ATPCO uses Red Hat OpenShift 4 to rapidly provide airfare data

ATPCO, an airline-owned organization, is the world’s primary source for airline retailing and pricing data. The company provides this data to a global network of airlines, data distributors, online travel agents, sales channels, and industry groups. To keep pace with dynamic, complex data needs, ATPCO’s technology teams migrated from a traditional, virtual machine (VM)-based approach to the cloud. After migrating from Red Hat OpenShift 3 to Red Hat OpenShift 4, ATPCO has reduced delivery times for new features and services, built a stable foundation for developer innovation, and improved scalability across major cloud platforms in just a few steps.

Benefits

• Reduced application deployment and service delivery times from months to minutes
• Created flexible foundation for developers to experiment and innovate 
• Achieved greater responsiveness and stability at scale

ATPCO was founded in the 1960s by a group of airlines to collect and distribute airfare information to the travel industry and technology partners, including data distributors, online travel agents, and industry groups. Its database now includes information on more than 289 million active fares for more than 400 airlines in 160 countries.

ATPCO faced increasing demand on its IT infrastructure, from constantly changing air travel pricing data to new fare elements to consider—for example, carbon impact, baggage allowances, seat pitch, and responding to the COVID-19 pandemic. Additionally, the company’s 2018 acquisition of rich content provider Routehappy demonstrated a need for improved scalability and application programming interface (API)-based integration with internal and third-party solutions.

The group initially tried delivering applications as microservices on virtual machines (VMs) and automating some processes to solve these growth-related challenges. However, establishing best practices for maintaining, monitoring, and securing services in this complex environment proved difficult, especially when demand for compute capacity and other resources changed frequently.

To continue to offer dynamic services and accurate information to its customers, the group sought to create a more agile, responsive cloud computing environment.

“We want to get out of the datacenter,” said Chris Blelloch, Head of Technology Platform Services, ATPCO. “By the end of 2022, we want all of our workloads to be running in public or private clouds, on a stable and sustainable platform.”

Building a flexible cloud foundation with Red Hat OpenShift and enterprise Kubernetes
In 2018, ATPCO decided to adopt Red Hat OpenShift as the new foundation for its IT infrastructure, running in Amazon Web Services (AWS). Based on Kubernetes, Red Hat OpenShift is an enterprise container platform for applications in hybrid cloud or multicloud environments. It provides ATPCO with operator-based service orchestration,  service discovery, and load balancing capabilities.

After its initial adoption of the platform, ATPCO migrated to Red Hat OpenShift 4 to expand its cloud environment to any of the leading public or private cloud platforms as desired, in just a few steps. The yearlong migration was done in four phases, each taking less than four hours.

“It was clear from the beginning that the Red Hat OpenShift roadmap would provide long-term support to meet our business and technology demands,” said Blelloch. “Even during our update to version 4, migration was painless and setting up clusters was automatic and easy. We are excited to be an early adopter of Red Hat OpenShift 4.”

The group engaged certified Red Hat partner Sysdig to enhance operational visibility and security for its Red Hat OpenShift deployment by improving collaboration between security, DevOps, and services team members. 

Reduced provisioning and service deployment times from months to minutes

ATPCO’s product teams have used Red Hat OpenShift to improve delivery times for new applications. Its previous VM-based provisioning approach required a month or more to create application environments—a process that now takes only minutes with automation. Additionally, production deployment can be now completed weekly in just 15 minutes using a continuous integration and continuous delivery (CI/CD) approach, compared to the former six-month timeline that required 2-3 hours for deployment.
“We can raise additional capacity easily within minutes. Instead of needing weeks to prepare hardware, run logs, and other tasks, it’s a much simpler and more automated process,” said Veerendra Akula, Platform Architect, ATPCO. “Compared to using VMs and on-premise systems, we can focus on rolling out new features faster to build solutions for customers, instead of configuration tasks that our OpenShift Operators now handle.”

“We can raise additional capacity easily within minutes. Instead of needing weeks to prepare hardware, run logs, and other tasks, it’s a much simpler and more automated process,” said Veerendra Akula, Platform Architect, ATPCO. “Compared to using VMs and on-premise systems, we can focus on rolling out new features faster to build solutions for customers, instead of configuration tasks that our OpenShift Operators now handle.”

Enhanced developer flexibility and freedom to innovate 

Development teams’ processes have also become more efficient with the shift from VM-based to container-based infrastructure. Developers can now build several experimental products to see what would work best for customers, without adding any resources, cost, or waste due to automated scalability. Operators and components, such as Red Hat OpenShift Serverless and Red Hat OpenShift Service Mesh, support automated scalability and a flexible, microservices-based approach.
ATPCO’s developers can also use their choice of languages, frameworks, databases, and public cloud vendors to deliver innovation that aligns with customer demand. “For example, we’re introducing a visual breakdown of the amenities built into airline pricing, such as whether or not a higher fee includes free bag check and extra leg room, versus a less expensive ticket that does not include those amenities,” said Blelloch.

Improved response to changing demand with greater stability at scale

Airline workloads are typically unpredictable, due to changing flight schedules, unexpected maintenance, weather conditions, and more. Previously, ATPCO had to prepare its infrastructure to support the maximum potential traffic in case of sudden increase in demand. However, monthly reports showed that CPU resources were severely underused at less than 5%.

With the enhanced scalability included in Red Hat OpenShift 4, ATPCO’s team can add or remove cloud capacity to its container environment to support demand in milliseconds. Now, CPU resources are used at close to 70% and cluster memory at 60%. In addition, its continuous deployment approach means ATPCO can use OpenShift operators to deliver updates, fixes, and new applications to production during business hours while drastically reducing downtime.

After success with Red Hat OpenShift 4, ATPCO is now focusing on increasing automation and finding technology solutions with more flexible pricing to remove any barriers to continued growth. The group is evaluating Red Hat Advanced Cluster Management for Kubernetes, an end-to-end solution for managing and securing Kubernetes containers across datacenters and public clouds. ATPCO is also testing Red Hat OpenShift Service on AWS, a jointly managed and supported solution that runs Red Hat OpenShift natively on Amazon Web Services (AWS) for easier subscription and pricing management.

“With Red Hat, we can meet our customers’ needs faster with new, high-quality services and features while still achieving our operational goals, such as reducing costs,” said Blelloch. “We also have insight into roadmaps and what future versions may include, and we can participate in the creation of new features that will help ATPCO continue to enhance our services.”