World Fuel Services Migrates 20 Data Centers to AWS in 2 Years, Cuts Costs by 50%

By 2018, World Fuel Services (WFS), a global fuel provider, had accumulated 22 data centers through business acquisitions, many running Windows and Microsoft SQL Server workloads. The high cost of operating these autonomous legacy systems pointed to a need to modernize technology, consolidate data centers, and reduce costs.

WFS launched an initiative to migrate all workloads to the cloud, choosing Amazon Web Services (AWS) because of its breadth of services and global footprint. WFS began with the AWS Migration Acceleration Program (AWS MAP) for Windows to access its tools, partner ecosystem, and 11 years of experience helping organizations migrate to the cloud. To help offset the cost of migration, WFS created a cloud economics team and engaged the expertise of Kinect Consulting, an AWS Advanced Consulting Partner. As a result, WFS migrated 20 data centers to AWS in only 2 years and realized a seven-figure cost reduction in 2019.

WFS provides energy procurement advisory services, supply fulfillment, and transaction and payment management solutions to the aviation, marine, and land transportation industries. Since its founding in 1984, WFS has grown through organic expansion and strategic acquisitions. The operation of independent systems for 22 data centers consumed many staff resources and made it nearly impossible to implement a comprehensive business strategy. WFS needed to consolidate its data centers to optimize costs and staff efficiency and to achieve scalability for growth.

In 2018 WFS evaluated multiple cloud providers, choosing AWS for its global footprint, its security and compliance capabilities, and its $22.6 billion investment in research and development to improve its services. “The worldwide locations of AWS enabled us to have regional compute distribution and enabled team members to tap partnerships established in previous roles,” says Avi Boru, director of cloud engineering at WFS. “Our leadership and engineering teams both wanted to work with AWS, so we agreed from bottom-up to top-down.”

IT and engineering staff throughout the company wanted cloud expertise, so WFS launched an initiative to introduce the team to the AWS approach. The company planned for 85 percent of its IT staff to become an AWS Certified Cloud Practitioner or higher by the end of 2021. The goal was to transform traditional engineers into cloud engineering teams focused on scripting and infrastructure as code. WFS also used AWS Training and Certification to flatten the learning curve of recent college graduates hired for the cloud service engineering team. “We chose new talent with fresh perspectives for our cloud economics team—entry-level engineers who had never touched AWS,” says Jason Astacio, vice president of cloud infrastructure at WFS. “We showed them the outcomes we were trying to achieve, and they started from scratch, using the success patterns AWS offered to cut run rates and costs.”

The cloud economics team identified migration bubbles—redundant infrastructure in the data centers and the cloud—that would triple costs. The cloud economics program focused on cost management strategies and used AWS Training and Certification resources to help mitigate the migration bubble. To migrate quickly, WFS used AWS MAP and AWS Application Migration Service (CloudEndure Migration), an automated lift-and-shift solution that simplifies, expedites, and reduces the cost of cloud migration. “AWS Application Migration Service helped us migrate faster and enabled us to release the overall program smoothly,” says Astacio.

The company began the migration with its application portfolio, migrating 10 applications in the first quarter, 40 in the next quarter, and 60 in the quarter after that. By the end of 2020, WFS was migrating 100 apps per quarter. WFS then began using Amazon Elastic Container Service (Amazon ECS), a fully managed container orchestration service. AWS Lambda, which enables WFS to run code without provisioning or managing servers, helped remove the staff’s responsibility for infrastructure maintenance. “AWS Lambda has always been our go-to because wherever there’s not an applicable AWS service available, we just write our code and make it an extension of our existing service,” says Boru. The company created and deployed custom resources to use alongside Amazon CloudFront, a fast content delivery network service.

The cloud economics team monitored processes throughout the migration, becoming a product team focused on cost showbacks and chargebacks. The team gained spending insights using tools like Amazon QuickSight, a fully managed business intelligence service that enables companies to create and share interactive dashboards. As a result, WFS cut monthly costs by 50 percent and avoided an additional increase in expenditures in 2020. “Amazon QuickSight helps provide context on our global costs across all regions and AWS accounts,” explains Boru. “Our usage is in place, but our costs aren’t spiking.”

WFS is assessing investments in renewable energy for possible future revenue. The company analyzes weather patterns and market predictions to decide whether to hedge or invest in wind farms. For this initiative, the company chose Amazon SageMaker, a fully managed service that helps developers build, train, and deploy machine learning models.

WFS is currently running an Oracle E-Business Suite solution for enterprise resource planning but plans to migrate this solution and surrounding applications to AWS. WFS is also looking for consolidation opportunities on Amazon Elastic Compute Cloud (Amazon EC2), which provides resizable compute capacity. WFS hopes to convert its Amazon EC2 instances for containers running on Amazon ECS and plans to tackle its two remaining data centers in 2021. “We’re not just consuming one service, but the whole fleet,” says Astacio. “It’s important for us to know that the services work together and that AWS continues to innovate its service offerings.”