Application Mode of data integration in ODS Projects

The operational data store (ODS), also called operational data storage, is a data set used to support daily global applications of enterprises. It is a data storage technology between DB and DW. According to the definition given by experts, it is a topic-oriented, integrated, current, and "volatile" data set, which reflects the instant of slicing at a certain time, A collection of business analysis systems and peripheral systems used to exchange data with each other, it is mainly used for Consistency Verification between the business analysis system and the peripheral system, and feedback from the business analysis system on decision-making support data of other peripheral systems, feedback data includes detailed information based on the customer's extended attributes. From the perspective of the role and implementation of ODS, ODS integrates the operation data of isolated business systems to provide a unified data view for enterprises and data sharing for ODS.

What is data integration? It refers to the synchronization between heterogeneous data. Heterogeneous Data refers to the synchronization between databases, files, and mails of different types and versions. To synchronize heterogeneous data, you must be able to accurately obtain the metadata structure of the data source and the ing between heterogeneous data, including syntax and semantics ing. Data integration can be considered as ETL (Bi is defined as extraction, conversion, cleaning, filtering, and loading), but more emphasis is placed on automated process management.

The implementation mechanism should include common integration modes: Split, merge, route, and PS. In addition, it should have the functions of modeling (metadata management) and governance (Governace.

What is the significance of data integration? It mainly implements data centers, such as enterprise Sid, or data integration of legacy systems. Based on the establishment of a Global Sid, you can build many meaningful things, such as portal/CMS, reports, data search, and mining. These can be broadly referred to as Bi, or business intelligence. Therefore, data integration has two levels of significance: 1) Operation Level 2) analysis level.

Currently, data integration products include Oracle/BEA's aqulogic DSP, Vitria's BW, and open-source eMule/ServiceMix/spring integration. However, as far as I know, open-source things have not yet supported RDBMS adapter, and some important functions are missing, such as monitoring and modeling.

Generally, the implementation of data integration can be divided into four steps:

1. Define the item and content of data interaction, such as BOM data between PDM system and ERP system;

2. Specify the Data Interaction cycle, once a day or once a week;

3. Select the interaction mode and use the database or middleware technology for interaction;

4. Interactive scheduling of ODS dataProgramData is uploaded or downloaded by the peripheral system to achieve data integration.