Oracle rebuilds Warehouse
New OWB merits serious consideration
Unsurprisingly, Oracle has announced a whole bunch of new products at Oracle OpenWorld this week. Some of these, like the new version of 11i, will get lots of media attention. One that will not is the new release of Oracle Warehouse Builder (OWB), code-named Paris. However, in its own right it is pretty significant and it merits some serious consideration.
Perhaps the first and most obvious new feature that will be available in the Paris release (which is scheduled to be available by the end of the year) is that you will be able to send data to third party databases as targets. You will, however, have to use Oracle Transparent Gateways to do this, which presumably means that you have to go via an Oracle database. However, as OWB uses the database as its engine this shouldn't be a hardship.
The second major new facility is data profiling. Just how clever this is I do not yet know (I am arranging a detailed briefing) but my guess would be that it is on a par with, or better than, the other ETL vendors that have introduced comparable capabilities over the last year, but not as advanced as the specialised product's from companies like Trillium Software.
There is some other nice stuff in OWB. For example, it understands OLAP and has support for such things as slowly changing dimensions; the cleansing and matching capabilities provided can be integrated with data mining algorithms for enrichment purposes; there is a mapping debugger for transformations; there are extensive new metadata features and expert capabilities; and so on and so forth.
Of course, OWB is about the commoditisation of ETL processes. That is, Oracle wants OWB to be the standard for moving data into an Oracle database (not necessarily a data warehouse) and for ensuring the quality of that information. Microsoft is doing the same with its DTS product and IBM is also pursuing the same path. However, previously the capabilities offered by these vendors have been small beer when compared to the specialist products from the likes of Informatica and Ascential. With the Paris release of OWB this product is starting to look grown up and I expect the same to apply to other database-based products as they reach new releases.
So, the question arises as to how successful these commoditisation efforts will be? The answer probably depends more on the success of Oracle, say, at persuading its users to consolidate on Oracle as opposed to the heterogeneous environments that most companies have. If we make the reasonable (but not always valid) assumption that users only want one ETL tool, then OWB can only become a commodity when the environment itself is dominated by Oracle.
However, there is another major issue. Most users still rely on hand coding for data movement and transformation rather than on ETL tools but, as products such as OWB become more feature-rich and easy to use, Oracle and its counterparts should be able to pick up many new users from their various existing communities. My brief look at OWB suggests that the product has now reached that stage. All Oracle shops should have a serious look at it.
Copyright © 2004, IT-Analysis.com
Sponsored: Benefits from the lessons learned in HPC