A financial company receives a high volume of real-time market data streams from an external provider. The streams consist of thousands of JSON records every second. The company needs to implement a s

Question

Options

• AIngest real-time data into Amazon Kinesis data streams. Use the built-in
• BIngest real-time data into Amazon Kinesis data streams. Deploy an Amazon SageMaker endpoint
• CIngest real-time data into Apache Kafka on Amazon EC2 instances. Deploy an Amazon
• DSend real-time data to an Amazon Simple Queue Service (Amazon SQS) FIFO queue. Create an

• A
  73% (24)
• B
  3% (1)
• C
  18% (6)
• D
  6% (2)

Explanation

Option A is correct because Amazon Kinesis Data Streams paired with Kinesis Data Analytics provides a fully managed, serverless pipeline with a built-in Random Cut Forest (RCF) algorithm specifically designed for real-time anomaly detection - no infrastructure to provision, patch, or scale manually.

Why the distractors fail:

• Option B (Kinesis + SageMaker endpoint) increases operational overhead because you must train, deploy, and manage SageMaker model endpoints separately, which adds complexity compared to Kinesis Analytics' built-in anomaly detection.
• Option C (Apache Kafka on EC2) requires you to provision, configure, and maintain EC2 instances and a Kafka cluster yourself - the opposite of "least operational overhead."
• Option D (SQS FIFO queue) is designed for decoupled message queuing, not high-throughput stream analytics. FIFO queues also have throughput limits (300–3,000 msg/sec) that would struggle with thousands of JSON records per second, and have no native anomaly detection.

Memory tip: For AWS exam questions asking about real-time streaming + anomaly detection + least overhead, think Kinesis Data Streams → Kinesis Data Analytics. The phrase "built-in" in option A is the signal - it means no extra ML infrastructure to manage. If you see EC2 or self-managed Kafka, eliminate it immediately for "least overhead" questions.

No community discussion yet for this question.