QuadX Uses Serverless Computing to Build Robust, Complex Systems

Although QuadX initially launched its business on Microsoft Azure, it migrated to Amazon Web Services (AWS) after a year. Engineers found that the AWS platform was more user-friendly and came with the right services to support the startup’s mission to “make the world both a marketplace and a playground.” According to Jojo Anonuevo, chief innovation officer at QuadX, “AWS provided us with the foundation to build our core platform, which enabled our business to scale beyond 10 times its growth.”

In 2018, QuadX began exploring serverless architecture to reduce computing costs, roll out services faster, and simplify scaling. An engineer from the QuadX team stepped up to become a serverless champion and self-trained on AWS Lambda. The team began experimenting with small pieces of existing microservices such as one-time password generation and label creation for its merchants booking dozens to hundreds of orders each day. The labeling use case became a proof point for the power of serverless. The team was able to convert print jobs that used to take several hours per day into smaller batch jobs that run in shorter time frames, thus saving costs.

By bundling Lambda on the backend with Amazon API Gateway on the front end, engineers are able to create and maintain APIs at any scale. The startup uses Amazon Simple Notification Service (Amazon SNS) and Amazon Simple Queue Service (Amazon SQS) in conjunction with Lambda for asynchronous event-driven processing, and Amazon Rekognition for optical character recognition to scan and validate shipping documents. “Beyond Lambda, we’re experimenting with AWS Step Functions for orchestrating complex workflows, along with other managed services,” adds Hamilton Chua, serverless technology engineer at QuadX.

Today QuadX is committed to a serverless-first strategy. All new services utilize this approach, and existing ones are being migrated over time. “Lambda promotes a new way of thinking,” Anonuevo explains. “We start to think of the critical capabilities we need to deliver first—things we can roll out in the serverless architecture to validate pieces of our design before spending months developing a new system.” Though the startup was founded on an agile, microservices approach to architecture, the introduction of serverless has ushered in what Anonuevo deems the next iteration of infrastructure: nanoservices with serverless. “Microservices need to be broken down further, and that’s our realization with this whole process,” he says.

Because teams can now unbundle, test, and enhance smaller portions of a microservice without interrupting the functionality of the entire service, features are more robust. Isolating service capabilities has become much easier with serverless technology. “It’s a combination of agile methodology and the granularity that serverless brings that has gotten our team to think more in terms of Lambda functions and nanoservices right off the bat,” Anonuevo explains. “We are seeing the wins in improved functionality from new builds, as we get feedback right away as opposed to waiting to build the entire microservice.” This encourages continuous innovation, as teams can easily try out new ideas.

Improved agility is not the only benefit of switching to serverless. Since transitioning key microservices, QuadX is spending just $600 on infrastructure each month by using Lambda and API Gateway, instead of $8,000 with traditional builds on Amazon Elastic Compute Cloud (Amazon EC2). “The team has also transitioned to a mobile-first strategy for their applications, an approach that tends to develop naturally with serverless architecture,” Anonuevo explains.

The impact of such changes on developers and engineers has been notable. Developers trained in the PHP programming language are now learning JavaScript and expanding their skill sets in other ways to become more well-rounded. QuadX has always maintained a dedicated DevOps team, but the DevOps methodology is spreading throughout the business to other areas, with staff viewing each task in terms of its smaller, “nano,” parts. “Our engineers are definitely thinking differently about how to build and code things with serverless in a quicker way,” Anonuevo adds.

Going forward, QuadX is looking to integrate more automation and machine learning into its operations and is experimenting with Amazon Lex for setting up chat bots. The company aims to serve as a leader in the domestic startup community and seeks to continually innovate as new cloud services become available. “We love the introduction of new technology, so we are the first to try any new AWS features we read or hear about,” Anonuevo says.

Furthermore, Lambda and serverless are now a building block for new services. “As we incorporate serverless technology on AWS, it has put us on the right track for building complex systems more efficiently,” Anonuevo concludes.