The Center of the UniverseXML, Web services, analytics, and other hot technologies have the leading relational DBMS providers working overtime to remain the best choice for managing all of your data. Here's a look at what IBM, Microsoft, and Oracle are doing
by Ken North Continued from Page 2 Integrated Business AnalyticsXML use is increasing in the BI and data mining communities, and competition among IBM, Microsoft, and Oracle remains intense. Object-relational technology enabled IBM and Oracle to embed business analytics in the database. XML plays a role for metadata services and exchanging models. The Predictive Model Markup Language (PMML) is an XML vocabulary for defining statistical models and sharing them between PMML-compliant applications. PMML enjoys support from IBM, Microsoft, Oracle, SAS, and SPSS. The Intelligent Miner plug-in for IBM DB2, for example, uses PMML. Grid ComputingThe GGF is a global community of practitioners and researchers that includes many early adopters of grid computing. GGF working groups and research groups have produced important grid computing specifications and best practices documents. Early adopters included the scientific community. Scientists are no strangers to massive data sets and computationally intense problems, and they found grids boosted tools such as the Basic Local Alignment Search Tool (BLAST) for comparing gene and protein sequences in public databases. Interest in grids increased when HP, IBM, Oracle, and Sun hopped on the bandwagon. IBM and Oracle are proponents of grid services defined by the Open Grid Service Architecture (OGSA) initiative. HP, IBM, Sun, and others are promoting on-demand computing based on a grid infrastructure. OGSA is a collection of specifications that provides for marrying Web services and grid services. Grid services, like Web services, build on XML technologies such as XML Schema and WSDL. The Binary XML Description Language (BinX) provides a vocabulary for describing the structure of arbitrary binary data files. Grid researchers are using BinX to describe large scientific data sets. Microsoft, IBM, and Oracle use the same grid services infrastructure technologies (TCP/IP, SQL, XML, XQuery, SOAP, and WSDL), but diverge in their platforms and higher-level technologies. IBM and grids. IBM released the Grid Application Framework for Java and the IBM Grid Toolbox, which includes the Globus Toolkit 2.2, custom installation scripts, and load balancing software (IBM LoadLeveler). The company made significant contributions to OGSA and the GGF specification for grid data access and offers a variety of grid solutions targeted at vertical markets such as the financial and healthcare industries. One of IBM's showcase grids is the Smallpox Research Project Grid built on DB2 UDB, Linux, and AIX. According to IBM, the Smallpox Research Project will perform 15 million DB2 queries per day. Oracle and grids. Oracle released a Grid Developer Kit based on the Globus Toolkit and supports bio-informatics grids with a modified version of BLAST and Linux server blade farms. One of Oracle's important grid undertakings is a bio-informatics grid that includes programs that use XML to export data from BLAST files and store them natively in Oracle Real Application Clusters. Oracle is grid-enabling its Collaboration Suite and Enterprise Manager and supports distributed XML processing. Enterprise Manager will provide coarse- and fine-grained grid monitoring with information about application clusters, advanced queues, streams, application servers, and databases. Oracle's showcase grid is a project done by Electronic Arts to support an online game. The Sims uses Oracle9i Real Application Clusters to support 35,000 simultaneous users and 30,000 SQL transactions per second. Microsoft and grids. Microsoft's showcase grid includes a federated database using Microsoft SQL Server and .Net technology. The SkyQuery grid includes three distributed astronomy data sources. Microsoft Research assisted with SkyQuery and reviewed the Data Access and Integration Services (DAIS) specification from the GGF. Jim Gray, manager of Microsoft's Bay Area Research Center in San Francisco, reports the GGF DAIS working group will include support for .Net-compatible row sets in future specifications. Microsoft paid Globus to develop a Windows XP version of its toolkit; however, the next version of the Globus toolkit isn't targeted for a Windows platform. Microsoft also funded the development of G2, a grid framework that uses the .Net Common Language Runtime and classes. In February 2003, Microsoft released a software developer kit for developing peer-to-peer applications for Windows XP. It includes a peer-to-peer API and support for multipoint communication, persistent peer groups, peer-to-peer name resolution, and distributed data management. The Data Access FutureA data access roadmap for the next decade will continue to have separate paths based on choice of operating system, programming language, and DBMS. Which APIs or objects will become the technology of choice for data access? For working with products such as DB2 and Oracle, many organizations use Java, JDBC, SQLJ, and tools such as Oracle JDeveloper or Borland JBuilder. For accessing Linux and Unix databases, C and C++ developers will use native APIs or ODBC. Microsoft is moving its developer community to OLE DB and ADO.Net, but it also offers a JDBC driver for SQL Server and an ODBC.Net provider. For writing database plug-ins, you can use Java, JDBC, or SQLJ with DB2 and Oracle, or C or C++ with DB2, Oracle, and SQL Server. Using a database plug-in, such as a DB2 extender, means taking a turn off the path of standard APIs. Java XQuery API, for example, is an execute-and-fetch API for Oracle users querying documents with XQuery. DataDirect Technologies and other middleware companies typically precede database vendors in releasing drivers or providers as new data access specifications become available. Looking to the future, it's important to provide connectivity solutions for Linux and .Net developers while continuing to support the Java community. Microsoft offers SQL Server for Windows and .Net platforms, but not for Linux. IBM and Oracle support Linux and .Net with servers and data access middleware. Oracle released a .Net data provider (ODP.Net) in December 2002. IBM released a beta .Net provider for DB2 in February 2003. The SQL standards committee has disbanded the SQL/CLI project group. Microsoft won't be adding new functionality to ODBC or the MFC ODBC classes and has dropped Remote Data Object from Visual Basic. Loss of support for ODBC is a problem for developers because it currently offers connectivity to more types of data than OLE DB, JBDC, or .Net. Grid data access. As grid technology emerged, researchers turned their attention to distributed data and grid data access. The OGSA project included middleware for integrating data from multiple data sources in a grid. The OGSA project worked in conjunction with the GGF DAIS Working Group (DAIS-WG). The DAIS-WG produced both a requirements document and a grid database services specification. IBM and Oracle have been involved in the DAIS-WG and participated in the development of both documents. It should be no surprise that the initial draft of the DAIS database services specification includes references to Java technologies. It suggested the grid transactions model will evolve from some combination of OMG OTS and Java Specification Request (JSR) 156 clients, and query results will be JSR 114-compliant WebRowSets. More recently the DAIS-WG agreed to investigate the WS-Transaction specification. According to Microsoft's Gray, the DAIS-WG has also decided to investigate a subset of ADO.Net for returning query answers. In January 2003, the OGSA Data Access and Integration (OGSA-DAI) project released a reference implementation of middleware for grid data access that provides access to data stored in DB2 UDB, MySQL, Oracle, and Xindice. Every Kind of DataSolving the data management problems of the 21st century requires the ability to exploit structured and unstructured data for Web services, BI, and other purposes. IBM, Microsoft, and Oracle will continue to compete with database products that support a variety of data types, including XML documents. Each company will try to offer the most complete platform for grid computing, component development, Web services, and application integration. Ken North [ken_north@compuserve.com] founder of Ken North Computing LLC, is an author and consultant. He teaches Expert Series seminars and is the XML and Web Services Editor for Dr. Dobb's Journal. RESOURCESGrid Data Access and Integration: www.ogsadai.org.uk IBM: www.ibm.com Microsoft: www.microsoft.com Oracle: www.oracle.com SkyQuery: www.skyquery.net
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| |||||||||||||||||||||||||||||||
