Azul Systems has released a new version of Zing, its flagship Java runtime, optimized for container and cloud deployments. Zing 19.07 eliminates the requirement for Zing’s ZST memory management kernel module, and extends its performance and throughput advantage when executing ad hoc and query-based workloads.
Sunnyvale, Calif.-based Azul bills itself as the only vendor focused exclusively on Java and the JVM. Zing is based on Oracle’s HotSpot, a core component of Java SE. It’s is a “no-pause” JVM designed to eliminate Garbage Collection (GC) pauses, a long-standing challenge for Java developers. This pauselessness, which Azul calls “generational pauseless garbage collection” (GPGC), enables Java app instances to scale dynamically and reliably. Azul CEO Scott Sellers has called GC “the Achilles heel of Java.”
Zing’s ability to perform complex memory manipulation functions and support memory allocation rates without pausing had been dependent on a kernel module — the ZST — which re-allocates memory pages from the operating system to guarantee memory access without paging or other performance-robbing processes. But using the ZST required root/administrator privileges to install and configure. By making the kernel module optional in this release, Azul eliminates the need for root/administrator privileges during installation, which greatly simplifies container and cloud deployments.
“With the release of Zing 19.07, use of the ZST module is now optional, making it far easier to install and deploy Zing across a variety of environments including container and cloud deployments,” the company said in a statement. “When deployed on properly configured supported versions of Linux, performance and latency characteristics of the C4 garbage collector [are] indistinguishable, with or without using the ZST.”
ZST will continue to be fully supported for applications requiring the lowest consistent latency or terabyte-scale Java heaps, the company said.
Starting with this release, Zing now comes bundled with Zulu Mission Control, the Java performance management and application profiling tool based on the OpenJDK Mission Control project.
This version of Zing also extends the runtime’s performance and throughput advantages for ad-hoc queries, which is a common use case in enterprise search and decision support applications — apps based on, for example, Apache Lucene, Elastic, or Solr — or analytics applications, such as those based on Apache Spark.
Other use cases include electronic trading, online advertising, real-time messaging, Big Data (Hadoop, Cassandra), and in-memory caching.