Research On Directional Password Cracking Technology Based On Personal Information

Cyberspace Security is an important barrier of user information.Identity authentication is an important means to protect user information security.Among them,text password is still one of the most mainstream identity authentication forms.When a password is used by more users,it is easier to crack.Such as "123456",a common national password,its anti-cracking ability is extremely low.All hackers will try this password first.Therefore,when users set passwords,they usually mix personal information,such as birthdate,name and so on.This is not only convenient for memory,but also enhances the specificity of the password and ensures the strength of the password.However,this part of information has a strong correlation with the user itself.Under the directional password cracking environment,the user’s personal information has become the most direct entry point.The main work of this paper is to construct a directional password cracking model based on personal information,and design an optimization algorithm of password generation.The concrete work and innovation of this paper are as follows:(1)In this paper,we analyzed the collected 8 public password data sets from the data source,password length,common passwords,personal information and long password structure,in order to explore the general law of password distribution,which provides direction and basis for password guessing.In addition,this paper further studied the "12306" data set which with contains the complete user’s personal information and obtained the distribution law of personal information module in password.(2)Aiming at the problem that the traditional directional password cracking model is too rough in the classification of personal information types and has a large number of repeated statistics,this paper constructs a more refined cracking model "PCFGs-N".This paper further subdivides the personal information module into 36 sub categories based on the existing 6categories,and sets the matching priority for each sub category to avoid repeated statistics.Therefore,PCFGs-N model can more truly reflect the distribution of personal information modules in passwords.In small and medium scale directional password cracking experiments,the success rate of PCFGs-N model is 18.9%～37.1% higher than the directional password cracking model "Personal PCFGs",and about 10% of passwords can be cracked with only 10 attempts..(3)In order to solve the problem of memory overflow in traditional password generation algorithm,this paper designed a memory-optimized algorithm Pivot-V.This algorithm uses variable dynamic nodes,delays the rhythm of low probability nodes entering the priority queue,reduces the number of nodes added to the maximum heap,and reduces the memory consumption.At the same time,the traditional PCFGs model is implemented in the single thread mode.The PCFGs-N model in this paper adopts the pre terminal structure based on parallel optimization.Multi process and multi thread architecture separates priority queue calculation,password filling operation and password output operation,and makes full use of the multi-core advantage of processor.The experimental results show that the password generation speed of the proposed Pivot-V algorithm is 3.9 times faster than that of the traditional Pivot algorithm.