Cowboy Migrates to AWS and Gains Scalability with Ease Using CloudNation

Cowboy provides highly connected electric bicycles (ebikes) for urban riders. Founded in 2017, the company now has customers in more than 2,500 towns and cities across Europe, completing more than 1.5 million rides per month. Cowboy’s bikes use smart features and cellular connectivity to provide advanced functions to users. As an anti-theft feature, the bike can alert the owner if it is being moved, jostled, or struck. The bike also incorporates a find-my-bike feature that allows users to locate a stolen bicycle. When a user logs in to the phone-based Cowboy app before a ride, the app also tells them about the bike's charge status.

Cowboy also uses connectivity to improve the safety of riders with automatic crash detection. When sensors detect a crash, the app sends a text message to the rider to check their status. If there is no response, Cowboy alerts the rider’s pre-programmed emergency contact. “I think we've solved a lot of worries—those safety fears, that users might have about buying and riding a bike—with these features,” says Zachary Diebold, lead software architect at Cowboy. “In addition to those safety features, our ebikes offer other capabilities that are attractive to younger riders and that produces a lot of data.”

Cowboy’s largest volume of data came from route tracking—including recorded routes, average speed, and ebike speed at any point. The system generated 25 pieces of data per second for every user. However, the volume of data being generated was growing too large for the company’s infrastructure to manage. “We had maxed out at the top tier of our cloud provider,” says Diebold. “We were going to hit our limit, so we needed a scalable solution.”

Cowboy had launched its infrastructure in the cloud using a provider that offered managed services in various tiers. The tiers didn’t provide the root access to servers that Cowboy wanted and provided limited access to logs. “It was a plug-and-play solution,” says Diebold. “But everything was fixed, with each tier offering certain configurations. We didn’t need a lot of RAM, we didn’t need powerful servers, but we had to pay for the highest tier because we needed the database storage. We had a vision that was bigger than our provider and it was becoming very limiting.”
 

The company approached AWS to learn how it could gain the capabilities it needed. After discussing Cowboy’s current situation and ambitions for the future, an AWS representative put the ebike manufacturer in contact with AWS Partner CloudNation.

“AWS contacted us and explained the situation and, after hearing about the tight timeline, we were ready to help,” says Bart Boonen, co-owner and chief technical officer at CloudNation. “Swift action was crucial. They needed something done quickly or their environment was going to affect their business. We promptly assigned two members of our team to collaborate with their team and immediately got to work.”

Working with CloudNation, Diebold’s team was clear about its needs. The team wanted a database that could scale and could be replicated in a data warehouse application so that it could run analysis on it without affecting performance. Cowboy also wanted deep insights into its data so it could be analyzed to drive better business decisions.

The database was migrated to Amazon Aurora, a relational database management system (RDBMS) built for the cloud with full MySQL and PostgreSQL compatibility. Cowboy’s solution also uses Amazon Simple Storage Service (Amazon S3), an object storage service offering industry-leading scalability, data availability, security, and performance. Elastic Load Balancing distributes network traffic to improve application scalability. The system is secured using Amazon GuardDuty, which combines machine learning (ML) and integrated threat intelligence from AWS and leading third parties to help protect AWS accounts, workloads, and data.

Diebold’s team didn’t just gain the environment it needed, it learned how to develop what it needed. “They build infrastructure as code,” says Diebold. “They were using that to start up machines on a testing space on a staging server and we would then use that to start up on production. We could see exactly what was built and then we were able to run that on the production account. We gained a lot of AWS knowledge from working with CloudNation.”