and a lot below that.