Academic Data Privacy-Preserving using Centralized and Distributed Systems: A Comparative Study

Data privacy has become a critical concern in a set of domains, including healthcare, education, traffic monitoring, etc., due to technology's high deployment and massive data collection. In education, academic institutions have started taking several precautions to prevent data misuse, especially students' information, unauthorized access to the institution's databases, and any security breaches that can negatively affect the institutions' activities and objectives and students' lives. Protecting student information has become a priority, especially with the emergence of online learning, to create a safe environment, foster trust, and comply with relevant laws. Existing data privacy techniques are mostly deployed in centralized platforms, which can increase the data processing complexity and response time. However, the emergence of distributed systems helped to improve the infrastructure's security and users' privacy. Also, it reduced the processing and transmission time while providing high-quality services. This paper proposes a comparative study of deploying distributed and centralized platforms while preserving education data privacy. The distributed system is developed using k-means clustering, while data privacy is ensured by applying the k-anonymity technique using both generalization and suppression. As a result, the centralized system outperforms the distributed one in terms of beta-likeliness, t-closeness, and delta-disclosure, with less suppression. Also, centralized platforms require less execution time and higher memory allocation than distributed ones.