An IoT-Centric Data Protection Method for Preserving Security and Privacy in Cloud

A malevolent utility provider may extract outsourced data from the cloud while storing, analyzing, and sharing the data among the involved entities to acquire sensitive information that can be misused. Therefore, data protection has become a challenging task that needs to be tackled appropriately. To address this pivotal and challenging issue, this article proposes a secure data protection method for preserving privacy in the cloud environment by effectively partitioning, partially decrypting, and analyzing the data that improves the model's efficiency while maintaining security. The model ensures the system's security and privacy by performing secure data storage, analysis, and sharing. The numerous experiments are conducted and achieved results demonstrate that the proposed method procures data privacy with high accuracy, precision, recall, and F1-score up to 96.85%, 96.65%, 96.85%, and 96.72% with a relative improvement up to 45.58%, 49.31%, 45.58%, and 48.46%, respectively, for varieties of datasets as compared to state-of-the-art works.

I. Introduction

The backbone of the cyber-physical system is Internet of Things (IoT) devices that generate a massive amount of data but do not have storage and computational power at their end. Therefore, data needs to be sent from on-premise to the cloud platform for various services [1], [2]. The cloud platform has emerged as a distinguished way to provide ample storage, computation, and sharing data with diverse stakeholders for effective utilization [3], [4]. However, it is not advisable to trust a third-party-based cloud platform, especially for the storage of sensitive data, because outsourcing data to the cloud causes the devices to lose control over it [5]. According to a survey conducted by Ponemon Institute and sponsored by IBM, the global average cost of a data breach is $4.35 million in 2022, which has increased by 2.6% and 12.7% compared to 2021 and 2020, respectively [6]. Due to these reasons, data protection has become a crucial challenge, and it attracted researchers to propose methods that can retain data privacy [7]. Most of the existing models are based on encryption techniques [8], [9], [10], and -differential privacy [11], [12], [13], [14], however, these models suffer from limited efficiency, utility, and accuracy, which could be improved. To the best of the authors' knowledge, no existing models established an effective balance between the accuracy and privacy of the outsourced data.