The Enterprise Resource Integration And Query Optimization

With the progress of enterprise information-management technology, it will cause "Information Islands" that there are so many information systems with different structure in one enterprise at the same time, so to a large extent the effect of information system will be hampered in enterprise management. For the enterprise data storage is separate, inconsistent and untrue, it will be harmful to the daily enterprise operation. "Information Islands" will separate the normal enterprise operation flow; on the other hand data analysis and data mining can't be done effectively in this case. Therefore it's very important to integrate enterprise resources.Integrated system is the key to enterprise information integration; meanwhile integrated heterogeneous data source is the kernel. The core of the integrated heterogeneous data consists of data access component, data transition and encryption component, communication component, which can be reused. Three different interface-schemes will be proposed to fit the three different type of data source.A message method, as a communication method in communication component, is proposed to integrate data and reassemble them in data warehouse asynchronously and promptly, and then to access non-structure Lotus files, as an example, is given. It will not have and effect on the normal business logic operation in integration sub-system, that message transmission mechanism is used to integrate heterogeneous data system, which belongs loss-coupling integration.After analyzing several data cube storage methods ,data cube will be saved to relation data base in star-model while constructing data warehouse. Though star-model is not as normative as snowflake-model, it has some redundant information in its database, however it has higher query efficiency by reducing query join-operation. Relation database has more advantage in technology maturity and adaptive capability than multidimensional data cube.Enterprise information resources are formed to save in data warehouse, so kinds of data query and mining can be done, and then the valid information can be committed to manager in order to help them make decision. Data cube is the object of On-Line Analytical Processing. There are billions records to be scanned, linked and converge counting in data query, however, response time should be as short as possible. An approach of improving data query is given to deal with this problem by putting data cube and recent query to memory. The approach can solve the problem of I/O bottle-neck and improve the query efficiency to some extent.