The promise of cloud-native was always elasticity: scale up instantly to meet demand, scale down just as fast, and only pay for what you use. But for enterprises running fleets of Java Virtual Machines (JVMs), that promise has been elusive. Startup delays and unpredictable performance have forced teams to overprovision, keeping servers always-on “just in case.” The result: wasted resources, rising costs, and missed sustainability goals.

Explore more in our Cloud Native hub.

In a recent episode of Java Reloaded, I spoke with John Ceccarelli, VP of Product Management at Azul, about how their new Optimizer Hub addresses this long-standing challenge. His message was clear: enterprises can finally break free from Java’s warm-up penalty—and save nearly 20% on cloud spend—without rewriting a single line of code.

Taking JVM Performance Off the Server

As Ceccarelli explained, the heart of the problem lies in how Java optimizes code. “The JVM has to actually run the code a number of times before it can optimize it,” he said. That means new instances start slowly, burn CPU on compilation, and deliver poor user experience during warm-up. Optimizer Hub changes that equation by moving the entire optimization process to an external service.

“When the JVM wakes up, it asks Optimizer Hub, ‘Help me out here,’” Ceccarelli said. The service has already cached optimizations from thousands of JVMs, so it can instantly deliver tuned code. The result: near-instant startup, smoother CPU usage, and the ability to scale confidently on Kubernetes or serverless platforms.

Lower Bills, Better Performance

Performance gains are only half the story. Ceccarelli broke down the mechanics of cost savings:
– Pauseless garbage collection removes stalls, reducing the need for idle reserve capacity.
– Azul’s JIT compiler produces code 40–50% faster than OpenJDK.
– Auto-scaling becomes viable, letting teams shut down servers instead of leaving them always-on.

Together, these improvements add up to about 20% savings on cloud infrastructure, while also improving end-user experience. “The best way to save money on a server,” Ceccarelli noted, “is to turn it off.”

Seamless Migration, Enterprise-Grade Security

For enterprises already invested in OpenJDK, migration anxiety is real. But Ceccarelli emphasized simplicity: “There are no code changes, no recompilation. You switch the JVM, hook it up to Optimizer Hub, and off you go.” Optimizer Hub runs inside the customer’s Kubernetes environment—not as a SaaS service—keeping performance data securely within their estate.

FinOps Meets Java

Ceccarelli, a FinOps Certified Practitioner, sees JVM optimization as a natural extension of cloud cost governance. “People never think of the JVM as a FinOps tool,” he said. “But just by swapping it, you can reduce cloud costs significantly without rehosting, refactoring, or re-architecting.”

With enterprises facing mounting pressure to optimize spend, Azul is positioning Optimizer Hub as both a performance solution and a financial lever. It’s a rare case where engineering and finance goals align perfectly.

Looking Ahead

The competitive landscape is shifting, with legacy tools fading and new approaches like Project Leyden experimenting inside OpenJDK. But Azul’s bet is that enterprises need solutions today, not experiments tomorrow. “We just wanted to tell the world this is real, this is here right now,” Ceccarelli said.

For organizations struggling with Java performance in the cloud, Optimizer Hub offers a practical, immediate way to unlock elasticity, cut costs, and simplify operations.