A manufacturing firm needs to ingest sensor data in near real time, persist it in nested JSON format, and support queries with sub-10 millisecond latency. The solution should require minimal operational overhead. Which architecture meets these needs?

The correct answer is to use Amazon Kinesis Data Streams and Amazon DynamoDB. Kinesis Data Streams is ideal for capturing high-throughput, near real-time sensor data, and it can persist data in its original format, including nested JSON. DynamoDB is a fully managed NoSQL database that provides single-digit millisecond latency at any scale, making it suitable for sub-10 millisecond query requirements. Its document model natively supports nested JSON. This combination offers minimal operational overhead as both services are fully managed. Incorrect options: Self-hosted Apache Kafka introduces significant operational overhead. Amazon S3 is an object storage service and is not designed for sub-10 millisecond latency queries on individual records. AWS Lambda is a compute service, not a streaming ingestion service for continuous data streams, and storing data directly in S3 does not meet the latency requirement. Amazon SQS is a message queuing service, not optimized for continuous streaming data ingestion like Kinesis. Amazon RDS is a relational database, which might not be the most efficient for nested JSON and can struggle to meet sub-10 millisecond latency at high scale without significant tuning and operational overhead.