i’ve been in training the last few days, taking a class on optimizing oracle performance. note that i’m not an oracle dba by any stretch, but surprisingly i’ve been able to follow 90-95% of the class. it’s funny because i sit here and learn things that i may possibly never use since i don’t administrate oracle databases, but for me it’s more about seeing what causes performance problems and how to attack them, allowing me to gain better insights into what customers face and how to help them address their true issues instead of blaming our storage systems.
an interesting concept brought up was CTD: compulsive tuning disorder, which is a phenomenon originally coined by gaja krishna vaidyanatha & kirtikumar deshpande in their book oracle performance tuning 101 (maybe more about this later, i think i might actually have this at work). generally speaking, it’s the obsession of trying to make things better by tweaking everything, looking for gains in everything that is there. the problem usually arises when you don’t have a clear picture of what to actually is best to do and worse off, what effect doing anything in particular will make.
the interesting methodology to counter CTD is to specifically attack the largest causes of delays in response time: first change the thing that will effect what users will notice the most. and not only that, you want to do this for whatever business process is actually the most important overall.
thus if it’s taking twenty minutes to generate a purchase order, attack that first, because that’s important to the business, and then look at what’s causing someone to feel that the twenty minutes is too long, and fix that first. who cares if some job that runs in the middle of the night runs an hour longer than it should. it might be an easier fix, but why mess with that until much later?
so what happens if you apply this methodology to your life? this means that instead of running around constantly worrying about everything in your life and trying to make everything little thing better, instead you should try and find the things that really matter to you, and make improvements in those things first.
sure, i could spend time trying to get my printer to work wirelessly or figure out a way to switch between my tivo and my dvd player without a manual switchbox, but isn’t it better trying to come up with a more effective process to manage my day-to-day workload? or finding a job that i’m more enthusiastic about?
there’s this book floating around called getting things done: the art of stress-free productivity, which either is a good example of this or the antithesis. i haven’t had a chance to figure out which. there’s also a blog called 43 folders which is someone’s compilation of tips and organizational methods loosely based on the book, which is interesting and intriguing at the very least.
the other thing that’s interesting to me about this optimization approach is that it serves as a reminder to look at problems from a different way from everyone else. don’t accept the conventional wisdom on anything, but approach it from another direction, something that makes sense to you.
normal oracle performance tuning espouses reducing large numbers of event counts and ignoring idle event times, both of which may be fatal errors. sure, reducing the number of events may be helpful, but if you’ve got some events that are taking way too long because of how they’re erronously constructed, reducing the other events won’t appreciably help.
doing things contrary to the popular method for the sake of productivity is oddly innovative. i loathe baseball, but this article in the new york times on ichiro suzuki was fascinating: it analyzes his various hitting swings, and finds that he has not just one, but several unconventional swings. this allows him to adapt to a wide variety of pitches and hitting situations, and now he’s on the verge of breaking the record for most number of hits in a season, a mark that has stood since 1920!
While hitters typically strive for one consistent swing, the 30-year-old Suzuki has developed a multitude of ways to beat a ball into the ground, flare it over a third baseman’s head, sneak it up the middle or smash it into the gap.
His variety of swings, unique even in his native Japan, has made him impossible to scout and difficult to defend, helping him capture the 2001 American League Rookie of the Year and Most Valuable Player awards and yielding a major-league-leading 231 hits this season. To eclipse Sisler’s record, Suzuki needs 27 hits in the remaining 19 games. He may be the only man who could swing it.
…
“The game is just different for this man,” Molitor said. “He sees spaces on the field and he guides the ball where he wants it to go, like he’s playing slow-pitch softball.”
…
His sense of timing and location has resulted in many of his bleeding singles and infield hits. Suzuki has 196 singles, breaking his American League record for singles (192 in 2001), partly because of ground balls that are placed in the perfect spot.
“He’s made an art out of it,” said Bret Boone, the Mariners’ second baseman. “I’ve seen the guy have five at-bats in a game and get four infield hits. No one in baseball can get a hit in as many ways as he can. He goes up with so many different approaches, and he decides pitch to pitch which one he’ll use.”
a couple of other quick thoughts:
Comments are now closed for this entry. Thank you for playing.