Web Services - Java Data Mining: Strategy, Standard, and Practice - page 270

messages to reduce the number of service calls to be made. JDM Web
services define 11 operations. This section gives an overview of
JDMWS operations and provides an example of an SOA application
integration scenario. The section also describes the details of the JDM
WSDL and explains how JAX-RPC can be used to enable JDM Web
services.

11.3.1 Overview of JDMWS Operations

The JDM Java API discussed in Chapter 9 defines mining classes and
interfaces with many methods for designing and developing object-
oriented applications. In contrast, JDMWS is designed to develop
service-oriented applications and as such defines a small number of
operations with large messages. The JDM standard maintains the
interchangeable object model between the Java and Web Services API
by defining a consistent data model using the JDM XML Schema dis-
cussed in Chapter 10. JDMWS operations are designed to be consis-
tent with those of the Java API. This enables a JDM API interface
provider to support JDMWS more easily. It also enables developers
familiar with the Java API to understand the workings of JDMWS
more readily.
JDMWS operations are categorized into three types: mining task
execution and monitoring, mining object management, and capabil-
ity discovery. Table 11-1 lists the JDMWS operations with signatures
and a brief description of each operation. In this table, the simplified
notations [in] and [out] are used to represent the service request
message and response message. For detailed syntax of these opera-
tions, refer to [JDMWSDL 2006].
The JDMWS executeTask operation is similar to the resouce.Con-
nection.execute method in the Java API. However, unlike the Java
API where the application must save input objects before executing
the task, JDMWS allows users to specify the contents of input
objects as part of the mining task. This is an important distinction
that allows applications to perform a complete mining operation
with one JDMWS call. For example, in the Java API, before the
model build task is executed, the task input objects (physical data,
logical data, and build settings) must be saved. In JDMWS, on the
other hand, users can specify the contents of all the input objects
within the build task itself to submit a mining task for execution.