A Comprehensive Data On The Frequency Of Index Recording Tiered Storage Methods And Study Implementation

Along with the social informatization process accelerated, the business data of communication the financial sector constantly update and perfect, The amount of data was explosive growth, the data volume of computer storage systems has reached PB middleweight, These data are stored in the form of a record or file, With the passage of time, The use of data value is changing, Inevitable within the storage system will produce large amounts of infrequently accessed data, These data accounted for a large number of effective storage space of the storage system.With the increasing amount of data, most enterprises have the idea to expand capacity on the basis of the original storage system, but reach a certain extent cannot be re-expanded, The companies have to spend the high cost to buy a new additional set of the same level of storage system with existing storage systems, on the face a new storage system can solve the problem of storage capacity expansion to increase, but the large amount of data storage systems the main bottleneck of the speed of data retrieval is disk I/O, To do so will not only fail to improve the retrieval speed but reduce overall storage system performance, This has caused a series of problems of the enterprises in the timeliness of the data processing rate is low, and enterprise customers access slow, How to solve the scale data storage of recorded data, improve the utilization of storage resources in the storage system, Solve the bottleneck problem in the I/O and storage capacity of conventional storage, Hierarchical storage is of great significance.The paper first describes the hierarchical storage mechanism, its necessity and advantages, domestic and foreign research status of hierarchical storage and the existing problems in the practical application, Next Review the life cycle theory which is closely related to hierarchical storage theory. Analysis of the existing four basic data storage structure and characteristics, A detailed overview of the typical three-tier storage architecture, as well as the characteristics of each level storage architecture.The paper analysis the basic data storage structure and hierarchical storage architecture, Followed analysis of the first-in-first-out, the optimal replacement algorithm, the most-recently-used, second chance common data replacement algorithm idea, Contrast the strengths and weaknesses of the algorithms after analysis of the characteristics of each algorithm, as the prototype of current typical data classification algorithm, On this basis leads to the analysis of several typical data migration strategy, analysis the Low water level method based on the storage space/the CACHE replacement algorithm based on access frequency and the algorithm based on the value of the life cycle assessment, contrast the pros and cons of the three algorithms。By analyzing the database applications storage characteristics of communications, finance industry, point out a number of issues that exist in both industry, Followed analysis storage level division method of traditional hierarchical storage ideas, Finally, point out the Storage level partition plan of business system of bank and design a Hierarchical storage system aim for data in form of record.The paper proposes a storage method of data classification based on integrated frequency indicators, The algorithm considering the storage space of the storage device、 Storage performance、Data correlation factors etc, Trying to balance the various influencing factors on data access, To achieve the ideal data access effect, the algorithm is simulated and tested in windows platform, prove that compared with two typical algorithms the algorithm is more efficient in grading and balance of the data storage system capacity and access speed.