The hawk swoops!

Or, more precisely, it has started to swoop. As this may be opaque I had better explain: Hawk is the codename for the next release (version 8.0) of what used to be the Ascential product suite and is now the IBM WebSphere Data Integration Suite. However, most of the product suite has yet to reach the Hawk stage, which will probably be in the autumn, the one exception being WebSphere DataStage TX, which has recently been announced.

DataStage TX is what used to be the Mercator product before that was acquired by Ascential. Although it was historically regarded as being in the EAI (enterprise application integration) sphere, Mercator always differed substantially from other companies and products in this area such as Tibco, SeeBeyond (now part of Sun) and WebMethods because the product was always independent of the messaging layer. In practice it sat, not necessarily uncomfortably, on the fence between data and application integration, which is one of the reasons why it was a good fit for Ascential.

Briefly, DataStage TX is a transformation engine that will map any sort of data (structured, semi-structured and mixed data) from one place to another or, more accurately, from multiple places to multiple places (because transformations are many-to-many) through a completely codeless environment. It is widely used in conjunction with SWIFT, EDI and other such documents.

The new release is huge in the sense that there are many new features and enhancements, many of which I will not have space to touch upon here, so I will just concentrate on some of the capabilities that caught my attention.

For many users perhaps the most significant enhancement will not be any of the fancy things that I will come onto in a moment, but the enhanced large file support. In previous versions of the product, Windows NT was a supported platform and this provided a lowest common denominator for the file sizes that could be supported. Support for Windows NT has now been discontinued (which is a bad thing) but the result is that the product can now support file sizes of up to 4Gb (which is better). The only exception is for CICS where there is a limit of 1Gb (this is because of CICS not DataStage TX).

Not far behind file sizes in significance is dynamic logging. When an error occurs in a financial transaction, it can cost you a lot of money. Previously, if this happened, you had to restart your application with logging turned on, which was a time consuming process, or you had to run logging permanently, which was expensive in performance terms. Now, you can turn on logging dynamically, find out what the problem is, fix it, and then turn off logging again.

Probably the sexiest new feature is the Visual Debugger, which is introduced with this release. This does pretty much what it says on the tin. As I have already mentioned, DataStage TX is a codeless environment but IBM has made the debugger look and behave like a conventional debugger for coding, so that there are breakpoints, stepping and so on. Almost equally appealing is the new Map Performance Profiler, which breaks down iterations and time on all mapping constructs so that you can see how individual parts of an application are performing.

Finally, I should mention that there are updated Trading Packs, which support the latest standards and that there is also an associated release of DataStage TX Trading Manager (previously Commerce Manager) which supports z/OS, which will be a relief to historic users of the product who were locked into an earlier version because Ascential/Mercator had stopped supporting the product on mainframes.

As I have said, there is a lot in this new version. If the rest of the Hawk release is as feature-rich, and if the same is also true of the forthcoming new release of WebSphere Information Integrator, then this autumn should prove to be very interesting.