Research Of Datawarehouse Storage

As the growth competition on the market,the requirements for information are going up day by day.It is extremely important to extract required information from massive data for making decisions.However,traditional databases are unable to meet the demands.With the development of database technology,data warehouse,which is a special and decision-supported technology for data storage,has come as the time required.The ability of acquiring data is one of the standards that judges whether a data warehouse system is advanced and success.As information that stores in the data warehouse consists of data,it is important to design a data storage structure,which is effective,rigorous,and in according with the features of data warehouse,to construct a successful data warehouse system.Data warehouse is an evolution of database.The data storage structure of a data warehouse could be similar to most relational databases.So we can improve the data storage structure of an existing database,in according with the features of data warehouse,and eventually change it into a data warehouse.In Chapter One,firstly,this paper briefly introduces the definition,functions and features of a data warehouse.Secondly,this paper analyzes the data storage of PostgreSQL and MySQL as open source databases.In the third time,this paper makes comparison on database and data warehouse,from aspects of data and system design. Finally,this paper analyzes the stats of art on data storage in a data warehouse, pointing out that there are so many studies on data warehouse application,and few on the bottom implementation of a data warehouse.Chapter Two discusses about the technologies of storing massive data on the side of data warehouse physical storage.As the growing computer technologies,storage system with high pace and huge capacity is required.To meet such demand under a proper price,storage hierarchy technology is used to make the storage system cost-effective.The storage hierarchy is consists of several parts,such as cache,main storage,secondary storage and tertiary storage.Data within the data warehouse is mainly stored on the magnetic disk which belongs to secondary storage.This paper analyzes the structure and features of a magnetic disk,and puts forward four strategies to improve this disk access performance,aiming to read and write the data stored on the disk rapidly and effectively.Chapter Three to Chapter Five is the core part of this paper.Most of this part is design and implementation,assisted with theory study.Referring to the existing data storage model of several commercial databases and according to the features of data warehouse,Chapter Three proposes a data storage hierarchy model based on data warehouse,designs and implements the logical storage structure of each level on the hierarchy,including the design of data block and data record,the implementation of record operation,the design of extent and segment,etc.The topic of Chapter Four is Index.This chapter introduces general sequential indices,but studies the BTREE index,which is a special sequential,in depth,including the design of BTREE index storage structure,the creation of BTREE index,and the implementation of BTREE index operation,etc.Chapter Five is divided into two parts:buffer and file.In the part of buffer management,this paper introduces the principle and two strategies of buffer management,then designs the logical structure of buffer manager,implements the block searching algorithm and proposes a new strategy for replacing a block in the buffer.In the part of file management,this paper briefly introduces the concept of file group,and designs the storage structure of file and file group.Finally,this paper makes a conclusion to the fruit of study,and makes a prospect to the research and development of data warehouse bottom implementation.