Azul Java learns to cut warmup times

Java software provider Azul has added a capability to slash warmup times for Java applications using the company’s Azul Platform Prime runtime.

The capability, called ReadyNow Orchestrator (RNO), “delivers the highest possible optimized code speed at warmup,” the company said, enabling improvements to operational efficiencies and optimization of cloud costs. RNO is included as part of the runtime, at no additional charge.

With this capability, Azul said it was looking to address a situation in which business-critical workloads using Java face a warmup problem. When a Java application is launched, the JVM must compile it into a form that can be executed by the machine or device running it. As the application keeps running, the JVM will recompile and further optimize important code to boost performance, essentially “warming up” over time before it reaches peak performance.

RNO records information about an application’s optimization profile and then uses this to shorten the warmup time the next time the application runs. Profile distribution is automated by delegating profile collection to a dedicated, customer-managed service. Rather than collect profile information on a single JVM, RNO monitors fleets of JVMs, learns from application usage what the best optimization profile is, and then serves the profile to any JVM requesting it. So applications warm up quicker.

Azul Platform Prime can be downloaded from azul.com. The new feature builds on Azul efforts to help businesses optimize rising cloud costs, Azul said. Devops teams can scale down the number of cloud compute instances they use to run Java applications during off-peak times and then scale them back up to meet demand. Organizations can reduce the average number of compute instances being used, thus lowering cloud compute costs. Azul Platform Prime formerly was known as Zing.