Browsing by Author "Larbey, Ruth Ama Mansa"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Distributed Database Architecture for Data Privacy Using Partitioning Algorithm (A Case Study of Social Security and National Insurance Trust)
    (2012-12-09) Larbey, Ruth Ama Mansa
    Over the past three decades, the trend in computing, database, networking and internet technology has enabled the collection of data as well as the access to and the utilization of this data in a way that was previously unimaginable. Each individual may exist in numerous databases in different geographical location from government agencies to purchases online to confidential medical records and financial information. Because this data is digitalized, it is stored in a way that is cheap, easy to locate, searched using sophisticated queries and easy to make perfect copies. Due to potential abuse and vulnerability of the data subject, this trend has raised data privacy concerns all over the world and motivated governments to pass laws and legislation that seek to protect the individual’s personal data. But government legislation alone is not enough and needs to be supported by technical solutions. Previous solutions have been based on data encryption causing a large overhead in query processing. Other approaches using distributed database techniques for secured database management services combine it with data encryption or some form of data encoding. This study proposes an architecture using database distribution technique without traditional encryption methods and a vertical partitioning algorithm for a near optimal partitioning of a database schema across multiple servers. The SSNIT database schema and legislation passed by governments representing five continents were used to provide a model to (1)identify privacy requirements from existing government legislation, (2) model confidentiality constraints from the privacy requirement, (3) feed the algorithm with the confidentiality constraints to come up with the vertical partitions. The challenge of singleton data confidentiality constraints and how it can be handled in the proposed architecture without any form of data encoding is discussed.