Oracle OLAP Services & Oracle Express Server - Oracle Statement of Direction - 2001
 


Background
For over twenty years, Express Server and its predecessors, Classic Express and Express MDB, have been the industry leading OLAP servers in terms of both analytical capabilities and performance.

Today, Express Server 6.3 provides analytical applications with a complete inventory of multidimensional query and reporting functions, and predictive analytical functions such as forecasting, modeling, allocations and consolidations, and scenario management.

Express Server 6.3 is also the undisputed OLAP performance leader. It holds the Analytical Processing Benchmark (APB) record, a measure of summary management performance, query performance, and concurrent user support.

Express Server has maintained its position as the industry leader in large part because it has evolved over time to keep pace with analytical requirements and application deployment models (first mainframe, then client server, and now Internet).

The New Platform for Analytical Applications
In Oracle9i, Express Server and Oracle database technology converge to form Oracle9i OLAP Services. Oracle9i OLAP Services expands the role of the Oracle database to become the platform for analytical applications on the Internet.

Oracle9i provides features previously found only in stand alone multidimensional databases, including support for:

  • Complex analytical queries based on a multidimensional data model.

  • Predictive analytical functions such as forecasting, models, and scenario management.

  • Rapid query response times for analytical queries.

  • Large numbers of concurrent users of analytical applications.

Because Oracle9i is part of the Oracle database, it also provides features and capabilities not available in stand alone multidimensional databases. These include:

  • Scalability. Oracle9i provides support for multi-terabyte data sets.
  • Open access, through an open Java OLAP API and SQL, allows organizations to leverage the investment in the data warehouse over the largest possible user community by supporting a wide variety of applications.
  • Oracle database user authentication, authorization, and data access methods provide an integrated security model for analytical applications and the data warehouse.
  • Integrated data warehouse and multidimensional modeling streamline to process of enabling the data warehouse for analysis.
  • High availability features of the Oracle database provide support for 24x7x365 availability.
  • Internet deployment options allow organizations to easily deploy Oracle9i analytical applications to large, geographically distributed user communities.

Oracle9i OLAP Services
Oracle9i OLAP Services fully enables the Oracle database as the platform for Internet based analytical applications. Key characteristics of Oracle9i OLAP Services include:

  • A full set of analytical query and reporting functions, and predictive analytical functions.
  • A Java OLAP API.
  • The native data store is the Oracle relational database. Significant enhancements have been made to the Oracle database to achieve Express-like OLAP performance and functionality.
  • All database administration tasks are performed using Oracle Enterprise Manager.

In addition, Oracle9i OLAP Services is the platform for Oracle Business Intelligence Beans, which provide data aware building blocks for analytical applications.

Java OLAP API
The Oracle Java OLAP API is designed from the ground up for OLAP. The Java OLAP API is object oriented and provides a simple means of expressing complex analytical queries. Since it's Java, it's made for the Internet.

The Java OLAP API provides:

  • A complete multidimensional object model.
  • The full range of mathematical, statistical, financial, and time series functions that are available in Express Server.
  • Support for asymmetric queries. For example, nested rankings.
  • Support for calculated dimension members ('custom aggregates') and calculated measures.
  • Support for incremental query construction.
  • Multidimensional cursors, which greatly simplify the task of rendering queries on tables, graphs, and other visualization tools.

Java Extensions for OLAP (JOLAP), to be part of Java 2 E. E., will provide the first industry standard OLAP API. Later releases of Oracle9i will fully support JOLAP, thereby providing a fully open OLAP platform for Internet based analytic applications.

Support for Planning Applications
OLAP Services provides not only query support, but also support for planning applications which generate data using a variety of methods such as forecasting, models, aggregations, allocations, and scenario management. Common examples include:

  • Corporate and government budgeting. In this type of an application, managers might allocate funds to lower levels within organizations, adjustments are made based on external data, forecasts, or financial models, and results are consolidated.
  • Manufacturing planning applications predict demand for a product using various forecasting and modelling methods, and data entered by a user.

In each case, users might generate multiple scenarios (for example, optimistic, pessimistic, and realistic scenarios) based on data from the warehouse or external sources, and the user's insight. The results are calculated and committed to the warehouse where they can be further analyzed using OLAP Services based or SQL based applications.

Query Performance
Analytical applications tend to require large volumes of data that are widely distributed throughout the multidimensional model. As a result, high performance cell access within the multidimensional model is the key to providing rapid query response for analytical queries. Consider, for example, ranking of products based on a market share calculation. This type of query might require data from both detail and total levels of the data model for several dimensions for many unrelated products, customers, and time periods.

Enhancements to the Oracle9i relational database, including new indexing techniques, optimizations for selecting hierarchical data, and new analytic functions provide high performance random cell access throughout the multidimensional model. This allows the Oracle database to provide query performance that is in keeping with the standard set by Oracle Express Server.

Summary Management
Summary management occurs within the Oracle relational database as materialized views. This offers several advantages:

  • The multidimensional model needs to be mapped only to the base level fact tables. Oracle's query rewrite features will automatically use pre-calculated summary data when it exists.
  • Materialized views automatically refresh summary level data when necessary.
  • Data stored in materialized views is accessible to both OLAP Services applications and SQL applications. Summary data does not need to be replicated in multidimensional cubes.

Enhancements to materialized views in Oracle9i include more generalized query rewrite, which supports a broader range of SQL statements, and grouping sets, which significantly improves manageability of materialized views.

Scalability
Oracle OLAP Services provides superior scalability characteristics because it is hosted on the Oracle database. This is a key advantage since it will allow organizations to analyze larger databases containing more detailed data over longer periods of time.

The Oracle database's proven scalability features such as partitioning and parallel query execution support, along with support for large SMP, NUMA, SMP clusters, and MPP hardware platforms allow for linear query performance scalability and linear summary management scalability.

Small databases can be efficiently supported on small hardware platforms, while multi-terabyte databases can be efficiently supported on larger, multiprocessor servers.

Business Intelligence Beans
The Oracle Business Intelligence Beans provide Java application components for use with Oracle9i OLAP Services. The Business Intelligence Beans are integrated into Oracle JDeveloper and provide a highly productive development environment for analytical applications. Business Intelligence Beans are standard Java beans, so they can also be used with other Java development environments.

Presentation beans such as Table, Cross Tab, and Graph provide the means to present data to end users. The Query Builder and Calculation Builder beans provide users with the means to build OLAP queries, and to define new data elements such as calculated dimension members and calculated measures (facts).

Query beans provide methods to query Oracle OLAP Services, while the Metadata Manager bean provides services for managing connections to Oracle OLAP Services and persisting user defined objects such as reports, graphs, saved selections, virtual dimension members, and calculated measures.

All user defined objects are saved in the Business Intelligence Beans runtime repository. Key characteristics of the Business Intelligence Beans runtime repository include:

  • The runtime repository is stored in the Oracle relational database and supports concurrent multi-user, read-write access.
  • All objects in the Business Intelligence Beans runtime repository are saved as XML. This XML definition of an analytical object can be applied to any deployment method (for example, Java or HTML) and to a variety of different devices (for example, a browser, PDA, or telephone).
  • An analytical object in the Business Intelligence Beans runtime repository can be accessed by many different applications, running on different instances of Oracle9i OLAP Services.

Metadata
Oracle OLAP Services uses Oracle9i's CWM-compliant repository for all metadata that describes the multidimensional data model and mapping of the data sources. The repository extends the Oracle data dictionary by fully describing the multidimensional model and the data warehouse. Key elements of the repository include:

  • Dimensions, including levels, attributes, hierarchies, and display names.
  • Cubes complete the definition of the star schema by describing facts, aggregation methods, and the mapping of fact tables to dimension tables.
  • Folders allow the cubes and folders to be grouped into subject areas.

Administration
Administration of Oracle OLAP Services is fully integrated into the database. There are two methods of preparing the database for analysis:

  • New OLAP tools in Oracle Enterprise Manager can be used to define dimensions, cubes, and folders. These tools create both database objects such as dimensions and materialized views, and objects such as cubes and folders in the CWM-compliant repository. Oracle Enterprise Manager will be used to prepare an existing data warehouse for analysis.
  • When Oracle Warehouse Builder is used to design and manage a data warehouse, the CWM-compliant repository is automatically maintained and the database is ready for use by Oracle9i OLAP Services applications.

Other administrative tasks are accomplished using Oracle Enterprise Manager. For example, OLAP Services tasks can be scheduled using Oracle Enterprise Manager's job scheduler.

Security
With Oracle OLAP Services, the Oracle database is the single point of all security administration.

Oracle OLAP Services authenticates users using Oracle database authentication and Oracle Internet Directory (LDAP). This eliminates the need to maintain operating system accounts for the purposes of authenticating analytical application users.

Access to elements of the multidimensional model (for example, dimensions and facts), is controlled by object privileges in the database. This allows multiple multidimensional data models to exist in the same database, while providing full control over which users have access to which models.

Access to individual data elements (for example, individual dimension members) is controlled by the database's fine grained access control features. This will be the underlying mechanism which will support the analytic applications 'scoping' feature.

Secure Sockets Layer (SSL) support provides a secure connection from the application user's browser to OLAP Services.

Mobile Applications
Oracle9i OLAP Services, the Business Intelligence Beans, and Oracle Internet Application Server (iAS) provide a solid foundation for building analytical applications that can be deployed to a variety of devices such as PDAs, mobile telephones, and pagers, as well as to PCs using Internet browsers or as Java applications. This allows users to conveniently access live data, while eliminating the need to maintain software and replicated data on the PC.

Improved Platform Support
Because Oracle OLAP Services is part of the database, it will be ported to all platforms that are supported by Oracle9i. As compared to Express Server 6.x, this significantly broadens platform availability.

Compatibility with Express Server
Although Oracle OLAP Services is part of Oracle9i, it does offer a high degree of compatibility with Express Server.

  • Express Server data files can be imported into an Oracle9i OLAP Services analytical workspace. An analytical workspace provides multidimensional structures that are compatible with Express Server database objects such as dimensions, variables, formulas, relations, and valuesets. Like Express Server data files, OLAP Services analytical workspaces can be persisted across sessions and shared by multiple user sessions.
  • All Express Server SPL code (for example formulas, programs, and aggregation maps) with the exception of some functions related to user authentication, interactions with the operating system, and XCA communications can be executed during an OLAP Services session in the context of an analytical workspace.
  • Data in an OLAP Services analytical workspace can be exposed to the Business Intelligence Beans and OLAP Services applications as a data source through OLAP Services Java OLAPI API. For example, a variable in an analytical workspace containing forecasted sales data could be exposed as a 'forecasted sales' data source through the Java OLAP API.

This compatibility fully protects your investment in Express Server data and SPL code by allowing you to use the data and code within the context of an Oracle9i OLAP Services application.

Oracle OLAP Services does not support Express Server's SNAPI, or XCA interfaces. Therefore, current releases of Financial Analyzer, Sales Analyzer, and Demand Planner will need to use Express Server 6.x. New versions of applications will be released for use with Oracle 9i.

Applications built using Oracle Express Objects Express Server 6.x must also use Express Server 6.x.

It has not yet been determined if applications built using Express Web Agent will be able to run on Oracle9i OLAP Services.

Migration to OLAP Services
Migration tools will be provided that convert Express Objects metadata (dimensions, levels, hierarchies, and measures) to the CWM-compliant repository and to move data from Express databases to an Oracle database.

All metadata in the Relational Access Manager repository is automatically upgraded to the Oracle9i's CWM-compliant metadata repository for use with Oracle9i OLAP Services.

A new Analysis and Planning Application will be developed using Oracle9i OLAP Services and the Business Intelligence Beans. This will be the successor to Oracle Financial Analyzer, Oracle Sales Analyzer and Oracle Demand Planning. This new application will support all the current functionality of the existing products with the addition of many new features and capabilities. All analytical objects and business rules created in Oracle Financial Analyzer, Oracle Sales Analyzer, and Oracle Demand Planning will be migrated to new application.

Oracle Express Objects customers can migrate to Oracle Business Intelligence Beans and JDeveloper, both of which are part of the Oracle 9i Developer Suite. This combination offers an unmatched set of application design and runtime capabilities. The strengths of Express Objects are built into the Business Intelligence Beans design experience and supported by the extensive Java development environment offered by JDeveloper.

Although there are no tools that can automatically convert Express Objects code to Java, JDeveloper and Oracle Business Intelligence Beans are highly integrated and offer a very productive development environment for analytical applications. Express Objects applications can be quickly replaced with Internet-ready applications that are easier to use and have more functionality.

A new version of the Express Spreadsheet Add-In will be released for Oracle9i OLAP Services. Express Server Spreadsheet Add-In queries will be migrated to the new Oracle Business Intelligence Beans and can be used with OLAP Services Spreadsheet Add-In. Where customers have used VBA to extend the Spreadsheet, these programs will need to be revisited to ensure that they comply with OLAP Services Java OLAP API.

Express Server 6.x
Express Server 6.x, and Express applications, are a safe investment for customers that presently require world class OLAP functionality and performance. Customers can confidently invest in Express Server and applications knowing that Express Server will be supported, maintained, and available for several years and that a clear migration path to Oracle OLAP Services is provided.

Oracle recognizes that it will take several years for the Express Server customer base to migrate to Oracle OLAP Services and will ensure that the transition is smooth and can occur on the customer's timetable. In the mean time, customers will benefit from new releases of Express Server, Web Agent, and applications.

In early 2001, there will be a new release of Express Server 6.x featuring a major upgrade to Express Web Agent. There will also be new releases of Express Server 6.x that will provide compatibility with Oracle9i. New releases of the Oracle analytical applications, based on Express Server 6.x, will be released in 2001.

As is always the case, new versions of Personal Express Server will be released at the same time as Express Server.

Source: Oracle Corporation