Migrating a Multi-Terabyte Astronomical Archive from Object to Relational Database Management System

Title: Migrating a Multi-Terabyte Astronomical Archive from Object to Relational Database Management System

Research Question: How can a multi-terabyte astronomical archive be effectively migrated from an object to a relational database management system?

Methodology: The researchers used the Sloan Digital Sky Survey Science Archive as a case study. They initially used an object-oriented database management system (OODB) due to its data organization capabilities, platform independence, query performance, and conceptual fit to the data. However, they found that the OODB had limitations in query support and data mining performance. After trying to improve the OODB with little success, they decided to migrate their data to a relational database.

Results: The researchers described the technical issues they faced with the OODB, including its inability to support their demands for features and bug fixes, and its failure to keep up with rapid hardware performance improvements. They explained that they had to abandon the OODB and migrate their data to a relational database. They outlined the steps they took to migrate the data, including creating a new schema, mapping objects to tables, and writing SQL queries to load the data into the relational database.

Implications: The researchers concluded that while an OODB might be a good choice for some data-intensive scientific applications, it may not be the best choice for very large archives like the Sloan Digital Sky Survey Science Archive. They suggested that migrating from an OODB to a relational database could be a viable solution for similar projects facing similar challenges. They also emphasized the importance of choosing the right database management system based on the specific needs of the project.

Link to Article: https://arxiv.org/abs/0403020v1 Authors: arXiv ID: 0403020v1