AZ-305 Design infrastructure solutions Practice Question

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Azure Table Storage

  Why it's wrong here

  Azure Table Storage is a schema-free NoSQL key-value store that scales horizontally across partitions, but each partition is handled with throughput limits and access is typically via OData/AtomPub. It does not offer Cosmos DB's global distribution, multi-region writes, or SLA-backed single-digit-millisecond latency, so while it is a cheap option for structured non-relational data, it falls short for high-throughput microservices and is less valid in this scenario.

• ✓

  Azure Cosmos DB

  Why this is correct

  Azure Cosmos DB is a fully managed, globally distributed, multi-model database that exposes multiple APIs, including Core (SQL), MongoDB, Cassandra, Gremlin, and Table. It provides comprehensive SLAs for single-digit-millisecond read and write latency at the 99th percentile, supports multi-region writes and automatic horizontal partitioning, and scales throughput independently as request units, which makes it a proven valid data storage solution for demanding microservices.

• ✗

  Azure Files

  Why it's wrong here

  Azure Files is a managed network file system that provides SMB and NFS shares, so it is designed for shared file access, lift-and-shift of on-premises file servers, and shared configuration data rather than transactional database workloads. Even premium file shares have bounded IOPS, bandwidth caps, and file-level lock semantics that are poorly suited to frequent concurrent updates, and a microservices data layer requires database-level querying, indexing, and low-latency access that Azure Files simply lacks.

• ✓

  Azure Cache for Redis

  Why this is correct

  Azure Cache for Redis is a fully managed, distributed in-memory key-value store based on the open-source Redis engine, delivering sub-millisecond read and write latency with rich data structures such as strings, hashes, sorted sets, and streams. With features like persistence, clustering, time-to-live, transactions, and pub/sub, it can serve as a real data storage layer for session state, leaderboards, rate limiting, and other hot-path data, which is why it is a valid data storage solution for microservices.

• ✗

  Azure SQL Database

  Why it's wrong here

  Azure SQL Database is a fully managed relational database engine with strong schema enforcement, ACID transactions, and a mature query optimizer, but those relational features add overhead in transaction logging, lock management, and query planning that limits predictable throughput under extreme microservice loads. It can be scaled with service tiers, elastic pools, and read replicas, yet it does not natively provide the same globally distributed, multi-region write latency guarantees as Cosmos DB, making it a less appropriate choice for this high-throughput data storage scenario.

Go deeper

Related to this question