Specifics about performance tuning and cloud computing figure importantly in the work of Java Champion Kirk Pepperdine, who told Janice J. Heiss that generic advice is useful but only if it is evaluated in terms of particular situations. "Performance tuning," Pepperdine contends, "is still an art where limits are stretched and one goes to places never gone to before. It's an experimental and creative realm with no one-size-fits-all answers to performance problems. It's highly localized."

One of the reasons the art of performance tuning is difficult, Pepperdine conjectures, is that those who might be expected to be good at it -- i.e., developers -- tend not to be because the skills called for in performance tuning are vastly different from those useful in writing performance code. Even the thought processes are different, he maintains.

Developers too often overlook the dynamic nature of information systems and, as a consequence, do the wrong thing when it comes to performance tuning, often writing far more complex code than the circumstances warrant. Testers, on the other hand, fare somewhat better because their thought processes are more congruent to what performance tuning requires.

According to Pepperdine, simplifying code frequently leads to the performance improvements being sought. This is especially the case, he continued, with Java, where runtime optimizers are at work and can be confused by complex code.

One of the most important factors involving the work of developers and performance tuning is the pace of their work. They frequently lack the time to go back and refine the code they have written, code that in itself often reflects a poor understanding of the initial problem.

Performance tuning in the cloud, like performance tuning involving large scale databases presents the same sort of sand trap for developers, which is further complicated by the absence of adequate tool sets for accomplishing the diagnostic work in either arena.

"All the variability in cloud computing makes it more difficult to define performance problems. You cannot go to your manager and say that you will find the problem by Tuesday and fix it by Thursday — which you can do with a number of problems," Pepperdine said.

Responding to Heiss's final question (what do you look forward to in JDK 7), Pepperdine said he hoped there would be better ways of managing concurrency in the multicore world, which is one way to ensure Java's continuation as the platform of choice.

More Information

Become Proficient in Java Performance Tuning Pepperdine course

Java Performance Tuning

JDK 7 Project

Java Performance Tuning Workshop

Chief Architect Previews JDK 7 [...read more...]

		fullsource
Print Email Current Profile Archive Blog Generate newsletter