380 likes | 756 Views
Evolution of Data Warehouse Conceptual Modeling. A Visual Tour Dr. Karen C. Davis April 2008. Overview. industry perspective research perspective architectures conceptual models research topics modeling ETL automated creation of schemas schema evolution and versioning.
E N D
Evolution of Data WarehouseConceptual Modeling A Visual Tour Dr. Karen C. Davis April 2008
Overview • industry perspective • research perspective • architectures • conceptual models • research topics • modeling ETL • automated creation of schemas • schema evolution and versioning
In the beginning … [O03]
Star Schema [K03]
Snowflake Schema [Z08]
Conceptual Models based on: • ER: ME/R, StarER • graphs: DFM • UML: YAM2, UML Profile
ME/R [SBHD98]
StarER [TBC99]
Dimensional Fact Model (DFM) [GMR98]
YAM2: upper level [ASS06]
YAM2: intermediate level [ASS06]
YAM2: lower level [ASS06]
A UML profile approach [LST06]
Additional Semantics • explicit hierarchies • symmetric treatment of dimensions and measures • multiple hierarchies • correct aggregation • non-strict • m:n fact and dimension • changes to hierarchy • uncertainty • data at different granularities • unbalanced hierarchies • irregular hierarchies • annotation of dimension values • multidimensional constraints • security • missing data estimation • sequence classification • progressive query answering • modeling metadata • schema evolution [PJ99] [HLSB02]
Extended Hierarchy Semantics [Banerjee and Davis 2007]
Modeling and Optimizing ETL [VSS02]
Using Design Patterns [JS05] • extending DDPs • more expressive target model [Deshpande and Davis 2008]
Using Ontologies [RA07]
ME/R to Implementation Cognos Powerplay [QAD Business Intelligence] [HSB00]
DFM to SQL [GR01]
Conceptual to Logical to PhysicalUsing a Model Gen Approach extend [ACB05] [Nicholson, Vaidyanathan, and Davis 2008]
Cross Schema Querying... Cross Version Querying…Schema Merging [GMR98, GLRV06]
Future Directions • lack of a standard conceptual model • modeling security • mining-aware design • semantic gap between conceptual and logical models • modeling ETL • design process • interoperability with metadata • emerging applications [RALT06]