a fact based look at hashing and sequences and collision strategies in data vault 2.0 standards.
a walk back through the joins, hashing, teradata, key selection, and arguments for and against sequence numbers. while this post is specifically related to teradata, there are some generic statements here that apply to all mpp solutions.
another walk through data vault 2.0 and hash keys.
a short entry explaining the thought process behind the switch in dv2 away from sequences over to hash keys. also describes some of the issues if natural or business keys are selected instead of hashes as primary keys in the data vault model.