Compute Grid provides a batch data stream (BDS) framework that includes pre-built code to work with popular streams like text, byte, database, datasets, and so on. You can implement an interface where the business logic for processing the stream is added. The pre-built code manages the opening, closing, externalizing/internalizing checkpoints, and so on.
A BDS framework pattern is a simple Java TM interface for a particular type of data stream that a user implements to insert business logic. The BDS framework has several supporting classes for each pattern that do most of the mundane tasks related to stream management. The following table shows the patterns that Compute Grid provides:
Pattern name | Description | Supporting classes |
---|---|---|
Used to retrieve data from a database using a JDBC connection. |
|
|
Used to write data to a database using a JDBC connection. |
|
|
Used to read byte data from a file. |
FileByteReader |
|
Used to write byte data from a file. |
FileByteWriter |
|
Used to read a text file. |
TextFileReader |
|
Used to write to a text file. |
TextFileWriter |
|
Used to read a z/OS dataset. |
|
|
Used to write to a z/OS dataset. |
|
|
Used to retrieve data from a database using OpenJPA |
JPAReader |
|
Used to write data to a database using a Java Persistence API (JPA) connection. |
JPAWriter |
BDS framework steps minimize the amount of work needed to create a new batch step by performing the bookkeeping tasks related to step management and delegating the business logic to a class implemented by the user.
Step | Description |
---|---|
A simple step that uses one input and one output stream. |
|
A simple step that uses one input, one output stream, and one error stream.. |
Step | Description |
---|---|
This policy provides a Compute Grid implementation of the ThresholdPolicy interface.. |
|
This policy provides a Compute Grid implementation of the ThresholdPolicy interface |