With the JSR, a batch job that looks like this via the Spring Batch DSL:

<job id="data" xmlns="http://www.springframework.org/schema/batch">
    <step id="import" next="report">
        <tasklet>
            <chunk commit-interval="100"
                   reader="itemReader"
                   writer="dataWriter" />
        </tasklet>
    </step>
    <step id="report">
        <tasklet ref="reportTasklet"/>
    </step>
</job>

will look like this:

<job id="data" xmlns="http://xmlns.jcp.org/xml/ns/javaee" version="1.0">
    <step id="import" next="report">
        <chunk item-count="100">
            <reader ref="itemReader"/>
            <writer ref="dataWriter"/>
        </chunk>
    </step>
    <step id="report">
        <batchlet ref="reportBatchlet"/>
    </step>
</job>

The major features not included in this release from a configuration standpoint are the parallel execution options including splits and partitioning.

To configure a job via the JSR's DSL, a developer can either embed the configuration in a Spring context or configure it via a stand-alone XML document as show above. In the above job definition, since the batchlet reportBatchlet is not defined as a bean in the same XML file, it needs to resolve to somewhere…

batch.xml

JSR-352 does not require a dependency injection framework in order to be used. However, there are a number of DI related features that the spec depends on. The first being the building of jobs via the DSL referring to named batch artifacts. In the example above, reportBatchlet can refer to one of two things: either a Spring bean defined in a context linked to that file or a batch artifact defined in the batch.xml file located in the /META-INF directory of the application. The batch.xml file provides a very simple method of defining name to class name pairing. An example batch.xml file that addresses the above case would look something like this:

<?xml version="1.0" encoding="UTF-8"?>
<batch-artifacts xmlns="http://xmlns.jcp.org/xml/ns/javaee">
    <ref id="reportBatchlet" class="org.springframework.batch.core.jsr.step.batchlet.BatchletSupport"/>
</batch-artifacts>

While release 3.0.0.M1 supports the batch.xml requirement, it is recommended that users use Spring context configurations instead due to it's more robust dependency injection and bean factory options.

Future releases will include full support for the entire expression language defined in JSR-352.

Beyond JSR-352

While JSR-352 defines a foundation for building batch applications that is similar to what Spring Batch already provides, it is important to note that Spring Batch goes much further than what is defined within JSR-352. For example, while the JSR defines an API for building batch artifacts like `ItemReaders` and `ItemWriters`, it does not provide any implementations. Spring Batch provides a comprehensive library of implementations that can be used with virtually no customization needed. Not only does using these out of the box components make you more productive, it gives you the piece of mind that you're using components tested in the most demanding production environments.

Spring Batch also takes batch processing much further than what JSR-352 outlines. Spring Batch provides a platform for easily building scalable batch applications. Via remote partitioning and remote chunking, Spring Batch applications can be scaled across multiple JVMs in a standardized and easy manor. Adding Spring Integration into your batch solution provides capabilities like message driven job launching, polling for files, upload/download and much more.

Finally Spring Batch is pushing the boundaries of what batch jobs do and where batch jobs are run. Spring Data for Apache Hadoop provides components to run Map/Reduce, Hive, Pig, etc jobs while using Spring Batch as an orchestration tool. In the coming months, we will be releasing tools to enable the running of Spring Batch jobs on Hadoop clusters via Yarn. All of the above can be wrapped by Spring XD which blurs the lines between Spring Batch, Spring Integration, and Spring Data by providing a single system for data ingestion, real-time analytics, batch processing and data export.

Conclusion

Spring has made a large investment into JSR-352 and is dedicated to implementing in with Spring Batch. This release constitutes our continued investment in this JSR as we take our first steps towards being JSR-352 compliant. We look forward to your feedback in the forums, Jira and in person at SpringOne2GX!