Minnesota Public Radio (MPR) is a media company providing regional services to 46 radio stations across Minnesota.
Modern Web Application Platform with AWS
Minnesota Public Radio's programming division, American Public Media, is one of the largest producers and distributors of public programming in the world. They collaborated with Improving to create a proof of concept to host their news, streaming music, and podcast sites in AWS.
While evaluating an aging active-passive disaster recovery plan, MPR decided to modernize its applications and approach by moving to a cloud infrastructure. This infrastructure would have to provide resiliency to serve their disaster recovery plan, scalability as their audience grows, and nimbleness as the technology driving their apps changes and evolves.
To demonstrate AWS as a viable solution, MPR and Improving decided the public podcast websites and MPR News should be hosted on the new infrastructure. Also, to prove private cloud capabilities, AWS should host several administrative applications. These apps would include a configuration manager, a schedule editor, and an image manager.
To accomplish this, Improving helped MPR build a Kubernetes cluster in EKS. The cluster would allow the applications to automatically scale based on workload and traffic. To host application data, Improving and MPR built a MySQL instance in RDS and an Elasticsearch instance at Elastic.co.
Improving also assisted MPR in setting up other supporting infrastructure, including: Identity Access Management, Virtual Private Clouds (VPCs), Application Load Balancers (ALBs), CloudWatch (logging and dashboards), CloudFront, and Terraform (configuration management).
The Business Benefits
The newer infrastructure enabled major cost savings. The reduction of required hardware meant that fewer servers would need to be purchased for these critical applications. Further, the move to AWS facilitated an easy transition from Akamai CDN to CloudFront. This move saved MPR thousands in annual costs.
Additionally, the new environment eased administrative burdens. As the websites no longer relied on an active-passive disaster recovery model, monthly testing and configuration drift were no longer major issues. The highly available nature of AWS meant that disaster recovery was mostly built-in. Also, configuration management via Terraform and Kubernetes kubectl tools meant changes to the environment could easily be tracked, documented, and rolled back if necessary.
Improving closely aligned with MPR, sending an engineer on-premise to ensure that the project was a best fit for the company. As an AWS Advanced Certified Partner, our collaboration made certain that the project was not only successful but also completed with best practices and future growth in mind.