Carrier Cuts Hours of Processing Time Using AWS Services for Its Electronic Cataloging Tool

The ECAT tool is crucial for Carrier’s sales operations in the commercial HVAC sector. When sales representatives need to evaluate and match equipment for a customer, ECAT lets them move quickly through different tools, calculate and select the appropriate unit, and find key performance characteristics to generate a recommendation about which HVAC family to suggest. A slow ECAT means less productivity for the sales team and decreased profits for Carrier. “Although we used a different provider for a couple of years, it just seemed more difficult to do things, and some things just didn’t work,” says Sotirios Koupas, associate director of software development at Carrier. “We started scaling dramatically, and the previous system wasn’t catching up. The cost kept increasing because we were using more containers. The challenge we faced wasn’t just the scalability of the users on the old system but the scalability of the engines too.”

Carrier had no prior experience using AWS and sought to learn quickly to improve its business and the services it provides. After attending AWS Experience Based Acceleration workshops for modernizing legacy window containers on Amazon Elastic Container Service (Amazon ECS), a fully managed container orchestration service that makes it easy to deploy, manage, and scale containerized applications, the company was ready to start the multiphase migration of all seven existing calculation engines to AWS in addition to building 15 more engines natively on AWS. Transitioning the ECAT from its former cloud services provider to AWS was no small undertaking. “The flexibility that we achieved on AWS is outstanding; we can do things that we couldn’t before. We did not know anything about AWS, so we learned from scratch. AWS made recommendations that tremendously improved the ECAT’s performance. AWS guided us on how to use the AWS services, and we knew our system’s engines and behavior,” says Koupas. “We worked hand in hand until we got it right.” The full migration is a three-phase process with Carrier working alongside experts from AWS, and complete migration is expected by the end of 2022.

The first phase of the ECAT migration finished in August 2021 and focused on using Amazon ECS. Using Amazon ECS, Carrier achieved automatic scaling, making ECAT processors 1.9 times faster and driving performance times down from 6.6 to 3.5 hours. “Our former cloud solution didn’t have automatic scaling. The AWS solution eliminated the need for individual engines, which let us begin scaling to any number that we wanted based on the request. We got a two-times-faster performance,” says Koupas. Carrier also used EC2 Image Builder, which simplifies building, testing, and deploying virtual machine and container images for AWS and on-premises applications. This helped Carrier minimize the footprint of its virtual engines and increase efficiency.

The second phase of migration was modernizing Carrier’s ECAT by decoupling microservices using Amazon Simple Queue Service (Amazon SQS), a fully managed message queuing service that supports microservice scaling and decoupling, distributed systems, serverless applications, and the ability to send, store, and receive messages between software components. With a global messaging queue migrated to Amazon SQS, Carrier could optimize the workload of container instances and increase usage of provisioned resources. This faster performance reduced processing time from the original 6.6 hours to 23 minutes, 16.5 times faster than before. “Moving to a global queue on AWS made it simpler to manage and manipulate task requests and send them to the appropriate container,” says Koupas. “We could reduce failures tremendously.” The company continues to monitor and maintain ECAT processes using Amazon API Gateway, a fully managed service that grants easy access to data, functionality, and business logic to create, publish, maintain, monitor, and secure API at virtually any scale.

This phase involved fine-tuning the scaling metrics and right-sizing instances with help from experts at AWS. Migrating to AWS helped Carrier make the ECAT available at the scale that it needed. As a result, Carrier decreased costs by 50 percent on average. Previously, it would have taken months, with up to $7,000 in costs to run the ECAT. After migration, the cost has been as low as $1,000 per month. “Reducing processing costs wasn’t the only benefit; we also gained the ability to scale services as needed based on our needs and the tool’s performance,” says Koupas. The company has gained the ability to scale the cost up and down based on need and performance, gaining elasticity and control.

Carrier plans to continue using AWS to power more services, tools, and models. “This project is absolutely ongoing,” says Koupas. “We’ve migrated just one engine at this point, and I expect to deploy 20–30 engines by the end of the migration.” Ongoing support helped Carrier successfully migrate ECAT processes from its former cloud provider to AWS, increasing speed and efficiency for field representatives. “I like the approach the AWS team took,” says Koupas. “They handheld the ECAT team until we could walk on our own. The customers are happy, and our systems are running more smoothly since we’ve migrated.”

To learn more, visit https://aws.amazon.com/ecs/.