Array DataBase Management Systems (ADBMSs) has recently become a promising approach that addresses the challenges of efficient managing scientific databases and ADBMS facilitates data-intensive project collaboration. Many ADBMSs have developed to address the challenges derived from data-driven research. Nevertheless, there is yet no broadly accepted standard database model for array data, thus leaving many topics to be explored and optimized.

In this thesis, we present an efficient strategy called TravelGraph for managing versions of an array database. Data versioning is a crucial feature for supporting scientific research and large-scale collaborative data-intensive work. By leveraging version metadata, the strategy facilitates better scientists’ collaboration and optimizes inter-versi...[ Read more ]

Array DataBase Management Systems (ADBMSs) has recently become a promising approach that addresses the challenges of efficient managing scientific databases and ADBMS facilitates data-intensive project collaboration. Many ADBMSs have developed to address the challenges derived from data-driven research. Nevertheless, there is yet no broadly accepted standard database model for array data, thus leaving many topics to be explored and optimized.

In this thesis, we present an efficient strategy called TravelGraph for managing versions of an array database. Data versioning is a crucial feature for supporting scientific research and large-scale collaborative data-intensive work. By leveraging version metadata, the strategy facilitates better scientists’ collaboration and optimizes inter-version queries. We transform the metadata of versions into a acyclic directed graph to demonstrate their provenance information, and further select a set of versions to materialize in order to fulfill the requirement which is tunable to users. The experimental results show that our proposed strategy reduces the resource burden arising from version storage and reduces the cost of recreating old versions.