MATHIS: a New Approach for Creating Views to Materialize in a Hybrid Integration System

Abstract

The vulgarization of information and telecommunications technologies has made the integration systems an immediate necessity. Among the latter, there are the hybrid integration systems. These offer a local database where it stores a part of data while integrating the other part virtually. This in the objective to increases system performance while ensuring a tradeoff between query response time and data freshness. The approach that has proposed a complete solution is one that creates the candidate views for materialization before selecting among them those that will be materialized. To do this, it selects the attributes most requested by users. Calling the algorithm k-schema, these attributes are organized in schemas, on which we based to build the candidate views for materialization. This approach suffers from a defect. It is because in the phase of the extraction of attributes of interest, we based only on their frequencies of appearance in the user queries. This will select some attributes, even if they do not respond to any selection criterion. Thus, they cause the elimination, in the selection phase, of views to which they were assigned. In this paper, we propose a new approach, which eliminates all attributes that do not respond to the selection criteria prior to the creation of views to materialize. We also propose a new solution to calculate the values of attributes relative to different criteria.
Copyright © 2013 Praise Worthy Prize - All rights reserved.