main content
1. From the perspective of the system and overall angle, consider the development of the data warehouse. The concept and content of the CIF. CIF Case - SAP BW IV. Data warehouse and enterprise application integration five. summary
References
Summary
The main content of this paper is to introduce the enterprise information factories and explore systemic and integral properties that should be considered when establishing data warehouses and enterprise applications.
Therefore, this paper is the latest developments of the main needle data warehouse, combined with SAP BW's actual case, this paper tells the concept, concepts and content of the enterprise information factory (CIF), and also expounds the way of thinking, we will use it when designing enterprise applications. A new perspective to view the data.
1. Consider the development of data warehouses from the perspective of the system and overall.
It is increasingly complex and larger and larger, people often dismantle the world to decompose and deal with the world, which obviously enables complex problems to be handled, but invisible, we pay for the price - I don't understand a series of consequences brought about by myself. This error is often repeated when developing enterprise applications. In addition to the data warehouse, we need to re-examine and think about the development of data warehouses and its relationship with enterprise applications in data warehouses.
Data warehouse is a topic, integrated, unrecognized (stability), data collection that is constantly changing (different time), to support business management - W.H.INMON
The essence of the data warehouse is the use of data support decisions. Its core value is whether its data model portrays the core business of the company. There is a point of view that the collection of data markets is the data warehouse, and the brief record can be placed in a data warehouse or place it. Data markets supported by the data warehouse. When a brief recording into a data warehouse is an ordinary application, it is to adapt to the departmental application when entering the data market. The most obvious solution is usually no effect; it may be improved in the short term, and it will only make things more deteriorate for a long time.
The company's activities are also a "system", which is also implicated in a subtle and interest-related action. It affects each other, which often exactly the time of the year. As a small part of the group, it is difficult to set it in it and want to see overall changes. We thus tend to place the focus in a certain fragment in the system, but never wants to have some of the most fundamental problems seem to have never resolved.
II. CIF concept and content
CIF (Enterprise Information Factory) is an architecture that looks at the data warehouse system in terms of the system, we can change the information, since information has become the focus of business competition, we can use information as a company Products, and the location of the production department-level product is "info workshop", a comprehensive management of enterprise information (system) is called "information factory."
Figure 1 CIF architecture
The architecture of the CIF is shown in the figure.
Therefore, the CIF has an infrastructure that must be able to absorb and support the enterprise information of different sources. It not only needs to invest huge resources on unstructured data from file management systems, internal and external websites, operating systems, groups, and emails, but also in traditional structured data (derived from data warehouses, data sets. City, operational data storage (ODS) has investment. Creation, use, exchange, and deletion of all information in the enterprise should be implemented in this infrastructure.
Enterprise Information Factory consists of existing processing-based systems that are connected between these systems. For the normal operation of CIF, each component on the production line must share a consistent and solid infrastructure, which has some specific Requirements: Flexible adaptation, low redundancy, upgradeable, scalable. ENTERPRISE DATA WAREHOUSE - located in the heart of the corporate information factory, including the following data
Granularization
historical
Integral.
Entire
Data created in EDW provides information on data mines, data mining, etc. In short, EDW is the basis for the operation of data warehouses and decision support systems. The data in EDW is granularized and has no redundant, and it is based on the consensus on the consensus due to different analysis results of CIF. EDW can be stored in the main memory (higher performance) and secondary memory (poor performance). Break the CIF main concept elements below
Operating system, data acquisition, data warehouse, main storage management, secondary storage, data mart, data submission (is a set of view processing processes that enable end users and their information system groups to establish and manage data warehouses, which contain the following process : Filtering, formatting and submitting data from the data warehouse to the data market), ODS, metadata management, exploration database (is an architecture of a decision support system, with a data harbor for detection and temporary processing, it Using data compression techniques to achieve access to the entire database in a short period of time). Data Mining Warehouse (is an analyst to test the environment in the exploration database, professional data mining tools contain Intelligent Agents to perform these tasks. Analyze applications (pre-designed, installed decision support applications, they usually It is necessary to customize the company's specific requirements. For example, risk analysis, CRM analysis, etc. External data (External Data is located in addition to external data except the corporate application, such as demographics, credit, competitors and fiscal information Generally, external data is purchased by a business to provide a businessman to a specific information.
