In the latest announcement coming over what we're terming "Database Week," Redis is announcing an OEM deal with Microsoft to implement Redis Enterprise as new tiers for its existing Azure Cache for Redis service. Microsoft, like AWS and Google, already had their own managed caching services based on Redis (often with a choice of Memcached), but the new agreement brings the richer features of Redis Enterprise to the Azure service. The announcement is for a private preview of the new offering.
Redis database is an in-memory NoSQL data platform that has long been utilized as a cache. As open source, it has been the basis of managed caching services provided by each of the major cloud providers. And the premium tier of the original Azure Cache for Redis service included backup to disk; clustering to multiple nodes; configurable firewall rules; multi-region replication, and other features such as data encryption in transit.
By contrast, Redis Enterprise is a superset that is more of a formalized database platform with features such as multi-cluster support, ACID transactions, automated failover, storage tiering to Flash, and various security features that are not part of the original open source core. It also includes support for Redis Modules, comprising SDKs developed by Redis and the community supporting additional data types such as JSON, time series, and graph; search; Bloom filters; machine learning and TensorFlow support; data pipeline building; change data capture; and other capabilities that extend the database.
The existing tiers of Azure Cache for Redis offer data replication with automatic failover, Redis Cluster for cache scale-out, data persistence, Azure virtual network integration, and a 99.9% SLA. The newly-added premium tier, based on Redis Enterprise, will add ability to run Redis on flash storage, additional data types optimized for analytics and machine learning, and the ability to configure active geo-replication for an increased SLA.
As we reported last summer, Redis is aiming to raise its profile from commodity provider of a NoSQL data cache to that of a multi-model database that doesn't just cache data, but persist it. The need to, in essence, get more respect as a data platform, surfaced a couple years back as Redis perceived that cloud providers were monetizing its own IP, and along with MongoDB, Confluent, CockroachDB, and others, began issuing its own novel licensing.
The agreement with Microsoft means that Redis Enterprise will be jointly sold. And to Azure users, it will still look like Azure Cache for Redis as it will use the same interface – it's just a premium offering that gets added to the portfolio.
There are some parallels with the Microsoft/Databricks partnership that took an open source service, tied it in with other Azure data services, and released it as Azure Databricks, in that this becomes an Azure native service that is jointly sold. But the Databricks offering went further with integration to a number of Azure data and storage services including Azure Blob storage and ADLS; the ability to pull data from Azure SQL Database and Azure Cosmos DB; query access from Power BI; and support for role-based access control managed through Azure Active Directory. While the Redis is a different creature – for instance, there probably is not as much need for integration with Azure blob storage, we could see it integrated as an in-memory operational tier complementing Cosmos DB, and we also believe that there could be real-time AI use cases if integrated with Azure Machine Learning. Those are just a couple examples of potential value-added synergies.
Azure is not the first strategic cloud partnership for Redis. Last year, Redis was one of the data platforms that joined Google cloud's open source database partnership. Like the Azure arrangement, there are joint sales and Google still offers its Memorystore service that (like AWS's counterpart) offers the choice of a more bare-bones Redis open source database alongside Memcached. So, there is the precedent of a premium/basic tier combo.
For Databricks, the Azure partnership was clearly an asset that helped propel it to unicorn status. While the market valuation upside in a COVID-19 is likely not to be as dramatic, taking advantage of 1 + 1 = 3 integration with related Azure data services wouldn't hurt.