Examkin

Soru

Zorluk: ZorSelecting and Designing Storage and Database Resources

An IoT telematics provider is designing a storage layer in a single Google Cloud region to ingest and store real-time sensor metrics from 100,000 connected vehicles. The system requires high-throughput write performance exceeding 80,000 operations per second with sub-10 millisecond write latencies, dynamic key-value access patterns, and seamless analytical querying capabilities in BigQuery without custom ETL pipelines. Which Google Cloud storage solution should the cloud architect select to satisfy these operational and analytical requirements at scale?

  1. Provision a Cloud Bigtable cluster and configure BigQuery federated queries to analyze the data in place.Cevap
  2. B
    Deploy a multi-region Cloud Spanner instance with strong global consistency to store the incoming sensor streams.
  3. C
    Provision a Cloud SQL for PostgreSQL instance encrypted with Customer-Supplied Encryption Keys (CSEK) for key management.
  4. D
    Store incoming telemetry data directly as JSON objects in Cloud Storage using IAM granular policies for access restriction.

Cevap

Cloud Bigtable with BigQuery federated querying is the optimal choice for high-throughput, low-latency time-series telemetry data.
Cloud Bigtable is Google Cloud's managed NoSQL wide-column database built specifically for high-throughput, low-latency workloads such as IoT sensor telematics. It easily scales horizontally to handle over 80,000 operations per second with single-digit millisecond latency. Furthermore, BigQuery can natively execute federated queries against Cloud Bigtable tables, enabling real-time analytics without needing an intermediate ETL data pipeline.

Adım Adım Çözüm

1
Analyze workload write throughput and latency requirements
Identified a single-region workload requiring >80,000 writes/sec at <10ms latency for unstructured/key-value time-series data.
Matching write scale and latency bounds narrows candidate database engines to high-performance NoSQL engines.
2
Evaluate candidate storage engines against the access pattern
Cloud Bigtable provides seamless horizontal scaling for high-frequency key-value and time-series telemetry writes.
Relational databases like Cloud SQL cannot sustain this write volume without sharding, while Cloud Spanner is built for multi-region relational ACID workloads rather than single-region key-value metrics.
3
Verify analytics integration requirement
BigQuery can query Cloud Bigtable directly via external tables without running complex ETL pipelines.
Meets the requirement for immediate analytical access without building custom data movement pipelines.

Anahtar Kavram

Selecting Cloud Bigtable for high-throughput NoSQL time-series workloads and leveraging BigQuery federation for analytics.
Tahmini Süre:2m 0s
Bu soruyu puanla

Konular

Selecting and Designing Storage and Database Resources