| Cipher encryption technology is the key problem in information security field,cryptanalysis technology is the reverse process of cipher encryption.The ultimate object of cryptanalysis is deciphering.Reversible encryption cipher can be deciphered quickly by searching its inverse transform while irreversible encryption cipher can only be deciphered by adopting rainbow table searching technology.For irreversible encryption cipher which has a large sample space,the decryption process needs a huge amount of calculation.Rainbow table construction and search time on CPU is long and the storage space is huge,which cannot meet the demands of practical system.Distributed GPU technology is a perfect method to solve these problems.Distributed GPU is suitable for irreversible encryption cipher for its supercomputing ability.Distributed GPU can construct rainbow table concurrently and search rainbow table by multithread operation.These characteristics shorten deciphering time and improve efficiency.Although the theoretical research of combing GPU and rainbow table for deciphering keeps consummating,practical application of the technology is still less.By in-depth analyses on the theories of GPU and rainbow table,it puts forward a system which can construct rainbow table quickly for irreversible encryption and optimize rainbow table multiply based on distributed GPU platform.The solution settles the following issues:(1)Calculating all rainbow table chains by concurrently computing;(2)Searching rainbow table by concurrently operated thread;(3)Realizing rapid decryption of irreversible encryption cipher of length from 1 bit to 11 bits;(4)Optimizing decryption system resources,search time and space efficiency.Chapter 1 of this thesis proposed application difficulty of deciphering irreversible encryption cipher by normal CPU system and the necessity for its application optimization;Chapter 2 explains working principle of the whole decryption system and each module,which is the application environment of distributed GPU;Chapter 3 discusses the key issues of the way to use distributed GPU in the system and how to construct rainbow table by GPU on the basis of chapter 2;Also discusses the test procedure of HASH function deciphering;Chapter 4 describes the research and test of rainbow table search by GPU to multiple bits ciphers and using multiple rainbow table chain lengths and gives a feasible comparison between the system and the existing CPU system.Analysis shows it gives a significant improvement of the decryption efficiency.Constructing and searching rainbow table by GPU can be a hundred times faster than CPU system,which can meet decryption demands in recent years;Chapter 5 of the text is a summary and explains future works of the thesis. |
PDF Full Text Request