After almost two decades in IT, at numerous clients in any and all sectors, system performance has run as red thread at just about every site that passed my path. Performance tests has been been made and sometimes even evaluated and acted upon in an acceptable fashion. At some sites there has been critical situations and at other there has been no problems at all. At the sites where problems has been identified, resources has (after a while) been poured in, in terms of bodies and silicone to rectify the problem. A quick solution has often been to "just" add more hardware. In some cases lots of time and money has been spend to analyze the problem and doing everything from specific performance enhancing fixes to complete re-designs and re-implementations of the systems in question to navigate around resource hogs.
Nowhere though, in all this time, at any of these locations, among all theses people, has the solution - or even idea of solution - existed to utilize a high-perfoming JVM. This could be due to that the existing available solutions aren't mature and good enough for the problems at large, cost or just ignorance.
Maturity certainly seems to have been a factor of the past but to date ignorance still appears to be the strongest.
After reading up about Oracle JRockit, I'm exited to try it out on a (or several) live system. This will one of the major goals of the remainder of the year. Expect results to be published here.