Multifunctional Password Recovery Platform Based On DCR System

Password recovery has always been the focus of computer network security and cryptography.With the development of Internet and encryption technology,user passwords are everywhere.Password is mainly used for identity authentication and data encryption which require a password to ensure the security of system or data,such as personal online banking accounts,enterprise management systems,national confidential information,etc.Password has been applied to individuals,businesses,national institutions and other fields.The development of password recovery and encryption technology are mutually restrained but complementary.With the research of password recovery technology,the commonly used methods for password recovery are brute force attack,probabilistic context-free grammar and mangled wordlist attack.With the rapid development of hardware,such as the emergence of the MIC coprocessor and GPGPU,as well as the promotion of high-performance cluster computing,password recovery time is greatly reduced.The specific implementation of existing password recovery methods are different from each other physically and logically,such as the single-functionality,different processing logic and flow,low computing power of single computing node,not only unable to make full use of the advantages of different password recovery methods but also unable to use computer cluster to shorten the password recovery time and improve the success rate of password recovery.In this paper,a multifunctional password recovery platform which supports cluster computing and three kinds of password recovery methods,password collision,complex expression and oclHashcat password recovery,will be implemented by studying existing password revcovery methods,usage of open source password recovery tool oclHashcat and using DCR(Decompose Computation Reduce)parallel framework.The platform based on DCR,which makes full use of not only the respective advantages of the existing password recovery methods but also computing capabilities of large scale computer cluster,improve the password recovery efficiency,success rate and provide appropriate password recovery method to be used in different scene through using the function of self-defining of computing task,task decomposition and reduction of computing results provided by DCR.At the end of this paper,we use the community forum software system WordPress as an example to verify the correctness of the three kinds of password recovery methods,testing their single machine performance,multimachine acceleration ratio,system efficiency,stability and reliability.