Explanation:
The recommended Google Cloud architecture for this scenario is to connect an IoT gateway to Cloud Pub/Sub and use Cloud Dataflow to process the messages (Option C). This approach allows the IoT gateway to batch messages from devices and send them to Cloud Pub/Sub, which can scale to handle varying loads. Cloud Dataflow then processes these messages, offering a scalable and managed solution. Option A suggests virtualizing a Kafka cluster on Compute Engine, which may not be as scalable or cost-effective. Option B, using Edge TPUs, is not ideal for long-term scalability. Option D, connecting Cloud Dataflow directly to the Kafka cluster, does not address the root issue of message volume spikes.
Ultimate access to all questions.
No comments yet.
You are managing a system that processes messages from IoT devices worldwide. Currently, a single on-premises Kafka cluster in the us-east region handles all messages, but managing sporadic spikes in message volume has become challenging and costly. What Google Cloud architecture would you recommend for this scenario?
A
Virtualize a Kafka cluster on Compute Engine in us-east and use Cloud Load Balancing to connect with IoT devices globally
B
Use Edge TPUs as sensor devices for storing and transmitting messages
C
Connect an IoT gateway to Cloud Pub/Sub and use Cloud Dataflow to process the messages
D
Connect Cloud Dataflow to the Kafka cluster to scale message processing