The Research And Implementation Of Optimized Character Space Searching Algorithm

As the development of password protection technology leading to larger and largercomputations, the character space searching becomes more and more difficult. In order tosolve the problem, it is necessary to optimize the searching algorithm for reducing thenumber of searches needed and inproving the searching efficiency. The traditional searchingalgorithms rarely apply the rules of passwords to the searching and also aren’t suitable forparallel searching which make the algorithms inefficient. That’s why we should focus on thecharacters’ distribution rules of the passwords and apply them to the searching algorithm tooptimize it, for purpose of reducing the number of searches needed and improving theabilities of parallel searching.This paper put emphasis on the research of brute searching strategy. Firstly we introducesome regular searching strategies. Then we propose an optimized character space searchingstrategy, which applies the finer-grained distribution information ofcharacters’ frequency tothe searching. The strategy combines several characters as a whole to calculate the frequencyin order to inprove the average effeciency of the searching.Secondly we inplement three strategies on the parallel computing cluster system basedon CPU+GPU heterogeneous architecture to test the effeciency of these strategies. Afteranalyzing the testing result, we find some distribution rules of characters’ frequency, andverify that the effeciency of searching strategy based on combined characters’ frequency hasbeen obviously improved.Finally, in order to enhance the effeciency of the system, we propose a two-stagedecomposer architecture and implement it on the the parallel computing cluster system.