Research Of A B/S-based OLAP System

OLAP(On Line Analytical Processing)is a decision support tool based on Data Warehouse. Mainstream database manufacturers all provided their solutions and products of Data Warehouse and OLAP. Concerning some questions and especial requirements of domestic users, this thesis designed and realized an OLAP system with B/S structure after deep research for Data Warehouse and OLAP. The main achievement of this thesis include the following aspects:1) A materialized view selection algorithm and a materialized view update algorithm are presented.2) The difference of physical schema is shielded from users through semantic objects who can manipulate their familiar semantic objects, and then obtain analytical results .3) Several key technologies which affect OLAP query efficiency are discussed, and relevant solutions are presented.4) The system is B/S-based which is three-layered, whose advantages is that it makes up for the disadvantages of the system with C/S structure. This structure is flexible and convenient for users, and is more suitable to practical requirements. Furthermore, this system is universal and extensible in which data sets that OLAP query needs can be obtained from Data Warehouse, Multidimensional database and OLTP database which can be built on different DBMSs.5) Web server clustering technology is discussed and a request dispatching algorithm is presented.