Back to newsletter 104 contents

A successful performance expert will walk into a customer's project on day 1, and if you give him or her full access to the people who understand the system, will have enough understanding of the system by the end of day 2 that they can start pointing out architectural bottlenecks, and that they can start describing the tests that need to be performed to identify the implementation bottlenecks. (I should note, for any potential customers out there using a performance expert, if you don't give them full access to the people who know the system, it can take a lot lot longer to understand the system).

A note from this newsletter's sponsor

I was wondering, recently, is that a skill that your average programmer picks up with enough experience, that ability to understand the system quickly and with sufficient clarity that you can so quickly understand its bottlenecks? Or is it a skill that performance experts have because they've needed to gain it, while others don't? I don't know the answer, I suspect it's a matter of enough exposure to different systems and sufficient interest to want to 'get it'.

What triggered my musing was doing a knowledge transfer on performance tuning to a couple of developers for their system. They already knew their own system, of course, so there wasn't a need for them to need to quickly acquire the system knowledge. Instead, they needed to acquire my analysis and tuning skills. The success of that kind of skills transfer comes down to how interested the developers are, and how much time they have to practise the skills I transfer. I've seen previous skills transfers result in people who have gone on to understand performance tuning to expert level; but I've also seen less successful transfers, usually where the developers were not give the time to bed in the new skills before they had to shift to something else.

Well, enough of me rambling, now on with this month's newsletter. As ever we have all our usual Java performance tools, news, and article links. Javva The Hutt tells us about his intern and the rogue monkey; there's a new cartoon at fasterj of synchronized methods in action; and, of course, we have extracted tips from all of this month's referenced articles.

A note from this newsletter's sponsor

News

Java performance tuning related news.

• java.nio.BufferPoolMXBean platform mbean management interface added in Java 7 to monitor direct buffers
• JEE 6 new features include: singleton beans; asynchronous session beans; EJB lite; asynchronous processing (supports AJAX) in Servlet 3.0
• Core Java Concurrency reference card
• memcpy of arrays into the JVM heap can be a race condition against array readers. Fascinating discussion
• JFX 1.2 performance update using Balls benchmark

Tools

Java performance tuning related tools.

• CollectionSpy profiler dedicated to profiling collections
• Sametime Alarm System for Server Performance, a real-time system monitoring application
• Processor Time Analysis Tool for Linux helps you identify Java threads that consume the most processor resources

A note from this newsletter's sponsor

Articles

• A Beautiful Race Condition (Paul Tyma/Malinator)
• Continuous Performance Testing (Kirk Pepperdine/pepperdine)
• Laurence Vanhelsuwé on Java Collections Pitfalls (Frank Sommers/Artima)
• Volatile Arrays in Java (Jeremy Manson/jeremymanson)
• Hibernate Performance Tuning (Ming Jiang/JavaLobby)
• Selecting the Best Java Collection Class for Your Application (Mark Grand/Developer.com)

Jack Shirazi

Back to newsletter 104 contents