Day II of the RittmanMead BI Forum 2011 started of with a presentation of Emiel van Bockel. He took us on a journey from table to strategy. Without knowing themselves the presentations of Michael Wilcke and Emiel their presentations where complementary. We are moving from informatization to socialization. The importance of the context is increasing. You should really know your customer and you must get your information straight. Earning money with information. If information has added value, people (or at least Emiel) are willing to pay for this. The integration of Oracle BI 11g and mobile increases the accessibility. Centraal Boekhuis is the first company, which migrated from 10g to 11g. New functionality like Actions, Maps, Visual Capabilities, Scorecarding but also the role based User Management and Hierarchies will give more possibilities to bring the information to the end user.
Next up was Stuart Wallace who wanted to help us fixing the Fixed Price. Most BI Projects (Processes) are subject to the Devils Triangle:
The biggest challenge is the scope. How can you manage a project and make sure you deliver; Fixed Time, Fixed Scope and Fixed Price.
Deliver in small chunks (agile iterations), keep the ROI of BI in mind (look further than the initial project scope) but make sustainable solutions.
Agile iterations bring us to next part of the day; Should all BI Projects be Agile? The discussion started of with a statement of an advocate and an adversary (Waterfall) of Agile. For me it was a very interesting discussing. I haven’t got any Agile experience yet, but I think it is a good method. Users ‘only’ know what the want, when they see their first dashboard. So you have to deliver value as soon as possible. Of course documentation is important, but do it during development instead of upfront. A physical diagram within OBI could also be sufficient as documentation. Sometimes you are designing a unknown solution for an unknown need. Agile should never be an excuse to deliver crap!! The most important thing is that you should think before you do something. On to top of that you should keep communicating. This way you could achieve better understanding for one another.
After lunch Mike Durran shed is light on Scripting the Management of an Oracle BI 11g System. Although Andreas Nobbmann already had covered this subject, Mike did pointed on some ‘new’ things. He first started out explaining the Oracle BI 11g Architecture. That made it a little bit more easy to understand when and why you should use certain MBeans to script these by using Phyton. For more details please have a look at the Oracle BI Systems Management API.
Tony Heljula had the privilege to close the Forum for the majority of the attendees, including me. Tony gave an thoroughly worked out overview of how to measure performance issues in Oracle BI EE and the Oracle database. If you consider the hardware a fact and fine, there few things on the software side you can check and fix;
- Gather Stats (%)
- Star Transformation
- Bitmap Indexes
- Removing Snowflakes (into the dimension, directly to the fact)
- Bitmap Join Indexes
- Parallel Query
Unfortunately I had to go, and missed out on the presentation of Stewart Bryson. Better luck next time.