|
|
BEA AquaLogic Data Services Platform (DSP) brings a service-oriented architecture (SOA) approach to data access. Data Services Platform enables organizations to consolidate, integrate, and transform as needed disparate data sources scattered throughout their enterprise, making enterprise data available as an easy-to-access, reusable commodity: a data service.
For client application developers, DSP provides a uniform, consolidated interface for accessing and updating the heterogeneous back-end data sources that comprise data services. This chapter provides an overview of Data Services Platform for client application developers. It includes the following topics:
| Note: | Data Services Platform was initially named Liquid Data. Some artifacts of the original name remain in the product, installation path, and components. |
The Data Services Platform (DSP) significantly simplifies how client applications access and use data. In a typical organization, data comes from a variety of sources, including distributed databases, files, applications from partners or e-commerce exchange markets. With DSP, client applications can use heterogeneous data through a unified service layer without having to contend with the complexity of working with distributed data sources using various connection mechanisms and data formats.
DSP provides a uniform, consolidated interface for accessing and updating heterogeneous back-end data. It enables a services-oriented approach to information access using data services.
From the perspective of a client application, a data service typically represents a distinct business entity, such as a customer or order. Behind the scenes, the data service may aggregate the data that comprises a single view of the data, for example, from multiple sources and transform it in a number of ways. A data service may be related to other data services, and it is easy to follow these relationships in DSP. Data services insulate the client application from the details of the composition of each business entity. The client application only has to know the public interface of the data service.
This document describes how to create DSP-aware client applications. It explains the various client access mechanisms that DSP supports and its main client-side data programming model, including Service Data Objects (SDO). It also describes how to create update-capable data services using the DSP update framework.
This guide provides information about how to leverage data services in your applications. For information about creating data services in WebLogic Workshop, see the Data Services Developer's Guide.
In the typical organization, data flows in a bidirectional manner from a wide variety of sources, including distributed databases, various files, applications from partners, or e-commerce exchange markets.
Creating an application that can access and update distributed, disparate data sources can be complex, challenging, and expensive: you must know how to use a wide variety of connection mechanisms and data formats, and how to use the variety of APIs required to access and update each back-end data source, for example.
Using DSP, data architects create data services that:
A DSP client is any process that consumes data services. A client application may be, for example, a Java program, non-Java programs such as .NET applications, BEA WebLogic Workshop applications, or JDBC/ODBC clients.
For client application developers needing to leverage such data assets, DSP supports multiple access methods (see Figure 1-1):
Whatever the client type, DSP gives application developers a uniform, services-oriented mechanism for accessing and modifying heterogeneous data from external sources. Developers can focus on the business logic of the application rather than details of various data source connections and formats.
Data services are created in WebLogic Workshop, BEA's integrated development environment for building and deploying many types of applications: portals, Web services, and integration applications, for example. The Data Services Platform application running under Workshop supports all aspects of Data Services Platform creation.
From a high-level perspective, a data service defines a distinct business entity such as a report that describes a customer and customer orders. The data service defines a unified view by aggregating data from any number of sources — relational database management systems (RDBMS), Web services, enterprise applications, flat files, and XML files, for example. Data services can also transform data from the original sources as needed.
In order to use data services, you need know only a few details, such as:
Data service client applications can use data services in the same way that a Web service's client application invokes the operations of a Web service. Rather than invoking operations from a Java client application, the data service client application invokes a data service routine.
| Note: | For complete information on creating data services see the DSP Data Services Developer's Guide. |
A DSP client application is any application that invokes data service routines. Client applications can include Java programs, non-Java programs such as Microsoft ADO.NET applications, BEA WebLogic Workshop applications, or JDBC/ODBC clients:
Regardless of the client type, DSP provides a uniform, services-oriented mechanism for accessing and modifying distributed, heterogeneous data. Developers can focus on business logic, rather than on the details of various data source connections and formats.
In your client application code, simply invoke the data service routine: the DSP engine:
These data objects conform to the Service Data Object (SDO) specification, a Java-based architecture and API for data programming that is the result of joint effort by BEA and IBM.
Data Services Platform administrators can control access to deployed DSP resources through role-based security policies. DSP leverages and extends the security features of the underlying WebLogic platform. Roles can be set up in the WebLogic Administration Console. (Refer to the DSP Administration Guide for information about the DSP Console.)
Access policies for DSP resources can be defined at any level — on all data services in a deployment, individual data services, individual data service functions, or even on individual elements returned by the functions of a data service.
For complete information on WebLogic security, see:
http://download.oracle.com/docs/cd/E13222_01/wls/docs81/security/index.html
Application developers can choose from several models for accessing DSP services. The model chosen will depend on the access mechanism you decide to used. The possible access methods are:
Each access method has its own advantages and use. Table 1-3 provides a description of each of these access methods and summarizes the advantages of the various models for accessing DSP services.
A Java interface for using data services. Returns data as data objects, providing full support for Service Data Objects (SDO) programming.
For more information, see DSP's Data Programming Model and Update Framework.
|
|
|
|
For more information, see Accessing Data Services from WebLogic Workshop Applications.
|
|
|
A data service can be wrapped as a Web service, providing the data service with the benefits of web service features.
For more information, see Enabling DSP Applications for Web Service Clients.
|
||
Client applications can use JDBC or ODBC to access DSP services using SQL queries. The DSP JDBC driver supports SQL-92.
For more information, see Using the Data Services Platform JDBC Driver.
|
Service Data Objects (SDO), a specification proposed jointly by BEA and IBM, is a Java-based architecture and API for data programming. SDO unifies data programming against heterogeneous data sources. It simplifies data access, giving data consumers a consistent, uniform approach to using data whether it comes from a database, web service, application, or any other system.
SDO uses the concept of disconnected data graphs. Under this architecture, a client gets a copy of externally persisted data in a data graph, which is a structure for holding data objects. The client operates on the data remotely; that is, disconnected from the data source. If data changes need to be saved to the data source, a connection to the source is re-acquired. Keeping connections active for the minimum time possible maximizes scalability and performance of applications.
To SDO clients, the data has a uniform appearance no matter where it came from or what its source format is. Enabling this unified view of data in the SDO model is the Data Service Mediator. The mediator is the intermediary between data clients and back-end systems. It allows clients to access data services and invoke their functions to acquire data or submit data changes. DSP serves as such a SDO mediator.
On the client side, information takes the form of data objects. Data objects are the basic unit of information prescribed by the SDO architecture. SDO has both static (typed) and dynamic (untyped) interfaces for working with data objects.
Static interfaces provide a programmer-friendly model for getting and setting properties in a data object. Accessors are generated for each property in the data type of a data service, for example getCustomerName( ) and setCustomerName( ) for a Customer data object. Static interfaces depend on a schema for type information.
The dynamic interface, on the other hand, is useful when a type is unknown or undefined at runtime. Dynamic interface calls are in such forms as:
get("CustomerName")
|
set("CustomerName", "J. Dough").
|
In keeping with the goals of a service-oriented architecture (SOA), data graphs are self-describing. The metadata API enables applications, tools, and frameworks to inspect information on the data contained in a data graph. The data is described by an XML schema, which describes the names of properties, their types, and more.
For details on using SDO, see DSP's Data Programming Model and Update Framework.
The SDO specification does not specify an update framework, but it does discuss the need for mediator services, in general, to handle updates to data objects; the SDO specification leaves the details up to implementors.
The SDO specification allows for many types of mediators, each intended for a particular type of query language or back-end system. DSP provides a Data Service Mediator, a server-side component of DSP's XQuery processing engine that serves as the intermediary between data services and client applications or processes.
The Data Service Mediator facilitates updates to the various data sources that comprise any data service. DSP's Mediator service is the core mechanism for the DSP update framework. The Mediator handles updates to relational and non-relational data sources as follows:
Developing client applications that invoke Data Services Platform functions and procedures presumes that a DSP project has been deployed or is being deployed. See the Data Services Developer's Guide for detailed information about developing data services.
Developing a DSP-enabled client applications encompasses these steps:
For example, if the data service is hosted as a Web service, you can develop a Web service client application using Java in conjunction with the service's WSDL file.
Similarly, if the data service is incorporated in a portal, business process, or Web application, your client application development process may take place entirely in the context of the server, as a set of pageflows or other server-side artifacts.
Figure 1-4 provides a conceptual overview of the various approaches, highlighting some of the relationships among various sub-systems and components.
Client application developers typically work with a small set of APIs, contained in JAR files. The APIs are primarily described through Javadocs (see DSP Mediator API Javadoc on page 1-13).
Data Services Platform APIs are contained the packages listed in Table 1-5.
XMLBean classes and interfaces on which the Data Services Platform SDO implementation relies. Also enables XPath expressions in untyped data accessors. 1
|
1A "query too complex" exception is raised if the |
The Data Services Platform Mediator API describes the routines needed by DSP client applications to invoke various DSP routines.
Client application developers will find Javadoc helpful for creating client applications that invoke data service routines. Data services developers and architects will find the Javadoc useful for understanding how to customize update behavior.
You can find javadoc for the Data Services Platform 2.1 Mediator API at:
Client applications built on earlier versions of Data Services Platform can continue to use the 2.0.1 mediator API routines. These are described in a Javadoc named javadoc-dsp201 and is available at:
Javadoc is also provided in ZIP file format; it is available for download from the DSP e-docs Web site:
Data service performance is the result of the end-to-end components that make up the entire system, including:
Before creating a client application for a data service, it is recommended that you benchmark performance of each underlying data source, and then benchmark the performance of the data service as you develop it. Use load-testing tools to determine the maximum number of clients that your data service can support.
In addition, you can use DSP's auditing capabilities to instrument your code, thereby gaining performance profile information that you can use to identify and resolve performance problems if they occur. For detailed information on DSP audit capabilities see DSP Administration Guide.
A compendium of technical and product information related to BEA AquaLogic Data Services Platform can be found in the introductory chapter of the DSP Concepts Guide.
|
