6 strategies implemented by data warehouse
When implementing the data warehouse, theoretical strategy guidance is required. With the development of data warehouse technology, the implementation of the data warehouse has developed from the original "auto-top" model to six models.
1, "Self-top" mode
In this "self-top" mode, the relationship between the data market and the data warehouse is unilateral, that is, data flows from the data warehouse to the data market.
2, "Self-bottom" mode
"Since the bottom up" mode is starting from the data fair to establish various departments or specific business issues, the global data warehouse is based on these data fairs.
The "bottom up" mode is characterized by less investment, and it is quick. Because it only needs less people to make decisions when establishing departmental data, solve smaller business problems. "Since the bottom up" development model can make a unit in the early stage of data warehouse development, or the proceedings of techniques can be assessed before making effective input.
3, "Parallel Development" mode
The "parallel development" mode refers to the establishment of a data market and the establishment of a global data warehouse at the same time under the guidance of a global data warehouse The "parallel development" model avoids blindness of each department in the development of their respective data marts, reducing data redundancy and inconsistency between each data. This relative independence in the "parallel development" mode is conducive to the construction of the global database.
The difference between "parallel development" and "self-top down" model is that it meets the requirements of various departments in the enterprise to establish a decision-support system for this department in a shorter time, so that they don't have to wait for the global data warehouse to establish well. Build your own data market. The experience gained during the establishment of a data market will help the optimization of data models of global data warehouses and the construction of global data warehouses.
The difference between "parallel development" and "bottom upward" model is that it uses a unified global data warehouse to guide the construction of various departments, and can solve the inconsistency between data and markets and reduce data. Data redundancy problem between the market.
But the user's demand is not still still, but as the emergence of new technologies and new applications is changing. The user's demand change is not only limited to the requirements of faster hardware, better data mining technology, more performance database software, more friendly graphical user interfaces also include changes in the information you need, and changes in the data required, this Some new properties, new tables or recommissions existing tables, attributes, will result in some new properties, new tables or reorganized tables, attributes. New demand will affect the data model of the data warehouse and data market.
"Self-top", "Self-Up", "Parallel Development" does not consider how to continuously react to the construction of the data mart and data warehouse. These three modes are only considering the needs of users in the process of building a data fair or data warehouse.
In order to solve the problem of user demand, three modes have been introduced in the industry.
4, "Feedback from the top down" mode
Feedback from the user's new demand in this mode is divided into two phases. The new demand for users in the first stage is constantly being fed back to the department's data mart, and the department data market produces its own demand variation based on the new needs of users. The second phase of the department data market feeds back its own demand and feedback to the global data warehouse, and the global data warehouse will make a corresponding change.
5, "Feedback from the bottom up" mode
In "feedback from the bottom up" mode, the change in demand after the global database is established, which will mainly reflect between data fairs and data warehouses. The data consistency between each department data is maintained and can constantly adjust themselves according to the user's feedback information, and the data warehouse established in this mode can reduce the demand change after investment. inconvenient.
6, "Parallel Development of Feedback"
In the "feeding parallel development" mode, in the initial stage of development, developers are mainly established under the guidance of a global data warehouse data model, and the issues encountered during the establishment process and Solutions and user opinions, feedback to the global data warehouse data model. While the global data model is constructed in guidance department data, it also collects feedback from developers and department users and adjusts themselves based on this information. After adjustment, the construction of the next stage global data warehouse is relatively smooth.
The ultimate goal of enterprise construction data warehouse is to allow enterprises to easily use the integrated decision support environment of the data warehouse to obtain valuable information to help manage managers make quick and accurate judgment on changing environments. And find a corresponding countermeasure. Therefore, the interface is friendly and powerful for the user's service front-end tool should be effectively integrated into this new data analysis environment. The data warehouse system is a data warehouse as the core, through data extraction tools, data conversion tools, query tools, report tools, analysis tools, data excavations, etc. to meet the needs of users.
Typical database system structure
