Azure Cosmos DB: NoSQL capabilities everyone should know about

Play Azure Cosmos DB: NoSQL capabilities everyone should know about
Sign in to queue


Microsoft Azure provides a fully managed NoSQL database service built for fast and predictable performance, high availability, elastic scaling, global distribution, and ease of development. As a schema-free NoSQL database, the service provides rich and familiar SQL query capabilities with consistent low latencies on JSON data - ensuring that 99% of your reads are served under 10 milliseconds and 99% of your writes are served under 15 milliseconds. These unique benefits make it a great fit for web, mobile, gaming, IoT, AI, and many other applications that need seamless scale and global replication. Come and learn about the NoSQL capabilities in Azure Cosmos DB that every developer should know about.



Session Type:

Tech Talk




Tech Talk A



The Discussion

  • User profile image

    Regarding the 'seamless' migration of Azure Storage Tables to a storage optimised tier of Cosmos DB Tables API - will this still support the current per transaction/per request billing model without requiring to pre-provision dedicated capacity?

  • User profile image

    By 'seamless' they meant no changes to code. From pricing perspective cosmos db (you will need something close to strong consistency, eventual consistency will most likely break your app) is much more expensive and depending what you are doing you might not notice any difference in throughput / latency.

    For example batch insert to table might be not be faster at all, you most likely won't benefit from secondary indexes because you designed you solution to work well with azure table (access only via PK/RK) which doesn't have secondary indexes etc

    Personally i would consider cosmos tables only when
    1) you are creating completely new solution and cost is basically not an issue
    2) you REALLY need fast reads or you need to combine multiple azure storage account 20k req/s to do whatever you are doing
    3) you need same data in multiple regions

Add Your 2 Cents