Cloud logo

Responsibilities

Developer, Designer

SpecialVFX@Cloud

1 minute read

Welcome to SpecialVFX@Cloud, a scalable cloud solution for special effects processing, developed as part of an IST Cloud Computing project by Miguel Rocha, Lucas Pinto, and Kenneth Brattli.

Project Description:

SpecialVFX@Cloud is a service hosted on Amazon Web Services (AWS), designed to support a hypothetical special effects studio by executing computationally-intensive tasks. The main functionalities include:

  • RenderImage: Generate photo-realistic images using ray-tracing.
  • BlurImage: Apply a blurring effect to images.
  • EnhanceImage: Highlight edges in images to enhance details.

Goals:

The primary goals of this project were to:

  1. Implement Elastic Scalability: Ensure the system can automatically scale by increasing or decreasing the number of workers based on the number and complexity of user requests.
  2. Optimize Cost Efficiency: Balance between cost and performance using a mix of EC2 instances and Lambda functions.
  3. Collect Performance Metrics: Instrument the application to measure the complexity of requests, aiding in load balancing and auto-scaling decisions.
  4. Utilize Cloud Environment: Deploy the system on Amazon Web Services (AWS), including EC2 instances, Lambda functions, and DynamoDB for metrics storage.
  5. Automate Deployments: Automate all cloud deployments and ensure resources are efficiently managed and deleted after use.

Overall Workflow:

  1. Load Balancer (LB): The entry point of the system that receives all web requests and forwards them to either an EC2 instance or a Lambda function based on estimated request complexity.
  2. Workers: Execute the requested operations. These are of two types:
    • VM Workers (EC2 Instances): Handle requests directly and can be instrumented to collect performance metrics.
    • FaaS Workers (Lambda Functions): Handle requests with fast startup times but higher cost per request.
  3. Auto-Scaler (AS): Monitors system performance and adjusts the number of active EC2 instances to balance load and cost.
  4. Metrics Storage System (MSS): Uses Amazon DynamoDB to store performance metrics, aiding in decision-making for load balancing and auto-scaling.

Milestones:

  • May 17th: Project Checkpoint (Optional): Submit your initial implementation and receive feedback on your progress.
  • May 31st: Project Submission: Submit your fully functional system along with a final report.

Documentation:

  • Check the project description here;
  • The final report can be found here.

Useful links:

  • Github Source code here;
  • Git-Lab Source code here.

Updated: