AZ-305 Question 82
Single answerYour company collects and processes large volumes of JSON data with varying structures in near real-time. The data must be stored in a way that supports quick reads and writes, scales globally with low latency, and allows for flexible schema evolution. Which Azure service would best meet this requirement?
- A
Azure Table Storage
- B
Azure Cosmos DB with the Core (SQL) API
- C
Azure Synapse Analytics (Dedicated SQL pool)
- D
Azure Data Lake Storage Gen2
Show answer and explanation
Correct answer: B
Explanation
Azure Cosmos DB is the recommended choice for storing and querying semi-structured data requiring global distribution, low-latency reads/writes, and schema flexibility. According to Microsoft documentation, Cosmos DB supports multiple APIs (including SQL, Cassandra, MongoDB), offers high availability, and automatically indexes JSON documents without requiring schema definitions. Refer to Microsoft’s official documentation on Cosmos DB (https://learn.microsoft.com/azure/cosmos-db/) for additional best practices and implementation guidelines for semi-structured data.
- A. Incorrect.
Option 1: Azure Table Storage. Although it can store key/value pairs effectively and handle semi-structured data, it lacks the rich querying capabilities, global distribution, and automatic scaling that are crucial for large-scale, low-latency operations. It's more suitable for simpler use cases or lower-volume workloads.
- B. Correct.
Option 2: Azure Cosmos DB with the Core (SQL) API. This is the best fit for storing semi-structured data like JSON with flexible schemas, offering global distribution, low-latency reads and writes, and auto-scaling. It also provides strong consistency choices, change feed support, and secondary indexes on JSON properties, which makes it highly suitable for real-time processing.
- C. Incorrect.
Option 3: Azure Synapse Analytics (Dedicated SQL pool). While this is powerful for analytical workloads and can ingest semi-structured data, it is primarily designed for large-scale data warehousing and batch analytics. It does not provide the same low-latency transactional operations and global replication as Cosmos DB.
- D. Incorrect.
Option 4: Azure Data Lake Storage Gen2. This is ideal for high-volume data storage and batch processing scenarios, especially analytics using big data services. However, it is not optimized for real-time reads/writes or rich indexing for varied JSON documents. It is typically used in conjunction with analytical engines rather than as a transactional data store.