Deep learning-based user authentication with hybrid encryption for secured blockchain-aided data storage and optimal task offloading in mobile edge computing

【Author】 Ganesh, N. S. Gowri; Balasubramanian, V.; Prasad, D. Venkata Vara; Velan, S. Senthil

【Source】WIRELESS NETWORKS

【影响因子】2.701

【Abstract】Mobile Edge Computing (MEC) integrates edge technologies and provides consumers with low-latency storage and efficient computing services. MEC is an approach that delivers services from the cloud at the network edge and is rapidly evolving. Internet of Things (IoT) devices can run their applications in real-time with minimal latency via MEC. However, task offloading is subject to privacy and security breaches such as data replication, confidential information leaking, and data tampering. Edge technology can provide enhanced security for data transmission and storage, enabling tamper-proof and transparency for IoT, when combined with blockchain. However, the effectiveness of data issue detection in the earlier schemes is relatively low, making the present information storage methods unsuitable for processing the obtained data in blockchain-assisted edge processing. The utilization of resources based on network, storage, processing, and privacy can be further enhanced through the Integration of Blockchain and Edge computing (IBEC). To tackle the privacy issues regarding task offloading, a blockchain-integrated secure data storage scheme is implemented in this paper. The data has to be stored that are collected from various IoT devices. The authentication of the user whoever tries to store that data, is verified with the help of the Adaptive Multiscale Dilated Bidirectional Long Short Term Memory (AMDBi-LSTM) model. This approach leverages the strength of Bi-LSTM to permit authorized people to access the data and enhance the security of the data. Moreover, adding dilation in the developed model can easily identify the authorized user sensitive data without the need for multiple steps and the multiscale helps to increase the security of storing the data. By combining these techniques, the developed AMDBi-LSTM model aids in enhancing the speed of detecting the data issue as much as earlier than the conventional models. The hyper-parameters in the AMDBi-LSTM are optimized using the Improved Nuclear Reaction Optimization (INRO) algorithm. Optimizing the parameters in the proposed model using the INRO algorithm helps to verify the authenticity of data, enhance the offloading task in MEC, and also reduce the data breach in the existing models. Once the confidentiality of the user is verified, then the data are considered for the encryption phase. The combination of Hybrid Attribute-Based Encryption with Elliptical Curve Cryptography (HABE-ECC) is used to encrypt the data. The integration of Attribute-Based Encryption (ABE) and Elliptical Curve Cryptography (ECC) has the ability to perform secure data transmission with the help of optimal key generation. Additionally, it consumes low power and memory during the process of encryption. The keys needed to perform data encryption are generated with the help of the same INRO algorithm. The encrypted data is digitally signed by the authorized user, and it is stored in the blockchain. The optimal task offloading is also executed with the help of the INRO algorithm. The authorized user-based secure data storage and secure task offloading are also achieved with the support of blockchain technology. The security offered by this system is compared with other existing algorithms to showcase the improved security provided by this approach.

【Keywords】Secured data storage and optimal task offloading; Mobile edge computing; Adaptive multiscale dilated bidirectional long short term memory; Hybrid attribute-based encryption with the elliptical curve cryptography; Improved nuclear reaction optimization

【发表时间】2024 2024 DEC 28

【收录时间】2025-02-05

【文献类型】

【主题类别】

--