Foundation migrated video processing to AWS
A Foundation case study shows a migration to AWS using MediaConvert for video processing, Kafka queues and PostgreSQL to scale thumbnail/GIF generation and CDN delivery across regions described. The stack emphasizes decoupled pipelines and managed media services to reduce ops burden while handling bursty newsroom workloads.
The AWS "Video on Demand on AWS Foundation" reference implementation deploys a CloudFormation template that provisions AWS Elemental MediaConvert, Amazon S3, Amazon CloudFront and AWS Lambda as a complete VOD workflow. docs.aws.amazon.com AWS Elemental MediaConvert bills by "normalized minutes" and added a volume-based tiered pricing model on August 7, 2025 to reduce per-minute costs as output minutes scale. aws.amazon.com For predictable baseline capacity the service offers reserved transcode slots (RTS) with a 12‑month commitment and fixed monthly billing, while on‑demand queues remain the scalable option for burst processing. docs.aws.amazon.com Amazon Managed Streaming for Apache Kafka (MSK) is available as a fully managed Kafka service for buffering and decoupling ingest streams, a pattern AWS and industry writeups cite for absorbing newsroom spikes and smoothing downstream MediaConvert jobs. docs.aws.amazon.com AWS hosts PostgreSQL options through Amazon RDS for PostgreSQL and the PostgreSQL‑compatible Amazon Aurora, with Aurora advertising up to ~3x throughput versus stock PostgreSQL in benchmarked scenarios. aws.amazon.com The Foundation pattern pairs S3 storage with Amazon CloudFront for global edge delivery, where regional CloudFront egress pricing ranges (examples) roughly from $0.02–$0.14 per GB depending on region and tier. docs.aws.amazon.com AWS case examples show hybrid choices—reserved MediaConvert slots for steady load plus on‑demand capacity for bursts—and one customer (Motive) reported an ~85% cost reduction after migrating to MediaConvert as part of their architecture refactor. docs.aws.amazon.com
