КРИПТОГРАФИЧЕСКАЯ КОНТЕЙНЕРИЗАЦИЯ ДАННЫХ В ОБРАБОТКЕ НЕЙРОННЫХ СЕТЕЙ ГЛУБОКОГО ОБУЧЕНИЯ

Abstract

Introduction: the existing development of methods for ensuring information security in the field of information technology allows you to control the confidentiality of personal data at variable levels. Hardware solutions can be portable cryptographic keys, as well as narrowly targeted biometric identification devices. Software solutions are methods of static verification using sequences and side-tools for deploying data virtualization systems. The main problem that stands out among the methods described above is the need for additional hardware that acts as tokens, infrared cameras, and biometric identification sensors. An alternative problem is the lack of security of software solutions, since the presence of zero-day vulnerabilities can reduce security, which leads to the possibility of interception of confidential data by third parties. Purpose of the research is expressed in the analysis and development of software methods and algorithms for ensuring data security based on containerization methods. The developed software solution allows you to perform application tasks in a secure container, while the implemented authentication methods support the ability to identify an individual both using sequences and using the Windows Hello application programming interface. Methods used: security methods and algorithms aimed at implementing a unique cryptographic algorithm. The results of the studied methods and algorithms for scanning a bounded shell allow us to variatively select the optimal outcomes of internal process instances along with the implementation of cryptographic methods. Application software is based on the joint use of virtualization tools and methods of interaction through neural network technologies. The isolation of this solution from alternative methods lies in the possibility of using the container without significant hardware resources and lack of primary knowledge in configuring the software module. Practical significance lies in the possibility of using methods and algorithms in the construction of secure jobs in small and medium-sized businesses.