News

Hazelcast Adds Support for Azul's Zing JVM

• By John K. Waters
• March 23, 2016

Java runtime maker Azul Systems is partnering with in-memory data grid specialist Hazelcast to provide support for Azul's Zing Java Virtual Machine (JVM). The JVM is now certified on Hazelcast Enterprise 3.6.

Azul's flagship Zing JVM is based on Oracle's HotSpot JVM, a core component of Java SE. It's 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. Scott Sellers, CEO and President of Azul Systems, has called GC "the Achilles heel of Java."

Hazelcast develops, distributes and supports a widely used open-source in-memory data grid (IMDG). The product, also called Hazelcast, is licensed under an Apache license that allows developers to include the grid in their applications. Hazelcast Enterprise is the commercially supported version.

The partnership, which certifies Zing for demanding in-memory workloads, is aimed at developers of memory-based webscale Java applications, said George Gould, VP of business development in Azul's partner alliances group.

"When we talk about the evolution of Java and we think about deploying an application that can use tens of gigabytes, Java can generally keep up," Gould told ADTmag. "But when you're trying to put gigabytes if not terabytes into an in-memory grid, you need to look at the JVM as an important part of that runtime stack. Zing provides that elastic Java layer that can simplify configurations and provide the necessary performance."

The two technologies are highly compatible, said Hazelcast CEO Greg Luck, but providing the increased availability, enhanced performance, and "ultra-scalability," required so serious collaboration.

"In the world of Ops, there are more deployment options than ever before," Luck said. "Different cloud deployments and different virtualizations, and then in Java itself there's the fairly standard implementations of the JDK. And then there's Azul, which has worked for many years to solve the GC issue. Both companies are innovators in the field of Java, and by working together we have created a solution that gives developers and DevOps teams with a very powerful and compelling platform."

The combination of Azul and Hazelcast greatly extends the operational lifetime and capacity of existing memory-intensive applications, Gould said. He cited a recent example, in which a system that was constrained by a legacy JVM could only support 40,000 line items, but was extended to support 1,500,000 in-memory catalogue entries simply by deploying Zing and with no change to the application.

"The commodity server has come of age and outpaced the the Java runtime's capability to keep up," Gould added. "If you deployed on a gigabyte 10 or 15 years ago, that was a big Java instance. Now we're talking several terabytes, and we can support that. And I think that's why it's become more relevant. And as people use more data, more cores, more horsepower, the user's expectation is that it'll work just as well as it did at a gigabyte. And providing that elastic scale is key. People just expect the JVM to work."