DynamoWave Chat - A Serverless, Real-time Chat Application

DynamoWave Chat is a modern, scalable, serverless real-time chat application built using AWS Lambda, DynamoDB, and API Gateway.

➡️ Core Focus: Enhancing scalability, performance, and security from a non-functional standpoint.

---

System Architecture & Components

---

Workflow Overview

1. Establish WebSocket connection: Two-way communication between the client and server.
2. ConnectHandler Lambda: Triggered when a connection is established, inserting the connectionId into ConnectionsTable.
3. Notification to Client: Once the connection is established, the client is notified.
4. SendMessageHandler Lambda: Iterates through connectionIds and sends messages to connected clients.
5. DisconnectHandler Lambda: Cleans up by removing inactive connectionIds after the session ends.

---

Services & Purpose

Service	Identifier	Purpose
API Gateway	WebSocket API	Enables real-time communication.
DynamoDB	ConnectionsTable	Tracks and manages active connections.
AWS Lambda	ConnectHandler	Records new connections for operational monitoring.
	DisconnectHandler	Removes inactive connections from the registry.
	SendMessageHandler	Handles reliable communication among connected clients.
	DefaultHandler	Notifies clients when the connection has been established.

---

Design Considerations

Availability & Reliability Improvements

1. Reserved Concurrency for Critical Lambdas: Critical Lambda functions have reserved concurrency quotas to ensure compute availability during peak times and prevent throttling.

   ➡️ Allocating resources ensures the application remains functional even during high load periods.

2. Data Durability via Point-In-Time Recovery (PITR): Enabled PITR for DynamoDB to restore data to any second in the past 35 days, ensuring data availability and fault tolerance, even in case of accidental overwrites or deletions.

   ➡️ Simplifies data recovery without operational overhead or over-provisioning costs.

3. Backpressure Resilience: Implemented API Gateway throttling and rate limiting to ensure backend services aren't overwhelmed during peak usage.

   ➡️ Defines a maximum threshold of incoming requests and caps client requests to avoid service downtime or DDoS attacks.

4. Resilience to Zonal Failures: While the application is resilient to zonal outages, to improve availability in production systems, DynamoDB Global Tables and regional redundancy with Route 53 DNS failover could be implemented.

   ➡️ Provides higher availability for mission-critical applications and geographically distributed user bases.

---

Code Optimizations for Lambda Reliability

1. Error Handling: Incorporated error handling mechanisms in Lambda functions to prevent cascading failures and ensure system stability.

2. Retry Mechanisms: Implemented exponential backoff for critical functions to recover from transient errors such as network failures or database operations.

   ➡️ Increases the likelihood of successful message delivery without overwhelming the system.

3. Graceful Error Recovery: Configured a Dead Letter Queue (DLQ) to capture and reprocess failed messages, ensuring zero data loss.

---

Performance Optimizations

1. Dynamic Auto-scaling for DynamoDB: Dynamically adjusts capacity based on fluctuating workloads to ensure cost-effective scalability.

   ➡️ Automatically scales based on demand, reducing costs during idle times and increasing capacity during peak times.

2. Custom Lambda Warmer: Implemented a Lambda warmer function to reduce cold starts and improve performance for sporadically used functions.

   ➡️ Configured a CloudWatch event to trigger the warmer function, maintaining low-latency performance.

---

Security Features

1. API Gateway Resource Policies: Enforced HTTPS-only requests to ensure secure transport for all connections.

2. KMS Encryption for DynamoDB: Secures all data at rest with KMS encryption, protecting sensitive user information.

3. Least Privilege IAM Roles: Pruned down IAM policies for Lambda service roles to enforce least-privilege access, minimizing risks of privilege escalation.

4. Throttling for DDoS Mitigation: API throttling prevents potential DDoS attacks by limiting the number of requests a user or bot can send in a given time.

---

Future Enhancements

1. WAF Integration: Adding a Web Application Firewall (WAF) on top of API Gateway to protect against excessive resource consumption and ensure application availability.

   ➡️ Managed and custom rules will prevent potential security threats.

2. CloudWatch Alarms & SNS Integration: Configuring real-time alerts for abnormal API usage patterns or security incidents via CloudWatch alarms and SNS notifications.

   ➡️ Improves monitoring and operational insights.

---

Contributions

We welcome suggestions to further improve the architecture or performance of DynamoWave Chat. Feel free to contact me at tanishka.marrott@gmail.com.

---

Credit Attribution

Special thanks to AWS for providing the foundational architecture guidelines for this project: AWS WebSocket API Chat App Guide.