This paper analyzes the importance of data removal. According to the existing data of the removal techniques and algorithms, it improves MD5 code fingerprint algorithm to calculate and optimize it, analyzes and recombines the fingerprint calculation of water level. Using the cache group to replace the single cache, we propose a new method to calculate the two water level fingerprints based on multi CPU to study and analyze the method. At the same time, it supports effectiveness of the method through the relevant tests and test data.