Design And Implementation Of GriDoc Document Visualization Manager And Client Data Access Driver

Data is an important resource for human society,and the use and management of big data play a role in fueling for the development of industries.Science and modern solutions base on big data.The main characteristics of big data are heterogeneous,distributed,massive and low value density.Driven by computer applications,people accumulate more massive and richer data.How to efficiently manage these discrete data to extract into domain-oriented integrated information-knowledge,becomes crucial.However,the current technologies focus only on the physical organization and the simple query-oriented logical organization of big data.And this organization is generally used to query,difficult to describe the specific application logic,and not suitable for knowledge-oriented management of big data,so the use is very restricted.In terms of above problem,this paper firstly researched GriDoc model proposed by professor QI Deyu and related implementation technologies,perfected GriDoc support system which consists of three major components: paragraph engine,GriDoc engine and GriDoc driver.Paragraph engine is a data access and release agent;GriDoc engine is responsible for processing user applications’ access and operation to GriDoc data;GriDoc driver provides software API to support applications’ implementation.Secondly,this paper unified GriDoc support system and application into GriDoc big data management system including paragraph engine,GriDoc engine,GriDoc driver and GriDoc manager.Next,this paper designed and implemented GriDoc client data access driver.The driver shields lots of underlying diversity and complexity of data management,and provides application developers with unified programming interface.The driver which is similar to the traditional database drivers is implemented using EITP protocol and referring JDBC design patterns.Then,this paper designed and implemented GriDoc document visualization manager above GriDoc support system.The document manager can connect GriDoc engine server and visually operate data in it while providing composite document data display,full-text search,data share and other business functions.And it is implemented using java multi-thread technology and buffer technology.Finally,this paper deployed and tested the big data management system,summarized the shortage of the system and carried on the forecast to the future work.