AWS Elastic Recode

Introduction (中文版本)

Transcode jobs are the most resource intensive workload for media / gaming studio customers, customers can achieve performance, cost and elastic of large scale transcode jobs on AWS.

This is A Media Solution that Maximizes Transcoding Flexibility and Scalability while Creating a Cost-effective Transcoding Option for AWS Customers

The major components include:

1. Web UI : interactive user experience
2. Transcode Docker Images: encapsulate the transcode & VMAF scoring details into docker image, so customer can get started quickly and adopt it according to your own business requirements
3. Large Scale Jobs Schedule & Execution: We take advantage of Kubernates, the opensource container orchestration platform, and AWS to provide large scale cost performance EC2 spot instances. Most important, we design and implement a reference control plane service to help customer manage the large scale transcode jobs.

Here is the current version solution architecture:

Base on the above architecture, the customer experience will be:

• All videos include source videos and output videos will be saved to Amazon S3 automatically
• The solution simplify and standardize the transcode profiles
• Customer can review the estimated dry-run statistic report before actually execution
• Customer can track the jobs status and drill down the detail logs through the realtime websocket tunnel
• The backend compute instances support CPU and GPU platform
• AWS EKS will auto scaled according to the number of submitted jobs

About the Workshop?

Created by AWS Solution Architect team China, this workshop aims to guide customer to experience transcode jobs using docker and how to manage & schedule large scale transcode jobs on AWS EKS.

In this workshop, you will learn:

• What are the inputs and outputs of a trancode job?
• What does VMAF scoring means and how it works?
• Submit and monitor jobs through our easy to use web UI
• Review and understand the estimated dry run results
• Run jobs using scripts to learn more details about the transcode progress
• Why we need a control plane for large scale jobs and how it works?
• Why spot instances matter to you, and how it works with AWS EKS?
• Run large scale jobs on AWS EKS and with our reference control plane service implementation

Security

See CONTRIBUTING for more information.

License

This library is licensed under the MIT-0 License. See the LICENSE file.