During this year’s digital Build event, Microsoft announced a new consumption-based pricing model for Azure Cosmos DB called ‘serverless’ that is ideal for spiky workloads and complements the existing provisioned throughput pricing model.
Furthermore, Microsoft also announced a few new capabilities in general availability and preview with autoscale pricing model, new features for Jupyter Notebooks, a new version of the Azure Cosmos DB Python SDK for Core (SQL) API, Azure Private Link and Azure Synapse Link for Azure Cosmos DB, and encryption and rest – next to a couple of upcoming capabilities such as Point-in-time-restore (PTR), a new version of the Java SDK, and an update to the change feed.
Azure Cosmos DB is Microsoft’s NoSQL offering on its cloud platform. Earlier this year in March, the cloud provider attempted to allow more developers access to Azure Cosmos DB with a free version. However, it did not give developers full access to the Cosmos DB service. To fix this issue, the company now offers a serverless pricing model for Azure Cosmos DB. With Azure Cosmos DB serverless, customers only get billed for resources consumed per operation, eliminating the concept of provisioned throughput, and significantly lowering the entry price for customers.
In the book of news of build 2020, it states that Azure Cosmos DB serverless pricing model offers per-operation compute pricing ideal for noncritical, intermittent and small workloads with moderate performance requirements. Consumption-based pricing bills customers for the total resources (request units) used for each discrete database operation.
Source: https://devblogs.microsoft.com/cosmosdb/autoscale-serverless-offers/
Thomas Weiss, program manager Azure Cosmos DB, said in a video around the database operations models for Azure Cosmos DB:
Azure Cosmos DB Serverless is going to be a pure consumption model. Where only RU/s consumed by your request will get billed.
And Deborah Chen, program manager Azure Cosmos DB at Microsoft, stated in a recent Cosmos DB blog:
Because serverless is a true pay-per-request billing model, it will lower even further the entry price for anyone who wants to start using Azure Cosmos DB or run small applications with light traffic.
Next to the announcement of the new serverless pricing model, Microsoft released a couple of updates for Azure Cosmos DB into general availability:
- Encryption at rest with customer-managed keys, offering customers the option to bring their keys for enhanced end-to-end encryption of their data.
- Azure Cosmos DB autoscale (previously called autopilot), another pricing model which offers SLA-backed single-digit millisecond latency and 99.999% availability guaranteed.
- The latest version of the Azure Cosmos DB Python SDK for Core (SQL) API. The 4.0 version includes usability improvements to the object model for better developer experience, support for new query features, bug fixes, and more.
- Azure Private Link for Azure Cosmos DB, allowing users to connect with Azure Cosmos DB from their virtual network via a private endpoint – which is a set of private IP addresses in a subnet within the virtual network.
And in preview:
- New features for Azure Cosmos DB built-in Jupyter Notebook support, simplifying development for developers to analyze and visualize their Azure Cosmos DB data.
- Azure Synapse Link for Azure Cosmos DB, which brings a tight integration between Azure Cosmos DB and Azure Synapse Analytics. Using the Azure Cosmos DB analytical store, a fully isolated column store, Azure Synapse Link enables no-ETL (extract-transform-load) analytics in Azure Synapse Analytics against globally distributed operational data at scale.
Reference: https://www.infoq.com/news/2020/06/azure-cosmosdb-new-updates/
