AWS Certified Solutions Architect - Associate

Explanation

Option D is the correct answer because it provides the best solution for decoupling and scalability requirements.

Analysis of each option:

Option A: Amazon Kinesis Data Analytics

• Kinesis Data Analytics is designed for real-time analytics and processing, not for message distribution to multiple consumers
• Not suitable for fan-out scenarios where dozens of applications need to consume the same messages
• Doesn't provide the decoupling needed between producers and multiple consumers

Option B: EC2 Auto Scaling group

• This only addresses scaling the ingestion application, not the overall architecture
• Doesn't solve the decoupling requirement between the ingestion application and dozens of consumer applications
• Consumer applications would still need to directly connect to the ingestion application
• Doesn't handle the fan-out pattern needed for multiple consumers

Option C: Kinesis Data Streams with single shard

• Using a single shard creates a bottleneck and limits throughput to 1,000 records per second or 1 MB per second
• The requirement mentions spikes to 100,000 messages per second, which a single shard cannot handle
• Storing messages in DynamoDB adds unnecessary complexity and cost
• Consumer applications reading from DynamoDB would need to implement polling mechanisms

Option D: SNS with multiple SQS subscriptions

• ✓ Decoupling: SNS provides complete decoupling between message producers and consumers
• ✓ Fan-out pattern: SNS can fan out messages to multiple SQS queues, allowing dozens of applications to consume messages independently
• ✓ Scalability: Both SNS and SQS are fully managed services that scale automatically to handle spikes of 100,000 messages per second
• ✓ Independent processing: Each consumer application can process messages from its own SQS queue at its own pace
• ✓ Reliability: SQS provides message durability and at-least-once delivery
• ✓ No bottlenecks: No single shard or processing component limits throughput

Key AWS services used in Option D:

1. Amazon SNS (Simple Notification Service): Publishes messages to multiple subscribers (fan-out pattern)
2. Amazon SQS (Simple Queue Service): Provides durable message queues for each consumer application

This architecture allows the ingestion application to publish messages to an SNS topic once, and each consumer application subscribes to the topic via its own SQS queue. This provides true decoupling, high scalability, and independent processing for all consumer applications.