AZ-305 Question 106
Select 2A manufacturing company collects high-volume streaming data from IoT sensors deployed across multiple global facilities. The data is both structured (e.g., order statuses, temperature readings) and semi-structured (e.g., logs). They want to ingest and analyze this data in near real-time, support ad-hoc queries, and incorporate machine learning for anomaly detection. Historical data should also be stored cost-effectively for long-term analytics. Which two solutions best meet these requirements for ingestion, storage, and analysis?
- A
- Use Azure Data Explorer for near real-time ingestion and querying, with Azure Data Lake Storage for archiving.
- B
- Use Azure Event Hubs for ingestion, store the data in Azure Table Storage, and connect Power BI using DirectQuery for advanced analytics.
- C
- Use Azure Synapse Analytics serverless SQL pools with Azure Data Lake Storage to query batch data and build machine learning models directly on data files.
- D
- Use Azure IoT Hub and Azure Stream Analytics for real-time ingestion, store data in Azure Data Lake Storage Gen2, and perform advanced analytics and machine learning in Azure Databricks.
Show answer and explanation
Correct answers: A, D
Explanation
Both Azure Data Explorer (Option 1) and the IoT Hub/Stream Analytics/Databricks pipeline (Option 4) directly address continuous ingestion and querying of streaming data, while also providing pathways for advanced analytics and machine learning. They are designed to handle large-scale, near real-time data flows and can integrate with Azure Data Lake Storage for cost-effective archival and further analysis. More information can be found in the official documentation for Azure Data Explorer (https://learn.microsoft.com/azure/data-explorer/) and IoT ingestion pipelines with Databricks (https://learn.microsoft.com/azure/databricks/).
- A. Correct.
Option 1 is correct. Azure Data Explorer (ADX) is optimized for near real-time data ingestion and analytics on large volumes of streaming data, especially logs and telemetry. It can connect to Event Hubs or IoT Hub to ingest data, and ADX queries can run quickly even on high-velocity data. Storing older or less frequently accessed data in Azure Data Lake Storage is a common pattern for cost-effective archival.
- B. Incorrect.
Option 2 is incorrect. While Azure Event Hubs is a scalable ingestion service, storing large volumes of data in Azure Table Storage is not ideal for advanced analytics or machine learning scenarios; it's more suited for key-value data with limited query patterns. Power BI DirectQuery on Table Storage is also limiting in terms of query performance and advanced analytical capabilities.
- C. Incorrect.
Option 3 is plausible for batch-oriented analytics but less optimal for near real-time ingestion requirements. Azure Synapse serverless SQL pools are excellent for on-demand queries over data stored in Azure Data Lake Storage, but they don't provide built-in real-time ingestion or advanced ML pipelines without additional components (e.g., Spark pools). While a capable solution, it may not fully address the streaming and immediate analytics needs described.
- D. Correct.
Option 4 is correct. Azure IoT Hub seamlessly ingests sensor data, and Azure Stream Analytics can process and aggregate data in near real-time. Storing the data in Azure Data Lake Storage Gen2 allows for scalable, cost-effective historical storage. Azure Databricks provides notebooks, Spark-based processing, and ML libraries, making it suitable for advanced analytics and machine learning scenarios.