Temporal and Flexible Data Warehouses

Temporal and Flexible Data Warehouses

연구 분야: Databases

학회: International Conference on Data Management Technologies and Applications

The present trend among companies involves a significant revamp of their data architecture, aiming to streamline data processes and phase out outdated systems. The advent of big data has profoundly influenced businesses, empowering them to adeptly manage and analyze vast datasets. In the realm of business intelligence, particularly in decision-making, data warehouses play a crucial role, leveraging OLAP technology for the efficient analysis of structured data. Constructed by amalgamating data from diverse sources, a data warehouse faces the challenge of accommodating big data-comprising unstructured, semi-structured, or structured data from myriad sources-where alterations in content and structure are frequent. To address this, our paper introduces a temporal multidimensional model utilizing a graph formalism for multi-version data warehouses, adept at assimilating changes from data sources. This approach relies on multi-version evolution for schema modifications and employs bi-temporal labeling for entities and their relationships to capture data evolution. Our proposal enhances data warehouse evolution flexibility, broadening analysis possibilities within the decision support system, and enabling adaptable temporal queries to yield consistent results. Building upon our prior work [6], where we presented the GAMM model emphasizing evolutionary data treatment, including dimensional changes, this study expands on the temporal labeling principle in our approach. It delves into various functions governing the evolution of temporal data, offering illustrative examples. We validate our approach through a case study demonstrating temporal queries and conduct runtime performance tests on graph data warehouses.