Implementation of a scalable cloud renderer. The system consists of three independently scalable primary components as outlined below.
For an overview and demo of the system feel free to refer to the YouTube video Building a Cloud Renderer with Kubernetes and AWS. The video walks through the architecture and shows the system in action.
React based interface where the user can create and schedule new rendering jobs. The system is simplified in that it renders a single scene, but the user can set the color of the primary sphere. The AWS Amplify library is used to interface with Cognito(Auth) and S3(retrieved finished images).
Nodejs express server operating as the scheduler/manager that communicates with the frontend, renderer, and DynamoDB table for job tracking. The scheduler determines whether or not to start a rendering job or place it in the queue. The system will automatically process backlogged jobs in a FIFO order.
Ray tracer based on Ray Tracing in One Weekend. Implementation in Go extended from Hunter Loftis' repo. Ray tracer was wrapped with a web api and extended to communicate with Amazon S3 and work with in memory .PNG files instead of file system .PPM.
- Extend the system to except a more complex scene definition
- Integrate a WebGL view in the UI to see a preview of the scene
- Extend system to see job progress in the UI