Back to newsletter 110 contents

The Oracle-Sun merger is through now, and Oracle announced a merger strategy (see our newsletter news items below for more details). It was good news for Java in that there was no real bad news while at the same time there was commitment to lots of Java investment from Oracle. As far as Java is concerned, Oracle seems to intend to put even more resources into it than Sun on its own did, which bodes pretty well for Java. The overall situation is summarized by a quote from Ellison: "Sun didn't make a lot of money from Java but we sure did". I expect the combination to continue to make money from Java, and that's almost certainly a good thing for Java - Sun's lack of significant profits from Java was always leading analysts to predict cutbacks on the Java side, now that will turnaround.

A note from this newsletter's sponsor

The only performance related aspect I could see was the intention to merge JRockit to HotSpot - surely a good thing as they each have pretty good features that the other lacks. The focus on garbage collection, multi-core processing and real-time monitoring is spot on - that is where it needs to be.

Oracle are likely to end up with more JVMs than Sun had (including several micro-kernel ones and real-time ones), but I would expect the reference free one associated with the open source project to continue to have as good performance as all of them, expect possibly for highly scaled applications - where enterprises are willing to pay extra to gain the benefits of handling those high scales.

My feeling is that this takeover is ultimately good for Java. That Oracle is maintaining NetBeans is testament to how much is being invested in Java - with three IDEs supported by Oracle (Eclipse, JDeveloper and NetBeans) it was surely a candidate for cutting back on. The JavaFX investment and the plan to combine with Javascript is unsurprising, that RIA space is where the desktop application competition is, and I think in a couple of years we will have a four-way Adobe/Microsoft/Oracle/Google competition for most of the developer market share here.

Now on with this month's newsletter. We have all our usual Java performance tools, news, and article links. Over at fasterj we have a new cartoon showing multi-threaded code; Javva The Hutt tells us about a colleague's interviews; and, as usual, we have extracted tips from all of this month's referenced articles.

News

Java performance tuning related news.

• Oracle Java Strategy Details 1
• Oracle Java Strategy Details 2
• Oracle Java Strategy Details 3
• Oracle Java Strategy Details 4

• performance improvements in the Java 6 Update 18 release
• performance updates in Java6 release 18
• jdk7 status update
• Asynchronous Support in Servlet 3.0

A note from this newsletter's sponsor

Tools

Java performance tuning related tools.

• A Pragmatic Way of Doing Root-Cause Analysis
• Wireshark network protocol analyzer
• Virtual Machinery's JHawk product generates vital Java code metrics allowing you to improve the performance, reusability, maintainability and overall quality of Java code
• New Relic RPM full featured production and devlopment stage performance monitoring, alerting, analysis and diagnosis tool
• Google Collections Library finally released (no longer beta)

Articles

• A Note on the Thread (Un)safety of Format Classes (Jeremy Manson/jeremymanson)
• The Missing Piece in the Virtualization Stack (Nati Shalom/natishalom)
• Java profiling with DTrace (Wayne Adams/wayne-adams)
• Overview of Performance Tuning Tools in Oracle Fusion Middleware (Olaf Heimburger/Oracle)
• Data scaling at Shopzilla volumes (Rob Roland/Shopzilla)
• Java Profiling: Under the Covers (Wayne Adams/JavaLobby)
• Minimize Garbage Generation: GC is your Friend, not your Servant (David Holmes/Sun)

Jack Shirazi

Back to newsletter 110 contents