Using DB2 Information Integrator
Briefing Note Yesterday, I give a brief overview of some of the facilities offered by IBM's newly announced DB2 Information Integrator, writes Phil Howard. Based on federation and replication, Information Integrator is at odds with the approach advocated by Oracle, which is more slanted towards centralisation.
However, we need to be clear what we mean by centralisation. If we mean the centralisation of transactional data or of content management data, then IBM is not in this camp. It believes (and I am inclined to agree) that centralisation is unnecessary, very costly and potentially risky in such situations.
On the other hand, there are clear advantages to having a single centralised data warehouse, for example, because you need to be able to support complex, multi-dimensional queries. In such a case it makes more sense to use a conventional ETL (extract, transform and load) tool to populate the data warehouse rather than attempt to federate it with transactional sources.
Similarly, complex transformational requirements may predicate an ETL-based approach, while a requirement for up-to-the-minute, consistent data may require a conventional replication environment.
IBM gives three examples in which it suggests that DB2 Information Integrator will be most useful. The first of these is to support analytical functions, where there is a need for real-time data analysis, a requirement for cross-enterprise views or for ad hoc enquiries across infrequently accessed data.
The second is in customer-centric environments (though this could also be to extend to suppliers, human resources and so on) where there is a need for customer self service and customer care details. And the third is to support migration and co-existence, as a result of mergers, acquisitions and re-organisations.
Of course, in this last example, part of the whole point of mergers and acquisitions is often to reduce administrative costs, which may mean getting rid of some IT resources so that database consolidation is a long-term requirement. Still, DB2 Information Integrator may still be useful in an interim capacity.
Actually, the scenario I have most problem with of those outlined is in the support for real-time data analysis. With companies like Apama specialising in providing real-time analytical capability, and ETL vendors introducing real-time data feeds into their products, I don't think this situation is as clear cut as IBM might like to think.
However, in a sense that is beside the point - it is clear that that are a variety of environments in which Information Integrator has advantages over other technologies - precisely what these environments are should become more obvious once the product is widely available and we have the experience of seeing it in action.
Phil Howard's Biog is here