Overview of MQ data format transformations

MQ message to SCA message architecture

The following points define the MQ message to SCA message architecture:

• When either an SCA message needs to be converted to a WebSphere MQ message, or vice-versa, you must configure one body data binding and may provide any number of header data bindings.
• A data binding is represented by the name of a Java™ class. The instances of that class can convert both from an MQ message to an SCA message and vice versa. The data binding classes are supplied with the product or may be written by yourself or some third party.
• IBM® Integration Designer includes a set of header data bindings. You may add custom header data bindings to this set on an individual import or export. Your user header data bindings will take precedence over the IBM provided header data bindings.
• When reading an MQ message, header bindings are chosen based on the format identifier in control information associated with the header. When no header binding can determine the message, the body binding is called. Body data bindings have access to the remaining portion of the unparsed message and also the entire chain of parsed header objects.
• When writing an MQ message, header bindings are again chosen based on the format identifier in the control information associated with the header. The body data binding can override the format identifier written into the MQ message.
• A function selector must be used with an export which, in conjunction with method bindings, will determine which body data binding is used.

Accessing MQ messages

To read MQ messages, an MQDataInputStream object is passed. It is based on the DataInputStream class. To write MQ messages, an MQDataOutputStream object is passed. It is based on a DataOutputStream class. These classes contain methods that simplify the reading and writing of types used in MQ message structures. The MQDataInputStream and MQDataOutputStream objects also can determine MQ Coded Character Set Identifier (CCSID) and encoding values. Upon entry to the read() or write() method, the Stream is initialized with the expected CCSID and encoding values. Data bindings for more complex structures may need to update these values during read or write operations. The Streams honor the current CCSID and encoding settings when writing or reading from the MQ message.

The MQDataInputStream and MQDataOutputStream interfaces can be found in IBM Integration Designer.

MQ data bindings in IBM Integration Designer

In addition to support for the MQ Message Descriptor (MQMD), the MQ data bindings supplied with IBM Integration Designer provide complete support for the MQRFH and MQRFH2 headers, which provides interoperability with JMS, IBM Integration Bus, and the publish-subscribe distribution method. Partial support is provided for messages with a format that begins MQH; that is, messages with a format for the form MQHxxxxx. These headers are exposed as binary data in SCA and Service Message Objects (SMOs). SMOs are used in mediation flow components.

Five body data bindings may be used with an MQ import or MQ export. String and binary data bindings treat the body as unstructured binary or text data. XSD types are provided describing the DataObjects used by these bindings. These data bindings require a business object definition matching that expected of the JMSObjectMessage and JMSTextMessage data bindings, as discussed in Overview of JMS, MQ JMS and generic JMS bindings.

Note that to use these body data bindings, which are discussed in Working with the simple JMS data bindings, you need to first add a schema to your module. To add the schema, open the dependencies editor and under Predefined Resources, select Native Body schema for Native Body DataHandler. Then save your selection and close the editor. In the navigation under Data, the appropriate business objects to be used with these body data bindings are created. When you create input and output variables for your operations, select the appropriate business type. For example, select a JMSTextBody if using text. When you generate the MQ data format transformation, a filtering mechanism will choose the appropriate data format transformation. In this case, wrapped text.

Additional bindings store a serialized form of the SCA message body in the MQ message. One serialized form uses Java object serialization and one serialized form uses an XML form. When either of these body data bindings are used with an import, the import includes an MQRFH2 header in the outbound MQ message with a property TargetFunctionName in the folder. This property is set to the name of the invoked operation. An export using the JMS-type function selector will look for a user property with that name.

Lastly, the MQ Service Gateway data binding can be used in conjunction with the Service Gateway interface. This data binding determines the type of the inbound MQ message and defers processing to the corresponding built-in data binding.

When a message is parsed, the header bindings are executed first, then the function selector and then the body binding.