DVLA was experiencing massively increased demand for the information, rising from 600,000 to more than 70 million requests per month. This was partly due to the discontinuation of the paper “tax disc” once displayed on UK windscreens to communicate that the owner had paid vehicle-excise duty, requiring the public to rely on DVLA online services to identify the tax status of a vehicle.

The organization was also being contacted by companies wanting to build applications that would consume vehicle-related data. Making the data open and available to appropriate uses had significant potential to simplify processes, reduce cost and risk, and contribute to economic growth—but DVLA would need to handle transaction volumes anticipated to measure in billions per month. “We needed to move to an elastic platform that would allow us to scale up and down based on demand,” says Lewis.

DVLA already had services running on AWS and, based on positive experiences, decided to evaluate Amazon API Gateway as a way to provide controlled data access. “We stood up a working prototype in just a few days using Amazon API Gateway,” says Lewis. “The speed with which we were able to deliver it was unprecedented.” Because Amazon API Gateway provides a complete service architecture, the team didn’t have to deploy servers or install, manage, or maintain software.

“When we deployed on premises, we had fixed capacity and had to procure more servers if we went beyond a certain limit,” says Lewis. “That model changed with the cloud, but even then, we had to provision capacity, manage images, create hardened builds, and so on. Now, using Amazon API Gateway, we’ve moved into a serverless world where the only thing we have to worry about is the code—the thing that creates the greatest value—and everything else is taken care of.” DVLA is now able to provide scalable, secure, programmatic data access to police, local authorities, and third parties building value-added solutions.

DVLA is also experimenting with the AWS Lambda serverless compute service. “We started by identifying some use cases that would make sense, such as generating reports from API Gateway data that get delivered to our management and operational teams,” says Lewis. “We’re very interested in the serverless model because it is only triggered when an event happens and you only pay for it when it runs.”