Design And Implementation Of An Optimization Framework For Querying Autonomous Data Cubes Based On Ontology

In recent years the capability to innovate and collaborate among differententerprises is more and more considered a crucial skill to react more dynamically tomarket changes and reduce risks. However, one of the main issues to overcome,especially in the management of temporary networks of enterprises (i.e., VirtualEnterprises, or VE), is the integration of heterogeneous data, and the need toevaluate common Key Performance Indicators (KPI) that are capable to measureperformances of the whole VE.BIVEE project has provided a platform for user to execute Multi DimensionalQuery for specific indicator. However, in such distributed scenario, in particular,besides heterogeneity due to the use of different terminologies, procedures andprocesses, the performance of such query cannot satisfy the need especially forincomplete Data Warehouse.This thesis introduces a new type of data structure–Cubette which only storesthe schema of aggregated data. For sparse and heterogeneous Data Warehouse,Cubette serves as a cache for information retrieval so that it is able to improve thequery performance remarkably. Its integration and connection with Ontology, WebService and existent module requires more investigation and development on relatedsubjects.Moreover, this thesis also describes the main functionalities of anOntology-based data explorer for KPI, aimed to support users in the extraction ofKPI values from a repository. Data produced by partners of a Virtual Enteprise aresemantically annotated through a domain ontology in which KPIs are describedtogether with their mathematical formulas. Based on this model and on reasoningcapabilities, the tool provides functionalities for dynamic aggregation of data andcomputation of KPI values through the formula. In this way, besides the usualdrill-down, a novel mode of data exploration is enabled, based on the expansion of aKPI into its components.