Just a week ago I went to the Rittman Mead BI Forum at Brighton UK. World’s finest Oracle BI experts were attending the forum with a lot of expert level presentations about several Oracle BI topics. First of all I have to say that I really enjoyed the setting and have to congratulate Mark Rittman and John Mead, because I think they really succeeded to organize an in-depth Oracle BI Experts forum.
A lot of attendees like Mark, John, Alex and Christian already blogged about the forum. I will try to give a impression from my point of view and how I felt during those two days. Because normally I am presenting on more management like BI seminars and I had the feeling that I had to present in the cage of the lion.

DAY 1

1. Craig was talking about predefined dimensions. This looks like a more dimension centric approach and a more business like information model. So Business Information Modeling will be the future!
2. To ETL or to ELT is a semantic discussion for me. I translated it to a architectural point of view for an enterprise warehouse. Extract, Transform and then Load, this approach is similar to a OLTP-ODS-DM architecture. To turn around the T and the L, you will get a OLTP-ODS-DWH-DM architecture. So we are going back to the, what I call the classic approach: enterprise warehouse architecture. This is great news!

Edward Roskes – Internals of Essbase
It was all about Essbase and cubes. This is a world I really have no knowledge of. What I understood is that cubes are the ultimate predefined calculations. But be aware of sparse and dense dimensions! How hard I tried, I still didn’t get it. Why would you predefine everything while the world is changing more and more. And how about changing dimensions, security like Virtual Private Databases options and how about the business information logic? So I still had a lot of questions. But I understand that if you predefine everything, use tools to click and tune and you have something that really could work fine. I thought it would be nice to compare cubes against my 3NF like approach.

Mark Rittman – BI Apps Optimization
Again a nice “click options”-like presentation by using the DAC administration tool features. What I really liked about Mark’s presentation is his database centric approach to use database features for performance like compression and materialized views. But I’m a little prejudice 🙂
A little nice thought I had: the DAC-Actions and XML, could that replace Designer tools?

John Minkjan – Cache management
I really liked John his presentation. John succeeded to talk 1 hour about caching to end with his final conclusion: put it off! Thank you John, let the database handle data and information!

Venkat Janakiraman – OBIEE and Essbase Security Integration
From now on I will call that Guy: Doctor Venkat. Some people called him the godfather of all obiee blogs. Well Dr Venkat really deserves that title! A very fast in depth knowledge presentation about security issues on several different BI servers. I already lost him after his first slide. What a speed and what a lot of technical knowledge. But then again, I was relaxed and didn’t really needed all information because we have 1 database, using Virtual Private Database options if needed. So ppfff, we don’t need all that knowledge because we ain’t got that problem.
But it made me scared. How about my presentation? Is my presentation enough on expert level ? I am not presenting tools and techniques. Scary ..

Adam Bloom – How to blow up your BI server
A very nice presentation about the path finding the source of a performance issue. What I learned: just push as much you can within the database. You have one single point of contact to search for and find the solution for your performance problems. I never thought of this argument, but it’s really an important one. Thank you Adam.

Andreas Nobbman – Scripting OBIEE – is UMDL and XML all you need
This presentation made me laugh a bit. Besides Business Intelligence I like to work in private time with my WordPress open source content management system. So after work hours I like to hack around using PHP on a MySQL database . Andreas gave me the same feeling. How wonderful you can hack everything together. But I need to separate private and business, these are two different worlds. But I was thinking: if you have a superb information model within the database, than you can easily generate the whole BI server using UMDL and XML. WOW: that would be nice!

[Dinner @ The Seafront] ^ [The night before] = TRUE

We closed day 1 with a very nice dinner. It was a hard day, a lot of techniques and things to talk about. I was sitting next to John. I told him that I was flabbergasted about the overload of tools and scripting presentations and that nobody was talking about information modeling. I should ask Mark for my security. All the experts will throw me of the roof after my presentation: users and information modeling. John advised me to express my more business/user point of view. So in the middle of the night I added some slides to my presentation to talk about the user, the information maturity model and the decision cycle. I really couldn’t sleep and was nervous. I felt like a strange duck with a weird title.

DAY 2

Focus, focus and focus. Always when I need to give a presentations it’s hard for me to concentrate on other presentations. So Mike, I am very sorry but I wasn’t really listening.

Mike Durran – OBIEE 11g
Well what I saw: OBIEE 11g looks really nice! I liked the graphics, the spatial options, but also to let users generate their own dimensions. Hmmm that could be very interesting. So 11g, please do come fast!
Emiel van Bockel – OBIEE The Rising Sun
Well here we go. I hope they don’t shoot me afterwards. So I started my presentation, half way I went to present about the user and came back to information modeling on the end. I was glad Mark asked me about performance, so I gave a live demo on our online production database: speed speed and again speed! This started a very nice discussion. Lucky for me John helped me out to translate all technical OBIEE questions. But I stand with in my opinion with good arguments and I really didn’t get good arguments why you shouldn’t start with information modeling on a more 3NF point of view. The only argument the experts could come with was performance. But good information modeling gives great performance opportunities. At the end a received a lot of compliments and I even ended on the 3rd place of best speaker. So I survived, succeeded and will continue on this matter.
Because this discussion, I promise I will write some nice modeling blogs to keep you all posted.

Maarten Jan Kampen – Oracle BI EE and Mapviewer
Maarten Jan presented us how to bring Obiee and Mapviewer together using a lot of web techniques like webservices and ajax, I really think it can be very interesting to use spatial for business intelligence purpose. But scripting is not my favorite. So maybe 11g can lead us to implement this nice new features.

Adrian Ward – OBIEE Global Implementations
A presentation how to load data into your datawarehouse, if you have a global company. Two things about this presentation:
1) I didn’t like the option to flush the cache every hour;
2) If you have a OLTP-ODS-DWH-DM architecture, than you there no loading problems because of the timezones.
So if you have a global company I suggest, merge John’s and my presentation together.

Anthony Heljula – Oracle BI & SOA Integration
This was a really “Love & Hate” presentation and therefore I just loved it J Wow that was a nice fast life presentation. Anthony showed us how to use Webservices, SOA tools, XML, iBots and mix it all together. But again a lot of click, copy&paste and pray for a good execution. Anthony show his expert knowledge in this area, great presentation.
Oh yeah, 1 thing: We as Centraal Boekhuis also have webservices up and running, but we use a generic database centric approach. The WSDL is something like
{function_call [varchar2], webKey [integer], input [type xml], output [type xml]}
The database:
1) checks the security by web authentication and the webKey
2) translates the function_call to a pls/sql procedure
3) translates the input XML to database variables
4) translate the database output to XML
So we don’t need to create all the different WSDL’s and have to think about security all the time. Just insert a new PL/SQL procedure into the database and we are up and running a new webservice.

Peter Brink – Integrating Cube Organized MV’s into the DWH
All about the new 11g webservices, more stability and easier to refresh MV’s. We experienced the same problems using MV’s. So I am very glad to hear that 11g has some nice improvements on this matter.

THE END & CONCLUSION

This were all presentations. My conclusion is:

1. There are a lot of security issues, especially if you are using more heterogeneous environments;
2. Due that heterogeneity you need a lot of administration and click tools;
3. The BI world is pushed into the XML world and that’s fun and flexible;
4. We would like to go back to the database because that’s is secure and stable.

In other words: Good luck!

My advice: stay into the database, information modeling is fun and flexible too. And you don’t have the downside of all the other XML and Tool stuff: Security, Performance issues and a Click syndrome (RSI) …

Mark and Jon: Thank you for this very interesting forum!