Azul Zing kicks Java up a notch

Azul Zing kicks Java up a notch

Summary: Java application performance depends upon the Java Virtual Machine being able to quickly create, use and then destroy pieces of data. Azul Systems has taken its decade of experience building Java acceleration systems and built a new JVM based upon Oracle's HotSpot technology.


Howard Green, Director of Marketing of Azul Systems, reached out to me and wanted to talk about the company's next exciting innovation, a new Java interpreter, called Zing, that is built upon upon what Azul Systems has learned over ten years of accelerating Java performance.

What does Azul say about Zing?

Zing™ is a 100% Java-compatible JVM based on Oracle HotSpot. Zing is optimized for Linux deployments and designed for enterprise applications and workloads that require large memory, high transaction rates, low latency, consistent response times and high sustained throughput, Zing is the only JVM that can elastically grow its application memory heap based on real-time demand and still guarantee response times. 

At its core, Zing uses the Azul C4 (Continuously Concurrent Compacting Collector) which eliminates "stop-the-world" pauses which limit the scalability of all traditional JVMs. Zing also includes a management and monitoring platform with a zero-overhead, always-on visibility tool called Zing Vision and the Zing Resource Controller management tool.

The Zing JVM is simple to install and requires no coding change to the application. Because Zing is both elastic and self-healing, tuning is typically reduced to just a few parameters. Simply point your application or startup scripts to use the Zing JVM, and you're running on the most elastic, scalable JVM with the fastest time-to-deployment for any size application.

Snapshot analysis

Approximately 10 years ago, I had the opportunity to meet Howard Green and key members of his team at a restaurant in, if my memory serves me well, San Mateo, CA. At that time, I was responsible for IDC's system and virtualization software research services.

While we enjoyed breakfast, the team discussed their plans to develop an appliance server designed to accelerate Java applications and services. The team was excited when they described the developed of a new type of multi-core processor designed to deal with interpreted languages such as Java. At that time, they hinted that this processor technology would be useful for other languages in the future. In time, the product described in that breakfast briefing was introduced and took Java execution to new heights.

Zing goes right to the heart of the challenges of accelerating a dynamic, interpreted language, such as Java. Programs create, use and then destroy little pieces of data during execution.  If not handled correctly, these unused and no-longer-needed pieces of data can choke the intrepreter. So, "garbage collection" performance is a key challenge for Java virtual machines. Azul had to think very creatively ten years ago when designing processors for Java. It is clear that what was learned then has been baked into Zing today.

If your organization relies on Java-based services or applications, it would be wise to learn more about Zing.



Topics: Virtualization, Cloud


Daniel Kusnetzky, a reformed software engineer and product manager, founded Kusnetzky Group LLC in 2006. He's literally written the book on virtualization and often comments on cloud computing, mobility and systems software. In his spare time, he's also the managing partner of Lux Sonus LLC, an investment firm.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • Maybe NASDAQ could use this?

    Quoted in the article:
    "optimized for Linux deployments and designed for enterprise applications and workloads that require large memory, high transaction rates, low latency, consistent response times and high sustained throughput

    Wasn't Java involved in the software glitches at NASDAQ that adversely impacted the Facebook IPO? If I recall correctly, the NASDAQ system runs on Linux.
    Rabid Howler Monkey
    • adverts now

      sigh, so this is a "we've got the fastest JVM that fixes all problems ever imagined in Java" (except the java design concepts themselves, of course) product. No doubt it'll cost a fortune and not provide the results that everyone currently running C/C++ code gets for free.

      Anyway, why isn't this piece marked as "advertising for Azul", its all fluff - 10 year old technology for running parallel tasks suddenly seems to be "baked in" to Zing. Pretty vague really, and how a C4 GC has any impact on low-latency processing is left to one's imagination, and if they are still running a Java interpreter, rather than a JIT compiler, performance is going to be terrible.
      • sounds like advertising indeed

        I mostly see praising here with the exception of C4 which I can tell you I'm eager to check out.

        First of all, C/C++ programs are not cheap to produce, they never were. you need good, experienced programmers for that. on the other hand, even monkey can code Java (I know it's a joke but really for a lot of stuff you only require an average IQ guy...). while cheap to produce, java programs always have and always will rely on garbage collection which traditionally requires a pause (TOTAL pause, the application is stopped entirely) to process. if you need to ensure a low latency for EVERY real-time request you better have only small pauses for garbage collecting (there are algorithms that perform decently) or NO PAUSE AT ALL, this is what C4 is supposed to feature.

        there you go, C4 if working as advertised, would be pretty good :)
    • Azul CEO suggested it

      bad coding was involved. it is known that bad coding can halt a system, regardless of programming language or operating system or system resources :D

      in an article, Azul CEO commented on the problem and also mentioned his JVM. but in my experience, no JVM can whitstand bad coding.