A service-oriented microservice framework for differential privacy-based protection in industrial IoT smart...

Area of Science:

• Computer Science
• Artificial Intelligence
• Industrial Internet of Things

Background:

• Advancements in AI, IoT, and edge-cloud computing drive smart industry transformation.
• Edge and cloud computing offer scalable resources for intelligent analytics.
• Data privacy and security are critical challenges for AI adoption in distributed environments.

Purpose of the Study:

• To propose a privacy-preserving, service-oriented microservice architecture for intelligent Industrial IoT (IIoT) applications.
• To integrate Differential Privacy (DP) mechanisms into the machine learning pipeline for enhanced data protection.
• To support both centralized and distributed deployments for flexible, scalable, and secure analytics.

Main Methods:

• Developed a microservice architecture integrating Differential Privacy (DP) mechanisms.
• Implemented and evaluated differentially private Radial Basis Function Networks (RBFNs).
• Tested models using real-world and synthetic IoT datasets across various privacy budgets (ɛ).

Main Results:

• The proposed framework maintains high predictive accuracy (up to 96.72%) with acceptable privacy guarantees.
• Microservice-based deployment achieved an average latency reduction of 28.4% compared to monolithic baselines.
• Differential privacy mechanisms effectively safeguarded sensitive training data.

Conclusions:

• The proposed architecture is effective and practical for delivering privacy-preserving, efficient, and scalable intelligence for IIoT.
• Microservice design enhances computational efficiency and reduces latency through dynamic service orchestration.
• Demonstrates feasibility of deploying robust, privacy-conscious AI services in IIoT for secure industrial systems.