Fingerprints of Electronic Files: Performance Tests
Information stored in digital documents can be lost during transmission, migration, or when the storage medium breaks down or is corrupted. To ensure that the data has not changed, you can perform a digital fingerprint procedure. Up to a small margin of error, a digital fingerprint is unique to each document and therefore verifies the integrity (unaltered state) of the document.
This Demonstration presents some results of using modified cellular automata to generate fingerprints of electronic documents based on their contents. It uses the binary data of the document as the initial condition. The result of the computation is a fingerprint of the document, which can have 128, 256, or 512 bits. This Demonstration evaluates the performance of different compression techniques on different file sizes and of generating different hash sizes compared with common hash algorithms. The timing axis on the graphic is plotted with a logarithmic scale.
Each test measures the time to process each compression technique, which is compared with common hash functions (MD5, SHA), with different hash sizes (128, 256 and 512 bits).
At every step, a cellular automaton rule is applied to the current state, and cells are dropped according to the compression technique being used.
This Demonstration presents some results of precision tests on research about finding an NKS way to generate fingerprints of electronic documents based on their contents, using the binary data of a document with an initial state and the result of computation as a fingerprint of this document.
More information about Document Management Systems: Wikipedia.
More information about Fingerprint Algorithms: Wikipedia.