Implementing a multiple-database xquery system cd.

2. FOUNDATION TECHNOLOGIES

XML, a derivative of SGML, is the driving technology behind the goal of this research. Used because it is an open-standard, tag-driven language, XML is not hindered by licensing or other monetary factors; it is completely free to use and extend. The XML documents we are expecting to use for the scientific community are data-centric. A data-centric XML document is one that is used as a data transport, meaning that it is not important to anything that uses this data that the data is stored in an XML format. Data-centric documents have a fairly regular structure and contain little mixed content. These characteristics mean that converting documents to XML and back is fairly mechanical. XML databases are used in the project because we wish to have XQuery be the query language for documents stored in these databases. The cross query system uses XML databases and XQuery to form a distributed query against all XML databases that it has knowledge of.

2.1. XML DATABASE

An XML database is a storage entity specifically designed for the efficient storage and processing of XML documents. A benefit of these databases is that they can be modeled on top of any preexisting database format, such as relational or hierarchical. The benefit of storing XML documents in an XML database, versus a relational database, is that an XML document may be partially structured in that the document has a structure yet is missing data. When a relational database is used to model a partially structured XML document, the result is either a large number of tables or tables with large numbers of columns with null values. Both of which are wasteful and inefficient.

2.2. XQUERY

XQuery is the language used for searching and retrieving data elements in an XML document. Since XQuery is not hindered by namespaces or schemas that affect the XML document, an XQuery can return data from documents that contain complex relationships with other documents. XQuery should disambiguate element names that occur in multiple schemas for a document; furthermore, it should be robust enough to remove knowledge of the underlying structure of XML documents. What makes XQuery different from SQL query languages is that XQuery can be used over any XML document; whereas, SQL queries can only be performed on relational databases. To have an XQuery operate on a relational database, the database must be modeled as XML; this case is called querying virtual XML documents. As of this writing, XQuery is still a working draft of the World Wide Web Consortium (W3C), and thus, is not an official standard. XQuery is a read-only query language at this point; data can only be selected and ordered, not updated.

2.3. XQUERY

A cross query system is one that distributes a user's XQuery across multiple, distinct XML databases. This system must ensure that the result from the distributed XQuery is equal to the result of that XQuery if performed on the same data located on the same database. This system can be modeled after any number of network architectures, such as star and fully-interconnected, to ensure a measure of fault tolerance. What remains is query validation; if a database is down in any network model, the user will not be given back a result for that portion of their query. A greater goal of the cross query system is to design it such that it is compatible with the Open Grid Services Architecture (OGSA). When the cross query system is grid enabled, the distributed XQuery will not be sent across multiple databases but instead to multiple grids, which contain those databases. This will then allow the grid to handle processing power, temporary storage, and network bandwidth that may be needed by the databases.