New! JVM Inventory, a feature of Azul Intelligence Cloud, accelerates Oracle Java migration and ensures ongoing compliance - Learn More

Support

English

English

Speak to a Java Expert Download Now

A Guide to Selecting a Performance Java Runtime

Cloud Cost Optimization for Java Workloads

Download Platform Prime for Free

Which Java runtime is the best fit for you?

Azul Platform Prime is the world’s fastest and most efficient JVM. It’s perfect for companies looking to optimize their Java workloads in the cloud. Use Platform Prime to reduce your cloud costs and meet performance SLAs.This comprehensive guide provides ten important factors to consider when choosing an optimized Java Virtual Machine (JVM).

Not all Java runtimes are created equal

Vanilla implementation of the Java Platform Standard Edition

The Hotspot JIT compiler in OpenJDK determines how fast
your code executes.
HotSpot has limited optimization levels, and that’s as fast as Java can possibly run – we think that’s definitely not enough!
Time and tuning to get to eventual code speed vs local CPU resources available for JIT compilation. We suggest a different approach

Built on OpenJDK and enchanced for faster warmup and peak performance

Azul’s Falcon JIT Compiler – faster code execution from improved optimization
Azul’s C4 Pauseless Garbage Collector – eliminate latency outlers, jitter and timeouts
Azul’s ReadyNow – accelerate warmup times to run faster, sooner
Azul Cloud Native Compiler – use the cloud to compile code for faster execution and greater efficiency

Top 7 Considerations for Using Java to Reduce Cloud Waste

Top Strategies Adopted by Azul’s Global Customers

1. Are you over-provisioning to address business requirements?

Azul customers agree that this is typically a universal behavior, because they need to meet SLAs.
But, many teams fail to optimize what they throw at it. Let’s face it, moving utilization up is very meaningful – and waste is sometimes necessary – but no always!
Our customers have switched from OpenJDK to Azul Platform Prime because it is an enhanced version of OpenJDK with superior performance, carrying capacity, and user experience that prevents over provisioning and requires less performance tuning and reduces how much you need.

2. Are you running code slower than it can go?

What if you could squeeze more out of every CPU? Are you running on a 10 year old CPU?
Our customers do that every day, without having to rewrite any code! They have found a way to use the fastest CPU for their dollar for older and newer chipsets?.
By choosing Azul’s optimized Java runtime they improved carrying capacity at higher utilization levels without sacrificing SLAs.
Azul Platform Prime accelerates Java warm-up times; delivers faster machine code; and eliminates latency outliers to run 2-5X more transactions through your infrastructure—without pauses, jitters, or timeouts.

3. Does workload inconsistency mean you need to provision more headroom?

This is the biggest potential for savings. The more predictable the utilization is, the more comfortable you feel.
Before Azul, many of our customers didn’t know when some loads will hit so they set utilization at lower levels to be more predictable and meet SLAs, and that lead to more waste within a typical business day.
Bottom line: workload inconsistency causes the need for more headroom to manage the swings.
Azul customers deliver less variability and more certainty under loads – and they push more typical workloads a little further. With Azul Platform Prime you can push your hardware harder and improve carrying capacity. Azul customers have delivered on performance SLAs even as loads increase, ensuring the fastest response time for latency-sensitive systems because Azul turns byte code into faster machine code. We also use the experience of previous instances at warm-up…by sharing the learning of prior instances we eliminate the need to relearn.

4. Are sprawling infrastructure software technologies driving your costs higher?

Azul customers have faced growing inefficiencies in infrastructure for modern big data technologies that are driving costs higher.
Let’s face it – with data processing – data grows….and there’s always going to be more of it.
Technologies including Kafka, Cassandra, and Solr/Elasticsearch that stream, store and search data respectively are common Java-based frameworks for distributed, cloud-based systems.
Azul customers have learned how to balance data processing needs with cloud costs – using Azul to bring compute efficiency for their data platforms. They use the most common Java-based frameworks…a lot of them are built and run by our JVMs. Azul’s performance runtime, Platform Prime, effects CPU use and memory. You can drive Solr and C* to higher utilization – and that means reclaiming compute nodes. With Azul Platform Prime your code runs faster and there is a raw speed benefit and carrying capacity

5. Are common services in your application and development platform not cost efficient?

There’s always a need to improve common services in your development platforms: including monitoring and logging using Kafka or Elasticsearch; evolving the golden images with permitted software for the organization; or optimizing Kafka throughput for connectivity and messaging services that access back-end systems
Before Azul, our customers were struggling to streamline costs and drive governance with adoption of FinOps practices.
Azul’s performance runtime drives applications that are unique to our customers. Azul Platform Prime is used by the best global engineering teams to continually improve performance levels – increase developer efficiency and productivity.

6. Is the cost to serve increasing faster than your business growth?

This is more important than it ever has been when running a digital business.
We all want to avoid a re-architecture for lift and shift Java workloads in the cloud.
Before Azul, our customers were struggling to satisfy business requirements, ensure a good customer experience AND lower cloud waste. As apps mature and data volumes grow, we are seeing a renewed focus on keeping engineering resources from eating into margins.
Azul reduces the inherited efficiencies that happen as you move into more Business as Usual. Using Azul Platform Prime reduces operational complexity and requires less engineering time, reduced triage time and fewer operational/support tickets. You free up engineering resources to deliver on revenue-generating services.

7. Are you not using what you pay for efficiently?

With cloud costs, our customers have become really good at reserving spot instances – it’s a good start.
But EC2 instances and AWS bills are becoming increasingly complicated as companies move to abstracted infrastructure like Kubernetes.
Our customers run complex, advanced systems – with a true cost that can be rolled up and buried into EC2 bills. Managing with utilization levels are emotional journeys for teams. Nobody wants to say the ideal number out loud.
Azul’s Platform Prime can run higher utilization levels and is willing to push harder to get more utilization (throughput) before auto-scaling with a new instance. The smallest improvements in utilization has huge impact on your cost – especially when you need consistency, repeatability, and predictability (same load levels, etc.).

Your Application Development Platform – Optimized

Run the Java-based infrastructure software that powers your internal platform more efficiently

Azul Platform Prime improves the efficiency and lowers the cost of running your internal development platform

Streaming and Events

Improve large-scale Kafka installations – giving you up to 45% more throughput and overall latency.

Search and Logging

Get better price and performance for large Elastic/logstash, Splunk, or Promethius installations

Containers and Container Orchestration

Enhance your pre-built container images to make systems and applications go further with lower latency, higher throughput under

Images are built using a Java Virtual Machine (JVM) that makes production code start fast, go fast and stay fast. You get the fastest machine code no matter what compute platform you are using.

Cost Optimization

A key element of FinOps framesworks: improve governance and cloud cost controls without sacrificing performance.

FASTER CODE = LESS COMPUTER CONSUMED = LOWER CLOUD BILL

Databases and Data Stores

Increase Apache Cassandra throughput and density while maintaining service levels.

Your Big Take-Away

FASTER JAVA = LOWER CLOUD COSTS

30%+ reduction in server footprint millions of dollars in savings

Would be running 5x infrastructure without Azul Platform Prime

Reclaimed 20%+ in CPU carrying capacity, improving gamer experience

85% reduction in cost per transaction with cloud-based finacial system

Improved performance and reduced total infrastructure spend by 38%

Reclaimed 50% of infrastructure used for electronic trading

MAINTAIN PERFORMANCE SLAS & IMPROVE CUSTOMER EXPERIENCE

Learn More

Immediate 95% reduction in trouble tickets, improving operational productivity

Removed 2,190 hours a year in performance tuning, accelerating fraud detection in microseconds at high volumes

Optimized 4 billion cell grids to improve productivity for geo-based simulation modeling

Drove 80% improvement in microsecond performance for trading platform

Delivering instant hotel pricing data from thousands of properties to customers

Enabled faster processing of massive datasets to drive real-time fraud analysis and protection

Looking into the future –
A view from IDC

IT teams are facing the challenges of a tough budget environment and the need for performing applications. Optimized Java virtual machines offer a solution to both.

Jevin Jensen, Research Vice President, Intelligent CloudOps