News from the world of data modeling
Written By: David Norfolk
Published: 17th March, 2014
Content Copyright © 2014 Bloor. All Rights Reserved.
I was bought up in the era of corporate database and data analysis/modelling, as a DBA in government healthcare and in merchant banking. In my book, understanding the structure and semantics of your data is one of the pillars of good governance. Then, in the 1990s, people forgot about data—and code—process was everything. Unfortunately, process changes all the time as the business environment changes, so governance gets harder, as we get more and more code, doing more and more clever things, with the business struggling to keep track of exactly what its automated systems are doing or, perhaps more of an issue, what they allow clever and inventive employees to do. Agile systems are good, but an organisation does need to be sure that its agile systems deliver business benefit efficiently—and that it is in compliance with regulations and that resources aren't being wasted on non-strategic or non-profitable (or, worse, fraudulent) business activities.
The answer, of course, is clever, near-real-time analytics. In the emerging world of big data everything is instrumented and the mass of information available to the organisation (both relational transaction data and non-relational logs and event-processing information and more) can be analysed in near-real-time and used to support decision making and good governance, with management-level exception reporting and dashboards. Well, there are still some technical issues there but a moment's thought will tell you that without a good understanding of data structures and semantics (used to support data quality and data governance processes), this is all built on sand. The good news is that data structures and semantics are relatively stable (the principles of double entry bookkeeping, for example, and the fundamental structure of accounting data, hasn't changed much in 500 years) so they can provide a rock around which to build good governance. Even in the mad world of derivatives trading, the fundamental accounting data really meant something, as the recent crash and consequent recession underlines; and data protection and privacy regulations (see here and here) are concentrating peoples' minds on the importance of data.
Perhaps this is why Wayne Williams, CEO at Embarcadero Technologies, can say: “Data architecture is now the biggest part of Embarcadero’s business and represents our significant commitment to the category. Every aspect of IT—from application development and business intelligence to security and user experience—is defined and driven by data architecture". Amen to that. That last sentence was true back when I was in DBA in the 20th century; I'm only a little concerned that I can report it as 'news' in the 21st century!
What is the real news is that CA ERwin, one of the most used modeling tools amongst the data professionals still practicing (and that's a growing group again) is being acquired by Embarcadero, one of the innovators in the data management space (see my blog here, for example). I have been following CA ERwin and, of late, its messaging has been moving firmly towards making modeling relevant for, and accessible to, business stakeholders in automated business systems development—a very good thing in my view. Data management generally, and data modeling in particular, should be a collaborative process involving both IT technicians and business analysts (back in the 20th century, it was often one of the few things really helping to bridge the gap between IT and the business).
I hope to talk to Embarcadero soon, about exactly what this acquisition means for its customers and for existing ERwin users, and I'll write something more about this acquisition after that. For now, Embarcadero has acquired a lot of ERwin customers, at least, if it can keep them (there are other data modelling tools available, such as Sybase PowerDesigner). Its ER/Studio is a strong product that already provides good modelling capabilities, although ERwin has been developing lately and is probabaly easier to use than ER/Studio. Embarcadero is promising to "drive greater investment, innovation and momentum in this critical technology segment”, so I'm cautiously optimistic. Embarcadero tells a strong story in enterprise development generally.