Whatever happened to Hawk?

This was the question I was asked a month or so ago at a
conference I attended: whatever happened to the Hawk release of
DataStage, QualityStage et al? This was not an unreasonable
question—Ascential announced Hawk before it was acquired by
IBM; I was briefed on Hawk over a year ago, with the expectation
that it would have been released long since.

First, a brief word about Hawk and its salient features. The big
news is twofold: to begin with what used to be ProfileStage and
AuditStage have been combined (and completely rewritten with a new
and very cool user interface—which will ultimately be adopted
by the other members of the product family) into a single product
called Information Analyzer; and secondly there is a single,
unified metadata layer that underpins all of the products and, as
we shall see, this applies not just to DataStage, QualityStage and
Information Analyzer. Of course, each of these elements also has
lots of new features (for example, QualityStage now has a resource
estimation function the like of which I have not seen before as
well as new performance analysis capabilities, while Information
Analyzer is now much more project focused) but I don’t have the
space here to discuss these in detail.

So, nice stuff, but why has it taken as long as it has? Well,
moving from an integrated to a unified metadata environment is a
major undertaking at the best of times, but IBM decided to extend
Hawk to be the foundation of its Information Server, which is a
much broader offering than just ETL and data quality. Hence the
delay, which finally ended at IBM’s recent Information on Demand
conference, where it was announced that the Information Server (and
thus Hawk) would be available from November.

Of course, this raises the question of what Information Server
is. Put simply, it is a unified platform for delivering data within
an SOA environment. It integrates not just DataStage, QualityStage
and Information Analyzer but also Business Glossary (which does
more or less what its name suggests), Federation Server (which is
what used to be Information Integrator but without replication,
which is now provided by Replication Server) and Rational Data
Architect which, apart from reporting and administrative
capabilities, is the only non-WebSphere branded part of this stack.
Underpinning all this is the WebSphere Metadata Server, which is
also used for the WebSphere Services Registry and Repository. There
are also reporting and administrative capabilities as well as the
ability to plug other products into the environment.

IBM’s view is that Information Server, as a much broader
integration stack than just the old Ascential products, is
effectively the equivalent of an application server but on the data
side. This is plausible but it remains to be seen whether the
company can establish Information Server as a market category in
its own right, especially as there are no other companies pushing
this concept at present. That said, IBM thinks that one of the
reasons behind Oracle’s recent acquisition of Sunopsis (which I
haven’t written about here precisely because Oracle doesn’t seem to
want to answer the questions that I have asked it about this
purchase) may be precisely because it sees a market here. In
practice, both Sybase (with its recent purchase of Solonde) and
Business Objects are both further ahead than Oracle in this
respect, as is Informatica, which has had a unified metadata-based
environment for some years.