An attribute or tuple timestamping in bitemporal relational databases

Much of the research on bitemporal databases has focused on the modeling of time-related data with either attribute or tuple timestamping. While the attribute-timestamping approach attaches bitemporal data to attributes, the tuple-timestamping approach splits the object’s history into several tuples. Although there have been numerous studies on bitemporal data models, there is no work contrasting these two common approaches in terms of system performance and ease of use. In this paper, we compared interval-based attributeand tuple-timestamped bitemporal data models by running sample queries to measure processing time, and then we evaluated their usability by using the same data. Our tests indicate that the attribute-timestamping model with one-level nested approach required less time and used less disk space; therefore, it is more appropriate for modeling bitemporal data.