SAA-C03 Question 456
Single answerA company is designing an IoT solution to process video data from thousands of surveillance cameras deployed across multiple locations. The solution should minimize latency and reduce the need to send large volumes of raw video data to a central AWS Region for analysis. Which approach should the company implement to achieve these requirements?
- A
Use AWS Lambda functions deployed in the central AWS Region to process video streams in real-time.
- B
Deploy Amazon Elastic Kubernetes Service (Amazon EKS) clusters in the central AWS Region to process video data.
- C
Implement AWS IoT Greengrass on edge devices to process video data locally before sending only necessary insights to AWS.
- D
Stream all raw video data to Amazon S3 and use AWS Glue to process the video files.
Show answer and explanation
Correct answer: C
Explanation
The best solution for reducing latency and minimizing the transmission of raw video data is to use AWS IoT Greengrass. Greengrass enables IoT devices to process data locally at the edge, allowing the system to perform real-time processing and only send necessary insights to the cloud. This approach is particularly effective for distributed compute strategies in scenarios where low latency is critical.
- A. Incorrect.
Using AWS Lambda in the central AWS Region would introduce latency because the video streams must travel to the Region for processing, which does not align with the requirement to minimize latency.
- B. Incorrect.
While Amazon EKS is capable of processing large amounts of data, deploying it in the central AWS Region would still lead to increased latency as video data has to be transmitted to the Region.
- C. Correct.
AWS IoT Greengrass allows edge devices to process data locally, significantly reducing latency and the need to send large volumes of raw video data to the cloud. This matches the requirements of the scenario.
- D. Incorrect.
Streaming all raw video data to Amazon S3 and processing it with AWS Glue is not suitable for real-time processing and would lead to high latency, as well as increased costs associated with data storage and processing.