Platform Prime enables 20% cloud savings with Java optimisation

A global enterprise has deployed more than 10,000 Java Virtual Machines (JVMs) using the Platform Prime solution with Optimizer Hub, resulting in reduced cloud costs and improvements in application performance.

Cloud deployments

The deployment, which covers a range of applications and microservices from eCommerce to payment and inventory management, makes use of Optimizer Hub to enable JVMs to collaborate and share performance optimisations across the fleet. According to Azul, this collaboration has resulted in significantly faster application startup and warm-up times, smoother application scaling, and a reduction in compute requirements that has contributed to cloud cost savings exceeding 20%.

Another company in the entertainment industry recently utilised Optimizer Hub to achieve a 25% to 30% reduction in CPU core and pod counts for its Java-based critical services. Optimizer Hub is included as an optional, customer-managed service within Platform Prime and does not require changes to existing Java applications or JVM-based workloads.

Addressing traditional JVM limitations

The scale and elasticity requirements of modern cloud-native Java applications present challenges in maintaining user experience, managing auto-scaling in response to demand, and efficiently handling frequent restarts during application updates. Traditional JVMs optimise performance in isolation, which can result in less efficient code execution and unpredictable behaviour. This often forces enterprises to over-provision cloud resources to maintain service levels, increasing costs. Recent survey data cited by Azul indicates that 71% of respondents report more than 20% of their cloud compute capacity remains unused.

By allowing JVMs across enterprise fleets to send and receive optimisation data using Optimizer Hub's centralised services, Platform Prime seeks to improve cloud performance, elasticity, resilience, and cost-effectiveness at scale.

Centralised performance optimisation

The Platform Prime suite includes Azul Zing, an OpenJDK-based Java runtime, and the Optimizer Hub. Optimizer Hub features two main services designed to enhance performance for containerised, cloud-native environments:

Cloud Native Compiler provides centralised just-in-time (JIT) compilation and caching, reducing the computational workload for each JVM by centralising these tasks. By doing so, it enables consistently faster code execution and allows JVMs to operate in smaller instances, which can further decrease cloud expenditure.

The ReadyNow service addresses JVM warm-up times by saving and reusing JIT compiler profiling and optimisation data across JVM runs. The ReadyNow Orchestrator component ensures that relevant application profiles are used efficiently, with the aim of supporting high service-level agreement (SLA) attainment, faster recovery from failures, and streamlined DevOps workflows, even during periods of high demand.

Customer and executive statements

"Java powers the backbone of the digital economy, but performance challenges with traditional JDKs have led enterprises to overprovision cloud resources and adopt complex operational practices," said Scott Sellers, co-founder and CEO at Azul. "With the centralised services of Optimizer Hub, Azul Platform Prime has eliminated long-standing tradeoffs of performance and cost by enabling JVMs to learn from each other and collaborate across entire fleets in production, delivering faster, smoother application experiences while driving down cloud costs by 20%+. Optimizer Hub is a must-have for every business running JVM-based, mission-critical workloads in the cloud."

Platform Prime and Optimizer Hub are positioned as solutions to address inefficiencies in conventional JVM deployments, aiming to deliver improved performance and lower infrastructure costs without requiring modifications to existing Java applications.