IOUG Live! 2003 Day #4 (Wednesday)
by Stephen Andert
Today was the last day of the Exhibit Hall, with the opportunity for free t-shirts and stuff. Since I would need a new dresser if I bring home any more t-shirts, the only time I spent in the Exhibit Hall was to talk to companies whose product I had an interest in. In some cases that meant I got to meet my Account Rep in person.
Big Bash or not, educational sessions started up again this morning and I could see any large reduction is head count from earlier in the week. I was able to squeeze in a few sessions today in spite of the time I took to rehearse my presentation one last time and test the setup in the Speaker Ready Room.
One of the presentations that I attended was titled "More Examples of Interpreting Wait Events To Boost System Performance". Roger and Terry were careful to avoid the debate with the Hit Ratio crowd, but made it clear that the proper use of the Wait Interface (WI) was usually, in their experience, the best approach to tuning. In this 90 minute session, they quickly covered an "Intro To WI" and then showed 5 real-life examples. In these examples, they discussed what the user reported the problem as, what the WI reported and what the problem was and how they solved it. This was a great series of case studies that anyone facing a performance problem should read to understand the methodology they followed.
Anjo Kolk presented a session on the Wait Interface. Like some of the other sessions I've attended, the rating of advanced certainly applied here. Anjo started out with statements that should be obvious, but frequently are not. We need to remember that the end user doesn't care about the hit ratio or what waits the database is experiencing. They want to get an order confirmation before the customer hangs up and calls the competition. He then dug deeper into the X$ tables than many dba's understand. He also demonstrated with some examples of how the biggest tuning problems are not technical, but organizational and political and need "special" solutions.
After Anjo's presentation, I went around the corner to deliver my presentation on "SQL*Loader Tuning and External Tables". How did it go you ask? I felt good about the presentation and like a pilot friend of mine likes to say, "Any landing you walk away from is a good landing". Seriously, I got several positive verbal comments from attendees and had more people in the room that I was expecting. There was also a lot of good questions from the audience, and one question in particular has given me an idea for doing some additional research in order to do an article for a technical publication. I'll know more is 3-6 weeks which is when the Conference Committee said to expect the results from the comment cards.
Gary Goodman gave a great presentation that I would highly recommend for any optimization project. He gave 5 rules for managers of optimization projects. Two of the best, IMHO, are that project goals need to be aligned with the business process that it is supporting and that you need to track the return on investment and document gains and improve your ability to forecast improvements and costs.
Tomorrow is the last day and I know I have mixed feelings. It'll be great to get home again and I know my family will be happy to see me. But I'm also a bit sad that the friends here will be parting ways soon and it will probably be at least a year before seeing most of them again. Also, I will miss the atmosphere where you can discuss a problem you are facing and have a good chance of including one of the genuine Oracle guru's (or godesses) in the discussion. I'm looking foreword to Brownies with the Board tomorrow for the Annual Membership Meeting. There is always interesting information shared there about the user group movement.
So long for now.
The conference is almost over. What is the best thing for you about a user group meeting or conference?
Thanks for writing these
Stephen, I want you to know I really appreciate your weblog entries about this year's IOUG, the more so because for the first time in several years I couldn't attend myself.