jeffreywangzhi/aws_multithread-statemachine

A fast, scalable, multithreaded download statemachine based on AWS, supporting .mkv, .mov, .mp4, .jpg, and others.

AWS_Multithread-Statemachine

A fast, scalable, multithreaded download statemachine based on AWS.

1. 10+ respective threads (lambda functions) download files from target file URL.
2. 10+ threads upload each chunk to the target S3 bucket.
3. Chunk merged and job finished.

• System Efficiency: 300+ GB/ minute
• Scalability: scalable with additional threads (AWS Lambda Functions)

Prerequisites

• Install AWS CLI : installation guide
• Install AWS SAM CLI : installation guide
• AWS Account Credentials : How to guide

CloudFormation Template Configuration

• Replace each aws-account in template.yaml with your own AWS account number. (Ctrl + F, aws-account)
• Config each Role in StepFunctions and Lambda Functions and grant corresponding permissions. (Ctrl + F, Role)

Building

• Build and deploy the sam application to AWS cloud.

$ sam build

$ sam deploy --guided

• Enter your target S3 name for storing the downloaded file.

$ Parameter TargetS3 [targetS3]: your-target-storage-s3-name

Usage & Input Format

{
  "src": [
    {
      "filePath": "https://download-file-1.ext"
    },
    {
      "filePath": "https://download-file-2.ext"
    },
    {
      "filePath": "{other-file-paths.ext}"
    }
  ]
}

Scalability

With the above solution, the system efficiency achieved 300+ GB per minute. The scalable system could also be further refined by thresholding the lambda functions (threads).

Who do I talk to

• Jeffrey Wang ([email protected])