You want to migrate an existing on-premises application to Google Cloud. Your application supports semi-structured data ingested from 100,000 sensors, and each sensor sends 10 readings per second…

Question

Options

• ADeploy the database using Cloud SQL.
• BUse BigQuery, and load data in batches.
• CDeploy the database using Bigtable.
• DDeploy the database using Cloud Spanner.

• A
  2% (1)
• B
  4% (2)
• C
  84% (42)
• D
  10% (5)

Explanation

Cloud Bigtable is the correct choice for this IoT sensor workload. With 100,000 sensors each sending 10 readings per second, the system must handle 1,000,000 writes per second of semi-structured time-series data - a workload Bigtable was purpose-built for. Bigtable scales to millions of rows per second with single-digit millisecond latency, supports semi-structured wide-column storage, and integrates natively with real-time analytics tools. Cloud SQL (A) is a relational OLTP database that cannot scale to this write throughput without extreme sharding and is poorly suited to semi-structured sensor data. BigQuery (B) is an OLAP analytical warehouse optimized for batch queries, not real-time millisecond writes; it would add unacceptable latency for live monitoring. Cloud Spanner (D) provides global strong consistency and ACID transactions but is optimized for transactional workloads requiring relational integrity - it is significantly more expensive and less optimized for pure high-throughput time-series ingestion than Bigtable.

No community discussion yet for this question.