You're on point man... yeah. Some of the backup solutions can achieve 97% deduplication. You can backup enterprise servers with terabytes of data over T1 connections and shit. It achieves it by loading a data file containing metadata into the clients cache. Then it reads the clients cache to see if the data already exists on the server. If it does, it moves on the the next chunk, if not, it sends it to the server and updates the data file.