Google Professional Cloud Developer Question 184
Single answerGoogle Cloud PlatformYour team is developing a streaming analytics application on Google Cloud that processes high volumes of real-time sensor data. The application needs to handle fluctuating data volumes efficiently while minimizing operational overhead. Which approach should you use to ensure scalability and reliability?
- A
Use Cloud Pub/Sub for ingesting data and Dataflow for processing the data with autoscaling enabled.
- B
Deploy a self-managed Apache Kafka cluster for data ingestion and use Compute Engine instances to process the data.
- C
Use Cloud Storage to store real-time data and run periodic batch jobs with Dataproc for processing.
- D
Run a Kubernetes cluster on Google Kubernetes Engine (GKE) with a custom data processing application to handle incoming data.
Show answer and explanation
Correct answer: A
Explanation
For a real-time streaming analytics application with fluctuating data volumes, it's essential to use services that offer horizontal scalability and minimize operational overhead. Cloud Pub/Sub is a fully managed messaging service built to handle high-throughput data ingestion, while Dataflow provides a serverless, autoscaling solution for stream processing. This combination ensures scalability, reliability, and ease of management, making it the most suitable choice for the given scenario.
- A. Correct.
This is the correct answer. Cloud Pub/Sub provides a fully managed, horizontally scalable messaging service that can ingest high volumes of data, while Dataflow supports autoscaling for real-time data processing, reducing operational overhead.
- B. Incorrect.
While Apache Kafka is suitable for data streaming, managing and scaling an Apache Kafka cluster requires significant operational effort. Similarly, Compute Engine instances do not offer built-in autoscaling for streaming workloads in the same way as Dataflow.
- C. Incorrect.
Cloud Storage is designed for storing large datasets but is not optimized for real-time data ingestion. Batch processing with Dataproc introduces latency and is not suitable for the real-time nature of the application.
- D. Incorrect.
While GKE provides flexibility for custom applications, managing and scaling a Kubernetes cluster requires substantial operational effort compared to a fully managed service like Cloud Pub/Sub and Dataflow.