Tring Image Processing API

A serverless image processing application built on AWS using Terraform. The system allows authenticated users to upload images through a web interface, stores them in Amazon S3 using presigned URLs, processes upload events asynchronously through Amazon SQS and AWS Lambda, stores image metadata in Amazon DynamoDB, and sends processing notifications using Amazon SNS.

---

Architecture

Frontend (HTML / JavaScript)
            │
            ▼
      CloudFront
            │
            ▼
     API Gateway (HTTP API)
            │
            ▼
        API Lambda
            │
            ▼
   Presigned S3 Upload URL
            │
            ▼
         Amazon S3
            │
            ▼
     S3 Event Notification
            │
            ▼
         Amazon SQS
            │
            ▼
    Processor Lambda
            │
     ┌──────┴──────┐
     ▼             ▼
 DynamoDB         SNS

---

Features

Authentication

• Amazon Cognito User Pool
• Cognito Hosted UI
• OAuth 2.0 Authorization Code Flow
• JWT-based API authorization
• User-specific image ownership enforcement

Image Upload

• Secure direct browser uploads using presigned S3 URLs
• No image data passes through API Gateway or Lambda
• Upload progress and preview support

Asynchronous Processing

• S3 ObjectCreated events trigger processing workflow
• SQS decouples storage and processing layers
• Processor Lambda extracts metadata
• DynamoDB updated after successful processing

Metadata Management

• Stores:

  • Image ID
  • User ID
  • Filename
  • Content Type
  • Extension
  • File Size
  • Processing Status
  • Created Timestamp
  • Processed Timestamp

Notifications

• SNS notifications sent when processing completes successfully

Frontend

• Cognito login/logout
• Upload image preview
• Image listing dashboard
• Image metadata viewer
• Full image preview modal

---

AWS Services Used

Service	Purpose
Amazon Cognito	Authentication and user management
API Gateway HTTP API	REST API
AWS Lambda	Business logic and processing
Amazon S3	Image storage
Amazon DynamoDB	Metadata storage
Amazon SQS	Event buffering
Amazon SNS	Notifications
Amazon CloudFront	Frontend hosting
Terraform	Infrastructure as Code

---

API Endpoints

Create Image Upload

Request

POST /images
Authorization: Bearer <token>

Body

{
  "filename": "example.jpg",
  "contentType": "image/jpeg"
}

Response

{
  "imageId": "uuid",
  "uploadUrl": "presigned-url"
}

---

List User Images

Request

GET /images
Authorization: Bearer <token>

Response

{
  "images": [
    {
      "imageId": "uuid",
      "filename": "example.jpg",
      "status": "COMPLETED",
      "createdAt": "2026-06-19T05:57:17.598344+00:00"
    }
  ]
}

---

Get Image Details

Request

GET /images/{imageId}
Authorization: Bearer <token>

Response

{
  "image": {
    "imageId": "uuid",
    "filename": "example.jpg",
    "contentType": "image/jpeg",
    "extension": "jpg",
    "fileSize": 29222,
    "status": "COMPLETED",
    "createdAt": "2026-06-19T05:57:17.598344+00:00",
    "processedAt": "2026-06-19T05:57:30.326139+00:00"
  },
  "downloadUrl": "presigned-url"
}

---

DynamoDB Schema

Table

ImageMetadata

Partition Key

userId

Sort Key

imageId

Local Secondary Index

CreatedAtIndex

Partition Key: userId
Sort Key: createdAt

---

S3 Object Structure

uploads/
└── {userId}/
    └── {imageId}/
        └── {filename}

Example:

uploads/
└── 2478c428-0091-70f3-8115-6ebfb24685e8/
    └── 92e0e3c9-a32c-41de-ac6d-7c67668e89c6/
        └── image.jpg

---

Processing Workflow

1. User authenticates through Cognito.
2. Frontend requests upload URL from API Gateway.
3. API Lambda creates metadata record in DynamoDB.
4. API Lambda returns presigned S3 upload URL.
5. Browser uploads image directly to S3.
6. S3 generates ObjectCreated event.
7. Event is sent to SQS.
8. Processor Lambda consumes SQS message.
9. Lambda retrieves image metadata from S3.
10. DynamoDB record updated with processing details.
11. SNS notification published.
12. Image appears as COMPLETED in frontend.

---

Frontend Features

Upload

• Image selection
• Local image preview
• Upload status indicator

Dashboard

• List uploaded images
• Display processing status
• Refresh image list

Image Viewer

• Full image preview
• Metadata display
• Processing information

Authentication Status

• Logged In indicator
• Logged Out indicator

---

Terraform Deployment

Initialize Terraform:

terraform init

Review changes:

terraform plan

Deploy infrastructure:

terraform apply

Destroy infrastructure:

terraform destroy

---

Project Structure

.
├── frontend/
│   ├── index.html
│   ├── scripts.js
│   ├── styles.css
│   └── config.js
│
├── terraform/
│   ├── main.tf
│   ├── variables.tf
│   ├── outputs.tf
│   └── modules/
│
├── lambda/
│   └── image-api/
│       └── lambda_function.py
│
├── processor/
│   └── lambda_function.py
│
└── README.md

---

Security Considerations

• Direct S3 uploads through presigned URLs
• No public bucket access
• JWT-based authorization
• User ownership validation
• IAM least-privilege permissions
• CloudFront HTTPS delivery
• SQS decoupled processing architecture

---

Future Enhancements

• Image thumbnail generation
• Multiple image formats
• Image transformations
• Metadata search functionality
• User-specific notifications
• CloudWatch dashboards and alarms
• CI/CD pipeline using AWS SAM or GitHub Actions

---

Author

Tarun Harish