A+E Networks Migrates to Serverless AWS Databases to Create Cloud-Native Apps

A+E Networks (A+E) is an innovative global media content company operating across multiple platforms in the digital landscape and home to 10 linear channel brands—including A&E, HISTORY, and Lifetime. It has an audience of over 500 million viewers in more than 200 global territories.

With an eye always on evolving, the company used Amazon Web Services (AWS) to continue to advance its linear experience and architecture, identifying that going serverless could facilitate expansion and further aid the company’s future growth while saving money and maintenance time.

A+E’s legacy on-premises database servers came with hefty overhead and required too much engineering effort to effectively manage. The company decided to pursue a service-oriented, microservices-driven, cloud-native architecture on AWS.

Enlisting the help of AWS Partner Virtusa, A+E lifted and shifted its 38 applications onto AWS—with two of those applications being refactored to use a service-oriented architecture that runs on serverless technology, including the purpose-built AWS databases Amazon Aurora and Amazon DynamoDB. The company is moving forward by building new applications on serverless only. 

As A+E expanded its reach both nationally and globally with international offices, it looked for a scalable new approach to its architecture. “Our engineering team acted like more of an operational team,” says Kevin Keeler, principal of cloud innovation and architecture at A+E. “The team members were focused on maintaining, patching, and adding new volumes at really just operational states.” The company knew it needed to migrate its applications to serverless, purpose-built databases for specific use cases that would enable it to create high-performance, scalable, and more functional applications faster. Such use cases would also help A+E use the right service for the right workloads instead of laboring under a legacy catch-all database solution on Oracle or SQL Server. “We knew the cloud was the future in terms of automation, speed to market, and the reduction of our operational workload,” says Jeffrey McGrath, vice president of data analytics and DevOps at A+E.

To migrate its applications to AWS, A+E first divided its dozens of on-premises applications into smaller groups to help organize the move. Then it did a systematic lift and shift of those  groups of applications to Amazon Elastic Compute Cloud (Amazon EC2), a web service that provides secure, resizable compute capacity in the cloud. Next, the apps were guided through a 40-point checklist of migration processes in 3- to 4-month sprints, going to different databases depending on the needs of the particular software. To optimize costs, A+E reengineered two apps from a monolithic architecture using Oracle to a microservices driven architecture using Amazon Aurora Serverless. When the network chose to reengineer an app, that decision prompted another offshoot sprint cycle to facilitate the change. “Aurora enabled us to spin up simultaneous environments to speed up development cycles,” says Keeler.

Moving forward, A+E has made going fully serverless and adopting a cloud-native, microservices-driven architecture a top priority and will continue to refactor its legacy applications, according to McGrath: “We ask, ‘Can we make the application more cloud native over time? Can we take a certain feature and make that serverless and microservices based and do that one feature at a time?’ It’s like paying our technical debt off every month as opposed to this massive rewrite of applications.”

In addition to running rearchitected apps on Aurora Serverless and Amazon DynamoDB, a fully managed database for internet-scale applications, A+E is building new cloud-native apps that run on AWS databases. One such app, Access, is primarily an internal system that indexes all A+E programs, including supporting materials such as cast and crew, promos, and credit documents. A+E’s programming and sales departments use Access to put together and sell shows. Access also serves as a sort of playground in which engineers can experiment with new services, such as natural language processing and artificial intelligence–powered video recognition. 

A+E decided to build Access on the serverless Amazon DynamoDB, with search results primarily served up through Amazon ElastiCache, which works as an in-memory data store and cache supporting submillisecond response times. The company also designed the app to run using AWS Lambda, which lets users run code without provisioning or managing servers, and A+E pays only for the compute time consumed. “We always joke around that it’s cheaper than getting a latte to run Access for a day using Amazon DynamoDB if you exclude video,” says McGrath. Time to market is always a concern for A+E, but in using an AWS serverless database and compute services, the company was able to get Access running quickly. “It took 6 months from the moment we got Virtusa involved to the moment we got our first release up,” says McGrath.

Operating in the cloud means A+E can execute disaster-recovery tests for apps like Access much more quickly, and when it releases an app, it takes minutes rather than hours. The idea of trying to build an app like Access on premises was absurd to the A+E team. “Now that we know better, we would never consider it,” says McGrath. In fact, without the serverless option, a lot of what A+E has been able to achieve since its move to the cloud would have been either impossible or completely cost prohibitive. 

Purpose-built databases from AWS deliver the agility and elasticity A+E needed to modernize while drastically reducing database overhead costs. At the same time, A+E’s engineers have more time to dedicate to business problems and tasks that move the company forward. “The cloud provided us around zero downtime, a multiregion base, disaster-recovery capabilities, and an immutable infrastructure,” says Keeler. Another benefit of AWS is that A+E can be anywhere it needs to be and no longer has to think about proximity to physical data centers. In the future, A+E intends to maximize the way it uses services like Aurora, AWS Lambda, and Amazon DynamoDB to shift even more of its maintenance burden onto AWS, effectively eliminating the time the company had previously spent simply keeping things running. “We’d rather just be in the business of coding and building business functionality,” says McGrath.

On AWS, A+E is prepared for an even more innovative future with its content available to viewers whenever they want and wherever they are.

All trademarks are the property of their respective owners.