You can use any of the standard hashing algorithms, but hashing can't technically guarantee uniqueness. Hashing is designed to be a relatively fast and/or small token to be able to see if one piece of data likely is the same as the other. It's fully possible for entirely different sets of data to produce the same hash, though being able to produce these algorithmically is very hard.
All of that aside, for checking likely identity, MD5 is fairly fast. SHA is more reliable (MD5 has been hacked, so shouldn't be use for security), but it's also slower.
