Uma abordagem para gerenciamento de consistÃncia em um ambiente de banco de dados heterogÃneos

AUTOR(ES)
DATA DE PUBLICAÇÃO

1999

RESUMO

A common problem within most corporations nowadays is the distribution of data along departmental and functional lines, contributing to the emergence of the so-called âislands of informationâ. The data are organized and managed by a mix of different DBMSs from different software vendors and different operating systems that use different network protocols. Usually, the total corporate data resource is composed of multi-vendor database servers, legacy systems, and relational and non-relational data sources. Unfortunately, these data sources have no ability to compare and relate data amongst themselves. In this thesis, we present an approach to solve the semantic heterogeneity among heterogeneous data sources. The task of resolving semantic heterogeneity aims at determining precisely the relationships between objects that are used to model similar information in different components. Moreover, another major goal is the detection of possible conflicts in structural representation of objects, which pose problems during the integration of such components. At last, the correct treatment and handling of these objects should be considered as well. In contrast to other proposals in the field of heterogeneous databasesâ integration, where a single global schema or partial global schemas (federated systems) are usually employed, we exploit the concept of navigation through the integrated databases. This method treats data conflicts only when a statement relates data from distinct databases at once. This means, we do not treat the global schemaâs construction beforehand, but on the contrary, we detect and cope with the conflicts during database navigation on the fly. We present in this work the problems we were faced with during the design and implementation of our integrated databasesâ environment, and furthermore, we discuss our solutions to these problems. Among these, we can emphasize: the use of intermediary tables to solve problems of schemasâ integration of the type âone-to-many attributesâ, i.e., an attribute in a database corresponds to many others in another database. In addition, we use an advanced methodology for proximity search in order to retrieve for the user similar information from the databases, in case the desired data value cannot be found for any reason. At last, by means of a special mechanism for consistency checking, we allow each user to individually define the level of consistency that must be enforced during navigation through the integrated databases. On the basis of this technology, we provide the users with a kind of expert system behavior when the mapping between correspondent data values cannot be made automatically and without inconsistencies

ASSUNTO(S)

consistency management ciencia da computacao banco de dados heterogÃneos databases gerenciamento de consistÃncia heterogeneous databases banco de dados

ACESSO AO ARTIGO

Documentos Relacionados