Study Of Storage Models In Native XML Databases

Native XML databases are generated with the extensive application of XML, they are specialized for storing and managing XML data. Native XML databases have features of common databases, but the internal storage is based on the tree-like structure of XML data.The XML data which are storaged in native XML databases are accessed with the technique correlated with XML. However, a native XML database is not an standalone database, and it can be bulit on existing database or proprietary storage format. Native XML databases emphasis on the logic views and what are provided to applications are tree-like XML data, while the formats of rock-bottom storage are transparent to the applicaitons.Thus, the storage models which are between XML data and different storage play an important role.The paper mainly researches the storage models of native XML databases. Different rock-bottom storage formats adopt different storage models. In the paper storage models based on relation database, object-oriented database, index files, compressed files are discussed in detail. XML data are deposited into native XML databases via these storage models and their tree-like structures are retained. On the contrary, different format data are reconstructed into original XML data via the storage models. dbXML is a procduct of native XML databases, it introduces document table storage model(DTSM). This model converts XML data into the symbol table and compressed table to store, it realizes a certain degree compression ratio and retains the structure of XML data. On the basis of analysing the DTSM, two improvement plans are proposed, one is to produce the symbol table based on document type definition and another is to compress the data of compressedTable with the adapted arithmetic compression algorithm. The improvement plans optimize DTSM in executing time and storage space.