Authors:
Noura Azaiez
and
Jalel Akaichi
Affiliation:
ISG-University of Tunis, Tunisia
Keyword(s):
Data Warehouse, Evolution Operations, Data Mart, Propagation Rules, Schema Versioning.
Related
Ontology
Subjects/Areas/Topics:
Decision Support Systems
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
Abstract:
Data Warehouse, accompanied with Online analytical processing, is considered as the core of the modern
Decision support systems. The emergence of new analytical requirements and changes in organization business
processes push the underlying information sources, destined to feed the data warehouse, to modify not
only their data, but also their structure. This, obviously, has a direct impact on Data Warehouse and its associated
Data Marts. Maintaining Data Warehouse structure becomes, therefore, a must; however, it is not sufficient.
In fact, evolutions performed on the Data Warehouse schema have to be propagated on the related
Data Marts in order to minimize costs, time-consuming and to guarantee the coherence of provided analysis
results; this presents our first vision issue for which, we aim to provide an adequate solution. Another issue,
which is as important as the precedent one, focuses on modeling a continuous temporal evolution phenomenon
and therefore reducing inconsi
stent Online analytical processing queries results. Indeed, data returned
by queries can be the result of an evolution phenomenon continued in several time intervals. Therefore, we
nominate the versioning approach as a solution to keep traces of Data Warehouse / Data Mart schemas’
modifications. Solving these two issues presents the key of organization Decision support systems durability
and its material prosperity.
(More)