Question 1

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

Accepted Answer

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.

Question 2

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

Accepted Answer

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.

Question 3

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

Accepted Answer

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.

Question 4

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

Accepted Answer

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

Question 5

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

Accepted Answer

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.

Question 6

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

Accepted Answer

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.

Question 7

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

Accepted Answer

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.).

A Guide to Selecting a Performance Java Runtime

Cloud Cost Optimization for Java Workloads

Which Java runtime is the best fit for you?

Vanilla implementation of the Java Platform Standard Edition

Built on OpenJDK and enchanced for faster warmup and peak performance

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

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

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

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

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

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

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

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

Streaming and Events

Search and Logging

Containers and Container Orchestration

Cost Optimization

Databases and Data Stores

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

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

Optimizing Java Application Performance

Getting Started – Three Simple Steps:

1. Estimated the size of your Java estate and top use cases

2. Confirm the technologies you use

3. Choose based on your business

Ready? So are we.