At the Think Big Data Warehouse conference, one of the other Keynotes that kicked us off was the address by Dr. Stephen Brobst, CTO of Teradata. Having to follow Claudia Imhoff is quite the challenge, but there presentation styles were so different and the topics/approach diverse from each other.
In discussing the trends in Business Intelligence (BI) Stephen noted some great high level observations which I will share below and add a little personal perspective to each:
Business Trends and BI
- More Decisions
- More Complexity
- More Data
- More Real-time
Technology Trends (I think these were selected from Forrester or Gartner)
- Actionable or Operational BI – real-time Enterprise
- Infrastructure Standardization
- Master Data Management
- Offshore Sourcing
- End-user Education
In the discussion he then focused on what the real-time enterprise is in the BI world. This is often called many different things and the problem with the definitions is that it connotated different things to different people.
So in the example of Real-time Data Warehousing, I think of automated systems like in a nuclear power plant or on a shop floor which are capturing emense amounts of data real time as closed loop systems leveraging the captured data as part of the feedback system. This is not generally what people are meaning, so Stephen is looking to change the definition to Right-Time Data Warehouse instead of Real-time Data Warehousing. I prefer the word active data warehousing. But both are good.
So what does this mean in this context? No business surprises as the information is there. This also means that there is a shift from batch. To define this change, you need to be able to do updates while the system is operating as opposed to batch overnight. So you need to have the data updated within the daily operating window (more than one time per day). So once you have made the shift you can then decide how often you want to load the data (turn the crank). Once you move to updates in less than 60 minute cycles you then need to change your mindset even further to the messaging architecture as well as the transformation/data integration layer.