Sift and Pass-through (or content based stub routing) – and why you need it
StuF 270005FDCU Visits (2403)
Within the context of a complex middleware development project, attempting to work on “next release” deliverables prior to gaining formal acceptance of the current release and reaching a “baselined” status can be a frustrating (and risky) exercise.
By their very nature, large development environments tend to be a little thin on the ground due to their associated licensing, hardware and maintenance costs -often projects will run only one “DEV” environment. As a customer goes through release one acceptance testing, keeping the development environment stable to facilitate efficient bug fixes will be paramount in order to reach the release one milestone and secure budget for release two (and onwards). This might result in a significant portion of the development team sitting idle until release one is signed off and work can finally start on release two enhancements.
It is at times like this that intelligent stubbing can really make a difference. IBM’s service virtualization capabilities facilitate the ability to quickly and easily create and run stubs (or virtual services) and to stand up both a stub and its counterpart live system component within the exact same environment at exactly the same time. All achieved with minimal configuration, risk or confusion and with no complex environment consistency overheads.
The behaviour is achieved with IBM’s “sift and pass-through” stubbing capability. Sift and pass-through stubs give a stub “first dibs” on a message transport or queue. If a running stub has been configured to intercept (or sift for) a given message payload type, or specific field contents within that payload, that stub will take responsibility for processing the message and providing all necessary ongoing response interactions, if not, the message is simply and reliably placed back on the transport from which it came in order for the deployed components within the system under test to process it in the normal manner (i.e. it gets passed through to the SUT). With sift and pass-through we can run the delivered code in the normal manner and intelligently mimic the behaviour of any new or differing release two components at the same time to provide new interfaces for release two developers to code and unit test against. The built in content based routing capabilities of the stubs will ensure all messages reach their appropriate end point destination seamlessly. All achieved without altering the delivered code in any way.
This is just one of the great features within IBM’s Rational Test Workbench and Rational Test Virtualization Server offerings which have been painstakingly designed to help development and test teams achieve more with less.