I envision a time when automated optimisation process will monitor your program every time it runs in production, automatically adjusting the tweaking parameters until it gradually learns and homes in on the optimal settings for running under various conditions using some sort of statistical multiple regression analysis and binary search, working somewhat like a numerical differential equation solver.
The problem is the optimum settings are not universal. They depend on the customer’s hardware, JVM (Java Virtual Machine) and data size and characteristics, e. g. one customer may always present data pre-sorted, or sorted in precisely reverse order. Different customers use different SQL (Standard Query Language) engines and different Servlet wombs.
Tweakables could be external to the Java program such as the optimum VM heap size or stack allocation. Too little and you GC (Garbage Collection) needlessly frequently, too big and your virtual RAM fills with paged out deadwood interspersed with tiny objects.
This page is posted
Optional Replicator mirror
|no blog for this page||Canadian
Your face IP:[220.127.116.11]
You are visitor number|