Microsoft Azure Solutions Architect Expert AZ-305 Practice Question

Azure Cosmos DB is purpose-built for globally distributed, semi-structured data. It supports multiple write regions, guarantees single-digit millisecond latency for both reads and writes, and provides a 99.999 percent availability SLA when multi-region replication is enabled. Cosmos DB automatically indexes every JSON property by default, eliminating manual schema management, and it can elastically scale to hundreds of thousands of request units per second on demand.

Azure Table Storage offers only eventual consistency beyond a single region and indexes only the partition and row keys. Azure SQL Database, even in Hyperscale, does not guarantee single-digit millisecond latency globally and requires manual index management for JSON columns. Azure Data Lake Storage Gen2 is optimized for analytical workloads, not low-latency operational queries, and provides no request-unit-based throughput scaling. Therefore, Azure Cosmos DB with the Core (SQL) API and multi-region writes is the only option that meets all stated requirements.