Project Description / Goal

The primary objective of this project is to create a serverless application that converts text-based blog posts into audio files (MP3 format) using AWS services. The application supports creating and retrieving posts through a RESTful API while ensuring scalability, minimal operational overhead, and cost-efficiency.

The application workflow includes:

• Accepting new blog posts via a web interface.
• Converting the post text into audio using Amazon Polly.
• Storing the audio in an Amazon S3 bucket.
• Retrieving post information, including links to the audio files, via the API.

Technology Stack

• AWS Lambda: Serverless compute for handling backend logic.
• Amazon DynamoDB: NoSQL database to store post metadata and URLs.
• Amazon API Gateway: RESTful API to expose the application's functionality.
• Amazon Simple Notification Service (SNS): Decouples the process of posting and audio conversion for asynchronous processing.
• Amazon Polly: Text-to-speech service to generate audio files in various languages and voices.
• Amazon S3: Static website hosting and audio file storage.
• Amazon CloudWatch: Monitoring and logging Lambda function performance.

Project Planning / Architecting

The architecture was designed around AWS's serverless offerings to minimize the need for infrastructure management.

Frontend: A static website hosted on Amazon S3 serves as the user interface.

Backend:

• API Gateway routes requests to Lambda functions.
• DynamoDB stores post metadata, including links to generated audio files.
• SNS enables decoupled and asynchronous workflows.
• Polly generates audio, which is stored in an S3 bucket.

Workflow for posting:

1. A new post is submitted via the static webpage.
2. API Gateway triggers the New Post Lambda function.
3. Post data is saved to DynamoDB.
4. SNS notifies the Convert to Audio Lambda function, which processes the text into an audio file.
5. The audio file is uploaded to an S3 bucket, and DynamoDB is updated with the file's URL.

Workflow for retrieval:

1. The user retrieves post details via the API.
2. API Gateway triggers the Get Post Lambda function.
3. Data is fetched from DynamoDB and returned to the user.

Project Journal / History

• Initial Setup: Defined the problem scope, selected AWS services, and created the static S3 website.
• API and DynamoDB Integration: Developed RESTful API endpoints and linked them with DynamoDB for storing post metadata.
• SNS and Lambda Coordination: Integrated SNS to decouple the posting and audio conversion processes.
• Text-to-Audio Conversion: Implemented Polly for generating audio files from text.
• Testing and Deployment: Conducted functional testing of workflows and deployed the application.

Obstacles

• SNS Message Deduplication: Resolved issues with FIFO SNS topics by enabling content-based deduplication.
• Cross-Origin Resource Sharing (CORS): Ensured proper configurations for seamless interaction between S3-hosted frontend and backend API.
• File Storage and Retrieval: Managed S3 bucket policies for public access and secure audio file hosting.
• Error Handling in Lambda Functions: Added robust logging and exception handling to troubleshoot failures during DynamoDB and Polly interactions.
• Integration Testing: Verified asynchronous processes (SNS to Lambda) and ensured data consistency between DynamoDB and S3.