Design And Implementation Of Metadata Management In Hierarchy Storage Management System

The rapidly increasing amount of digital information is adding complexity to the information storage and management,thus hierarchy storage management(HSM) play more and more important role on data storage management.HSM deals with the classification and valuation of information and then stores the data in the most appropriate storage.It provides the most effective management services to the information from its creation to its disposition.High scalability,most efficient management and transparently access to the dynamically stored data has been the great challenges to a large scale data management system.This thesis then analyses the architecture and design criterions and presents the design and implementation of the Wisdata HSM in SAN environment.The WISHSM system reaches the goal of manage the information efficiently relying on its high scalability and performance,attained by researching on key technologies of file reparse,data migration,data migrating back,and optimizing database performance and implementing them in the system.The design and implementation of metadata management is then presented in the thesis in detail.WISHSM designs the application-level metadata for hierarchy storage management,according to data classification and information valuation.The file system monitor collects application-level metadata information,update and maintain these information according to file operation(such as file creation,file modification,file delete and so on).The metadata information is transmitted to metadata server through application server agent,and then is saved into database by metadata management to offer data migration.It is demonstrated from the performance evaluation that policy cache technology improved the performance of migration policy execution by an order of magnitude.Besides,Policy Metadata Container is designed to improve the poor performance in the initialization process of the policy cache.The time spent on initialization decreased by 16 to 20 times due to the performance evaluation.