Blockchain based cloud service security architecture with distributed machine learning for smart device traffic record transaction

Abstract

AbstractIn today's world, the transaction through the smart device has received greater attention and configures numerous applications which can efficiently process huge traffic records on a growing demand for service centers from the edge of the networks. Due to these immense growths, the concern raises in critical transaction records in terms of system security threats and efficiency issues in the smart devices. However, existing methods failed due to security attacks during the tenure of access transactions and aggregated services. Recently, blockchain technology enables service centers depends on various platforms to share transaction records. But, it is difficult to store the transaction record because of its size. To address these issues, we proposed a SECure LearningChain (SEC‐LearningChain) design based on the integration of blockchain technology, machine learning (ML), and cloud computing primitives are applied together for a secure data transaction in a Peer to Peer network as well as efficient data sharing service. This approach consists of four design models: First, an attack detection model detects the attack using threshold‐based anomalous traffic detector in the transaction network. Second, a mold blockchain transaction network model is designed based on the cryptographic hash and encryption to deal with threats and validate the identity verification process for a secure transaction. Next, the large‐scale transaction record is optimized and trains the ML model for the output prediction. Finally, the cloud assessment model manages the stored transaction records and easily share the accessed services across different cloud platforms for each service center. Furthermore, we prove that the SEC‐LearningChain design resists transmission control protocol flooding attack, denial of service attack, and falsify attack. Experimental results demonstrate that the performance of the SEC‐LearningChain achieves more number of transactions in each blocks over existing schemes.