AZ-305 Question 104
Single answerA retail company processes millions of structured and unstructured transactions daily. They plan to store all data in Azure Data Lake Storage and want to perform near real-time analytics with minimal cluster management overhead. They also need a scalable Spark-based solution that can handle both batch and streaming workloads. Which service should you recommend?
- A
Use Azure Databricks with auto-scaling ephemeral clusters
- B
Deploy and manage Azure HDInsight with Hadoop and Spark
- C
Migrate all raw data to Azure SQL Database for analytics
- D
Set up an on-premises Hadoop cluster and connect via ExpressRoute
Show answer and explanation
Correct answer: A
Explanation
Azure Databricks provides a serverless, managed Spark environment that is ideal for big data and advanced analytics in the cloud. It integrates seamlessly with Azure Data Lake Storage and supports both batch and streaming workloads without the overhead of manually managing clusters. More details can be found in the Microsoft Azure Databricks documentation (https://docs.microsoft.com/azure/databricks).
- A. Correct.
Correct. Azure Databricks offers a fully managed Spark environment with auto-scaling clusters, which simplifies both batch and real-time analytics. Its integrated environment and near real-time streaming support are well-suited for this scenario.
- B. Incorrect.
Incorrect. While Azure HDInsight can run Spark, it requires more hands-on cluster management. It doesn’t offer the same level of simplified, fully managed service that Azure Databricks provides.
- C. Incorrect.
Incorrect. Azure SQL Database can handle structured data, but it’s not ideal for large-scale unstructured or semi-structured analytics scenarios, and it lacks native Spark support for distributed processing.
- D. Incorrect.
Incorrect. An on-premises Hadoop cluster would add significant overhead for capacity management and maintenance, and it doesn’t leverage the benefits of a fully managed Azure service for near real-time analytics.