CASE STUDY: AutoMakers Inc. 1M connected cars, 100GB/day telemetry. Req: Predictive maintenance, real-time driver dashboard, monetize data. CEO: Data is new engine. CFO: Cut 3rd-party IoT costs. CTO: Highly scalable ingest. Tech: MQTT ingest, stream processing, ML models, 7-yr cold storage, handle intermittent connectivity. Constraints: Anonymize data, low vehicle compute, strict analytics budget.
How should you architect the highly scalable ingestion layer for MQTT telemetry data from 1 million cars?
GCP PCA · Question 15 · Domain 5: Managing Implementation and Ensuring Solution and Operations Reliability
CASE STUDY: AutoMakers Inc. 1M connected cars, 100GB/day telemetry. Req: Predictive maintenance, real-time driver dashboard, monetize data. CEO: Data is new engine. CFO: Cut 3rd-party IoT costs. CTO: Highly scalable ingest. Tech: MQTT ingest, stream processing, ML models, 7-yr cold storage, handle intermittent connectivity. Constraints: Anonymize data, low vehicle compute, strict analytics budget.
How should you handle the intermittent connectivity of vehicles to ensure no telemetry data is lost when they reconnect?
Answer options:
Increase the RAM on the vehicles to store data locally for months.
Configure Pub/Sub message retention to hold messages until the processing pipeline acknowledges them.
Use Cloud SQL to buffer incoming connections.
Require vehicles to use Dedicated Interconnect.
50 questions · hints · full answers · grading
Expert