This content has been marked as final. Show 2 replies
Within the JBI environment, the XML message is typically passed around as a DOM Document (wrapped as a TRAX DOMSource), so the document is parsed only once. This should be very quick, even when evaluating XPath functions to find parts of the document repeatedly. DOM and OTD aren't that dissimilar, so you should be comfortable with it.
When sending the XML message "across the wire", you are forced to serialize to XML again. This is the foundation of interoperability and loose coupling.
Sometimes converting the XML to a more convenient form (different schema) can help make it easier/quicker to run queries against.