Towards a Modular On-Premise Approach for Data Sharing

Area of Science:

• Computer Science
• Data Privacy
• Cloud Computing

Background:

• Increasing data generation from interconnected devices raises scalability and privacy concerns.
• Public cloud solutions offer availability but lack user control over data, posing privacy risks.
• Regulations like GDPR heighten the need for robust data protection measures.

Purpose of the Study:

• To propose a modular cloud-of-clouds architecture for private data storage and analysis.
• To enable secure data sharing and computation without compromising user privacy.
• To reduce upfront costs associated with data infrastructure.

Main Methods:

• Implementation of a cloud-of-clouds system for persistent data storage.
• Integration of usability modules for secure data sharing and private computation.
• Application of MultiParty Computation (MPC) and K-anonymization for intrinsic privacy.

Main Results:

• A novel system architecture enabling private data storage and controlled access.
• Demonstration of secure data analysis and machine learning model training via private computation.
• Validation of the system's ability to maintain data privacy and user control.

Conclusions:

• The cloud-of-clouds approach effectively addresses privacy and scalability challenges in data insights.
• The system provides a secure and cost-effective solution for managing sensitive data.
• Advanced privacy-enhancing technologies are crucial for enabling trustworthy data sharing and computation.