Pipelines provide the ability to perform data processing and validation operations on incoming and outgoing messages. Incoming messages use receive pipelines, and outgoing messages use send pipelines. The canonical use of pipelines is to normalize inbound data into XML and denormalize outbound data from XML into whatever the desired format may be, although this is by no means mandatory. A common misconception is that BizTalk is always required to normalize data into XML format; this is not true.
BizTalk provides four pipelines in-box that are suitable for most scenarios.
The following pipelines are available on the receive side:
XmlReceive — The XmlReceive pipeline contains the XML Disassembler pipeline component.
PassThruReceive — The PassThruReceive pipeline contains no pipeline components.
And the following pipelines are available on the send side:
XmlTransmit — The XmlTransmit pipeline contains the XML Assembler component.
PassThruTransmit — The PassThruTransmit pipeline contains no pipeline components.
As described in Tables 2-1 and 2-2, pipeline execution is split into multiple processing stages, each of which is designed to perform a specific type of processing. Note that each stage can optionally contain pipeline components, as required. Pipelines are covered in more detail in Chapter 4.
Messages flow through these pipeline stages and pipeline processing, like all processing in BizTalk, is stream based. BizTalk uses this to avoid having to load the entire ...