The world continues to be driven by data, and all industries are awakening to one reality: optimizing IT infrastructures for peak performance and cost-efficiency. Airlines like American Airlines are no strangers to facing this constant challenge.

Containerization technology, like Kubernetes, is created for just this purpose – allowing optimization and efficiency at scale.

Intel has been a strategic technology partner to American Airlines for the past 25 years. Consequently, Intel’s suite of enterprise software solutions, specifically Intel Tiber App-Level Optimization, have played a key role in helping American Airlines leverage Kubernetes to achieve new heights of scalability.

While we have previously covered American Airlines’ success with Intel Tiber App-Level Optimization on Databricks optimization, this piece dives into the Kubernetes optimization journey discussed at KubeCon Paris.

Taking Flight with Kubernetes

Kubernetes is an open-source system for automating containerized application deployment, scaling, and management.  For airlines like American Airlines, Kubernetes offers a way to streamline complex IT operations and ensure applications run smoothly without disrupting their complex ecosystem and tech stack.

However, achieving optimal performance within a Kubernetes environment requires careful consideration of hardware and software.

This is where Intel’s suite is a pivotal solution for most companies with these struggles.  

But why is Intel talking about application performance when they’ve made a considerable name for themselves as chip-makers?

Over the years, we’ve dramatically increased our investment in software development and cloud solutions. At the end of the day, what matters is the value that you’re getting out of our hardware/what matters is how much application performance you are getting out of that.

Markus Flierl, Corporate VP at Intel Software.

According to Gartner, 85% of AI and machine learning projects fail to produce a return for the business. 

Meanwhile, investment in public cloud spending is estimated to exceed $1 trillion (yes, trillion with a “T”) by 2027.

With these trends in full view, Intel is hard at work with cloud providers to offer the latest and greatest software stacks available on the market seamlessly integrating with existing ecosystems. With their strategic acquisition of Intel Tiber App-Level Optimization in early 2022, Intel Software’s suite of tools has bolstered best-in-class app performance optimizations that reduce cloud costs.

Through services such as Kubernetes and container optimizations, Intel offers these tools through Intel Tiber App-Level Optimization as a way to make these groundbreaking solutions available to all of their clients years before the “next generation of open source.”

The 25 Year Partnership Between Intel and American Airlines

American Airlines has been a long-time technology partner, and their recent foray into Kubernetes optimization is a testament to their commitment to innovation.

Vijay joined Markus on stage to offer his review of Intel, specifically Intel Tiber App-Level Optimization’s Kubernetes optimization, and how it has positively affected American Airlines as a whole. He begins by underlining the strength of the partnership with Intel and the solutions they have provided over the years while also building a community within the IT community based on trust, security, and resilience.

With an unpredictable market like the travel industry, American Airlines had to adapt. They saw trends in blended travel (remote workers traveling for BOTH business and leisure) and adopted more premium seating, with options to upgrade plans once purchased. Of course, these changes would have otherwise strained their current workloads and resources on their IT infrastructure for coordination and management.

We heavily rely on amazing partnerships like Intel to optimize our application runtime, and we have been doing this for the last 20 plus years.

Vijay Premkumar, Sr. Manager Cloud Innovation at American Airlines

Focus and Goals for Kubernetes Applications

American Airlines constantly strives to optimize its IT infrastructure for peak performance and cost-efficiency. Managing a complex network of applications and data demands a modern approach. 

While cloud adoption offered scalability and flexibility, American Airlines faced challenges in optimizing their cloud platform for containerized workloads running on Kubernetes. 

Their key objectives were to:

• Promote engineering excellence through streamlined application management.
• Improve operational resilience by ensuring application performance and stability.
• Strengthen collaboration and automation for efficient IT operations.
• Maximize cloud efficiency to optimize resource utilization and reduce costs.

The next step was to set clear priorities that would provide a secure, resilient and cost effective self-service Kubernetes platform for application teams. Specifically, the team set KPIs for 99.99% availability SLA for production, 3x shared cluster capacity, and 100% self-service accessibility for application teams to seamlessly integrate.

The American Airlines tech stack is varied and unique, so any 3rd party tools are required to be complementary to existing technologies and initiatives. In order to achieve their focus of cost optimization and goals to promote engineering excellence, improve operational resilience and maximize cloud efficiency, any addition to the Kubernetes tech stack was required to fit in cleanly with existing tools including Rancher, Argo, Kuma, Runway, and their own custom in-house operators.

Searching for a Kubernetes Optimization Solution

Given the strength of the partnership, American Airlines turned to Intel for a solution that could address these challenges within their Kubernetes environment. The answer came in the form of Intel Tiber App-Level Optimization, an autonomous optimization service that continuously analyzes and optimizes resource allocation for virtual machines or container workloads. 

We started the implementation (non prod) about 12-14 months ago, and today it’s 100% in our production, about 1,300 plus clusters have been successfully implemented and 37% fewer resources. This results in about 23% cost savings in our Data Lake solution. It has been a fantastic journey.

Vijay Premkumar, Sr. Manager Cloud Innovation at American Airlines

Four key reasons for deciding to optimize their Kubernetes with Intel Tiber App-Level Optimization were highlighted:

• Ease of deployment: It was a simple integration process because Intel Tiber App-Level Optimization was already deployed in their Data Lakes job cluster. Additional activations on Kubernetes clusters required minimal heavy lifting, zero code changes and provided quick time to value.
• Complementary to existing tech stack: As outlined above, American Airlines’ Kubernetes stack is complex and includes a variety of tools. Intel Tiber App-Level Optimization was able to plug in without any disturbance to the ecosystem and provided added value on top of existing tools.
• Multiple points of impact: The technology packs a one-two punch of runtime optimization improving performance on the application level and capacity optimization autonomously rightsizing pod resources.
• Strengthened collaboration and automation: With Intel Tiber App-Level Optimization’s autonomous capabilities, developers would be free to work on new features and innovations, instead of manual configurations. Continuous and autonomous solutions also enable their dynamic Kubernetes environment to increase velocity at scale without disruptions.

These characteristics translated well into American Airlines’ goals with Kubernetes, one of the airline’s fastest-growing environments. 

Achieving New Heights With Kubernetes Optimization

Over the last few months, since activating Intel Tiber App-Level Optimization in their non-production cluster, American Airlines has seen the following outcomes:

• Between 40-45% cost savings
• 30% job time reduction
• 20% throughput increase
• Achieved SLAs

They were also able to see impact in terms of visibility and customization of clusters. The Kubernetes team was able to pick and choose the configuration that was right for their unique infrastructure.

A Look Ahead: Clear Skies for Continued Success

The success story of American Airlines and Intel with Kubernetes optimization paves the way for a future of efficient and high-performing IT infrastructure in not just the airline industry but all sectors that require optimization to their tech stack without interference.

As we look ahead, Markus and the team at Intel will continue to focus on these critical areas for the future:

• Data Lake Optimization: Intel Tiber App-Level Optimization optimizes resource allocation within your data lake, ensuring smooth operation for data analytics workloads.
• Kubernetes & Containers Optimization: Specifically designed for containerized environments, Intel Tiber App-Level Optimization optimizes resource allocation within Kubernetes clusters, eliminating over-provisioning and wasted spending.
• Runtime Optimization: Intel Tiber App-Level Optimization continuously analyzes resource needs at runtime, dynamically scaling and prioritizing workloads for optimal performance. This includes:

This is only the beginning of the journey. Going forward, a lot of what we’re doing across all these different areas and thousands of engineers and optimizing the different parts of the stack will be through [Intel] Intel Tiber App-Level Optimization, and so we expect that over time that some of these optimizations that we can do through granular will actually increase in magnitude.

Markus Flierl, Corporate VP at Intel Software