There is a question that comes up once a while … comparison between WebSphere Transformation Extender and DataStage. What is better for what? In this blog I wish to describe some of my understandings based on some experts whom I hear. In several blogs I am talking about DataStage. So first let me introduce WebSphere Transformation Extender (WTX).
WebSphere Transformation Extender (WTX)
WTX is actually a family of products all based on the common universal transformation engine. It can run standalone in its own right, and has editions tailored to ESB , BPM , and B2B products in the WebSphere portfolio. It comes in different editions each suited to a different role or environment, supports a range of platforms and has a range of optional Packs that accelerate solution delivery.
It accepts file documents or electronic messages as inputs, parses them using data dictionary like definitions, and transforms them into one or more output messages for onward routing to target applications, both internal or systems of external trading partners and suppliers. It can convert most any formatted content XML, and EDI like formats and does this without any coding or scripting by the developer. Instead the developer uses drag and drop technology and a library of pre-built functions to create processing rules similar to how you would use macros in your favorite spreadsheet application.
WebSphere Transformation Extender vs DataStage
- DataStage is “typically” going to make more sense when there are high end parallel requirements for transforming “massive” quantities of data.
- WTX is “typically” going to make more sense when there are complex transactional requirements (creative units of work)
- WTX is “typically” going to make more sense when there are complex hierarchical formats at the source and at the target (say EDI). This is an increasingly grey area also since the xml Stage was enhanced several releases ago, but still is often a deciding factor.
- WTX “typically” appeals to more technical types and DataStage to “data professionals who are less technical”. You are more likely to find “semi-technical” teams in a data warehousing scenario who prefer a classic ETL tool paradigm like DataStage or its other competitors.
- There are more real-time protocols available “out of the box” with WTX.
- DataStage and its metadata are more geared for data warehousing, decision support and governance scenarios — when integration with Cognos, Business Objects and other such tools becomes important, and when the audience (users supporting the application) are immersed in those technologies.
- WTX is going to make the absolute most sense when you are dealing with an “out of the box” subscribed metadata scenario (Type Trees provided by IBM) — SWIFT, EDI, Edifact, etc. …especially the legacy binary ones.
A site needs both kinds of tools — for the audience building and supporting the application, as well as the functionality itself. Where necessary, there are also hybrid solutions (DataStage invoking TX or vice versa), most of which are using the older MapStage or just doing things directly with JavaPack.
DataStage 8.7+ can use TX with an accelerator that runs on top of the Java Integration Stage. More information on this can be found here.