Java runtime solutions company Azul and Apache Cassandra company DataStax have partnered to allow DataStax Enterprise (DSE) customers to use the Azul Zing "Java supercharger" virtual machine.
With this news, Zing is now a certified Java Virtual Machine (JVM) for DataStax Enterprise (DSE) — Zing is good for real-time deployments that need to leverage large in-memory datasets and caches.
As Dr. Dobb's readers will know, core big data technologies such as Cassandra, Solr, and Spark are written in Java requiring a JVM for runtime execution.
Azul Zing is said to be the only JVM that implements "pauseless garbage collection", which provides highly consistent Java runtime performance independent of an application's memory requirements.
Through this partnership, DataStax users get low latency, real-time solutions for demanding applications requiring an ever-increasing amount of in-memory data such as fraud detection, website personalization, payment systems, and time-critical decision support.
"DataStax is the distributed database management of choice for enterprises, and together with our partners we offer innovative solutions that complement our technology," said Matt Rollender, vice president of Infrastructure and Ecosystem Development, DataStax. "We are pushing the envelope with Azul Systems by delivering an incredible boost in performance for JVMs."
"Companies depend on real-time big data systems to maximize revenue and mitigate operational risk," said Scott Sellers, CEO of Azul Systems. "Zing was created to allow Java applications and open source databases like Cassandra to support high throughput, real-time, and low latency use cases even with massive in-memory datasets. We are excited to be working with DataStax to bring these benefits to more enterprises."
Zing is the best JVM for real-time Cassandra deployments. Zing allows Cassandra to operate more consistently by eliminating JVM-caused response time delays. With Zing, each Cassandra node can scale to use 1 TB of in-memory data while remaining capable of delivering maximum response times below 20 milliseconds — a level of response time performance unmatched by traditional databases.