≡ Menu

The future of Data Warehouses


An article caught my eye in the Teradata Magazine this month – Steve Brobst, CTO of Teradata, outlined 4 areas he thinks will drive data warehousing:

  1. Sensor Technology
  2. Pervasive BI
  3. In Database Analytics
  4. Non-Traditional Data Types

I don’t disagree with Steve at all – these are all critical trends for data warehousing. A couple of them, however, strike me as having broader impact. While non-traditional data and in-database analytics (further discussed in this article on the SAS / Teradata partnership) will drive the data warehouse in new directions, the other two will also change what people do with the data warehouse.
Sensor technology, such as RFID information, will dramatically increase the volume of data and the need for timely storage/retrieval for sure. But it will also make it increasingly impossible for human decision-makers to use the information. The need for information systems to decide and act – for decision management in other words – will rise rapidly. Decisions needed in real-time based on large volumes of data that is being updated constantly cannot be delivered by people – it just won’t scale. I have discussed this before and I stand by what I said – real time, automated sensor data can only be taken advantage of by real-time, automated decisions.
Pervasive BI is another topic that makes me think. The idea that everyone will use traditional BI tools strikes me as inherently flawed. Clearly the tools that work for knowledge workers and managers won’t work for call center representatives, let alone for unstaffed channels like ATMs or websites. Again, decisions will need to be automated in a way that uses insight derived from data if better and more data-driven decisions are to become truly pervasive as I have said before.
Lots of change is coming for the data warehouse but the way data is used is going to be the biggest change of all – from reporting to decision management.