SAP-C02 Design for New Solutions Practice Question

A logistics company is designing a new event-driven system to process shipment tracking updates. The system must ingest millions of events per second from thousands of IoT devices, retain the events for up to 7 days so multiple independent consumer applications can replay them, and deliver each event to consumers in the order it was received per device. Which AWS service should a solutions architect recommend as the ingestion and retention layer?

⚠ Common exam trap

The trap here is selecting Amazon SQS FIFO because it preserves order, while overlooking that SQS messages are deleted after consumption and therefore cannot be replayed by multiple independent consumer applications.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Amazon Kinesis Data Streams with a retention period of 7 days and partition keys based on device ID.

Kinesis Data Streams is built for high-throughput, ordered event ingestion with configurable retention, and it allows multiple independent consumers to read and replay the same records. Partitioning by device ID guarantees that all events from one device are processed in order by a single consumer. This directly meets the ingestion rate, 7-day retention, replay, and per-device ordering requirements.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
Amazon Simple Queue Service (Amazon SQS) FIFO queues with a message retention period of 7 days.
Why it's wrong here
SQS FIFO queues preserve order within a message group but do not support replaying messages after they are consumed and deleted, so multiple independent consumers cannot re-read the same events. SQS also has a default throughput limit per FIFO queue that requires batching or partitioning to reach millions of events per second. It is a queue, not a replayable log.
✗
Amazon EventBridge with an event bus and archive and replay configured for 7 days.
Why it's wrong here
EventBridge archive and replay can retain events, but the service is optimized for routing events to targets rather than for high-throughput ordered streaming. Ordering per device is not guaranteed, and the default throughput is far below millions of events per second. EventBridge is better suited to application integration events than to high-volume IoT telemetry.
✗
Amazon Managed Streaming for Apache Kafka (Amazon MSK) with a topic retention of 7 days and a single partition per topic.
Why it's wrong here
Amazon MSK can retain data for 7 days and supports replay, but a single partition per topic caps parallelism and throughput, making millions of events per second unachievable. Ordering per device would also be lost because all devices would share one partition. MSK is viable but requires choosing an appropriate partition key and partition count, which this option does not.
✓
Amazon Kinesis Data Streams with a retention period of 7 days and partition keys based on device ID.
Why this is correct
Kinesis Data Streams ingests high volumes of events and retains records for up to 365 days, with 7 days meeting the requirement. Multiple consumer applications can read the same stream independently and replay records within the retention window. Using the device ID as the partition key ensures all events from a given device land in the same shard and are delivered in order.

Go deeper

SAP-C02 Design for New Solutions Practice Question

Why each option matters

Option-by-option breakdown

Related to this question