Veiklos duomenų integravimas: problemos, ypatumai ir metaduomenų svarba

Abstract

This article explores the integration process of an enterprise data from the viewpoint of metadata management. This article also presents modern problems of data integration, various important aspects of integrated information management and new integration needs. Moreover, the reasoning on the metadata used for the administration of data integration, and metadata approach (integration and architectural capabilities), which easily allows collecting, processing, organising and, most importantly, integrating activity information from different data sources and applications. Lastly, this paper presents an empirical research in the enterprise of the data integration view (i.e. in the domain of human resource information management). The current information management situation of human resources is also graphically revealed and shown in this paper. This is a good practical tool for the development and optimisation of human resource processes and data integrations.
Today, when many enterprises do not live and exist without data integration, the integration quality (of activity information process and information) could be improved by using more flexible methods and effective decisions of data integration. Therefore, integration has to change and new integration activities should be developed or the existing ones should be changed, e.g. choosing (or creating their own) specific management tools, practices or techniques of integration that help defining, identifying and understanding the integration processes and the content of integration. Otherwise – to provide the necessary knowledge on integrated business objects and data, whereas ensure trust of data users in the integrated data. This adequate form of integration management (and information creation) can be suitable to support the environment of today‘s business world (i.e. when the complexity of integration increases and there is a thriving requirement for information integration needs), integration based on (controlled by) activity (enterprise) metadata and its management.
This research showing that metadata – an essential element of data quality and safety integrity – can help constantly optimise and improve the information services and adapt them to the real (local) business needs of the enterprise. However, it is a daunting task – the metadata requirements are difficult to obtain and organise, because metadata is often very specific to different enterprise sectors, perceived differently and always need to be collected and exploited. Moreover, one must need and be able to deeply understand the semantics of the data, the scheme of the data source and to describe the schemes (a unique metadata). One also needs to work with several types of applications or tools (with modelling, integration, acquisition, storage, analysis, provision). Despite these and other difficulties in the work discussed, the necessity fact – the integration process needs to be provided with metadata information elements (e.g. with domain knowledge of activity/process, knowledge of activity expert). This is so because without metadata information, data integration cannot be designed effectively and qualitatively.