A good day yesterday at IOD. Here's some notes from the sessions I attended
Keynote
With the terabytes and petabytes of information swirling around us, managing those information assets is a critical competitive differentiator. The keynote session made the challenges clear: managing data growth, optimizing existing infrastructure, providing solid analytics, protecting customer privacy, integrating, presenting federated information coherently. (there's more, but you get the idea)
It wasn't that long ago customers were interested primarily in accumulating (and analyzing/reporting) information. I seem them now taking the next evolutionary step -- focusing on how to protect, optimize, and govern whilst still leveraging that data for insights and innovations. A tall challenge.
Data Quality Assessment Accelerators for SAP
I've taken this class before, and attended to see if there was anything new.
As always, Mark Oberholtzer knows how to make the SAP migration challenges clear. Get an initial data assessment early, preferably before committing to the final migration timeline. Anyone who's lived through a couple SAP projects knows the hit your project will take when the data issues surface later in the game.
Nothing new from Mark in terms of process, but the reports have improved quite a bit. And (very cool) there'll be integration with Cognos reporting for next year.
Integrated Data Management: Product Strategy and Vision
Curt Cotner described the new and upcoming features for data management.
Data Studio:
From an administration perspective, many features remind me of Oracle's grid control (but supporting more than one database engine). In fact, I'm seeing an ever stronger emphasis on muti-db support in all the toolings. Since few of us are lucky enough to work with only one production db engine, this is a big help.
pureQuery:
Ok, this is cool. I recently spent a few days at a client whose apps are built via Hibernate. We used the Optim toolset to reverse engineer some of a business object's data representation -- which made the dba all kinds of happy.
Abstraction layers like Hibernate and iBatis are great for the application developers, but a real challenge for the DBAs. One dba gave an example of a 3-table join that was rewritten by the framework into over 1500 distinct select statements. And then of course there's the challenge of trying to figure out _which_ app issued a particular sql statement against the database -- particularly when the app developers never actually see the true sql.
Enter pureQuery. pureQuery will sit between the applications and jdbc.
- It gives us back tracability to the app.
- We'll be able to do a jdbc capture to identify known sql -- and provide more performant overrides where appropriate.
- We'll be able (where appropriate) to reject 'unknown' sql statements, ensuring new sql is reviewed before implemented on critical systems. Note to security guys -- helps mitigate sql injection risks as a nice side effect.
- If we bind to static query plans, the average performance gain is ~15%.
- If we recode to work directly against the pureQuery api, performance gains average ~25%.
Optim SAP Data Privacy
This wasn't a public class. We announced "early access" of the SAP Data Privacy offer for testing a few months ago. Since we had all the Sales Engineers in the same city, the product guys gave us a training session on the architecture and future growth.
Very, very cool. When you think about the number of SAP sites whose "test" system is an unmasked clone of PROD, the need for a privacy solution becomes obvious.
Oh, and did I mentioned it's SOA-enabled? More on this once I get my hands on the software for testing.
