A distributed computing model for big data anonymization in the networks.

Area of Science:

• Computer Science
• Data Science
• Information Security

Background:

• Big data's rapid growth presents significant challenges for IT infrastructure and computing capacity.
• Publishing big data for analysis risks individual privacy, necessitating robust anonymization techniques.
• Apache Spark offers a scalable, in-memory computing framework ideal for large-scale data processing.

Purpose of the Study:

• To propose an efficient parallel computing model for privacy-preserving big data anonymization.
• To leverage Apache Spark's resilient distributed dataset (RDD) programming for enhanced data anonymization.
• To address runtime, scalability, and performance issues in large-scale data anonymization.

Main Methods:

• Developed a three-phase in-memory computation model for big data anonymization using Apache Spark.
• Implemented partition-based data clustering algorithms to support the λ-diversity privacy model.
• Utilized RDD transformations and actions, incorporating City block and Pearson distance functions.

Main Results:

• Achieved efficient parallel implementation of a novel big data anonymization computing model.
• Demonstrated the model's effectiveness in preserving the λ-diversity privacy model.
• Provided a comprehensive guideline for applying Apache Spark in privacy-preserving big data research.

Conclusions:

• Apache Spark is a suitable framework for implementing scalable and performant privacy-preserving big data anonymization.
• The proposed three-phase in-memory model effectively handles the complexities of large-scale data anonymization.
• The Spark-based implementation offers a valuable tool for researchers in the field of data privacy.