Redis Labs, Microsoft offer massive dataset support with enterprise cache service

Like Redis cache? Wait till you’ve tried it with flash


Sponsored Redis Labs has introduced the first native integration of Redis Enterprise Software on a major cloud platform. Developed jointly with Microsoft, Redis Labs announced Redis Enterprise on Azure Cache for Redis, offering an up to 99.99 availability SLA and the ability to support datasets up to 15TB.

The Azure Cache for Redis Enterprise service has been in private preview since May, but has now moved into public preview, giving developers access to two new tiers, Enterprise, and Enterprise on Flash, via the Azure Portal.

Microsoft already offers Azure Cache for Redis, a managed Redis cache service, based on its own implementation of v4.0 of the in-memory data store, and pitched as a complement to Azure database, K8s and App Services.

But the new Enterprise tiers expand upon the existing Azure Cache for Redis tiers and are based on Redis Labs’ Enterprise implementation of v6.0 of Redis. Redis Labs is the official sponsor of the Redis open source project, and Redis creator Salvatore Sanfilippo previously headed its open source team and remains on the company’s technical advisory board.

The collaboration brings a raft of enterprise level Redis features to Azure, enabling higher SLAs, and additional functionality including three Redis data modules, all integrated natively into the Azure Portal. Further enterprise class features will come as the two new tiers move to general availability in 2021, including support for Redis Labs’ Active geo-distribution, and support for hybrid cloud setups.

Redis Labs’ Director of Strategic Alliances, Cassie Zimmerman, said “For years developers have turned to Redis for its unbeatable responsiveness and ability to scale in their applications. The introduction of Redis Enterprise technology native on Azure Cache extends upon the range of use cases developers can use Redis for, while providing next level resiliency and performance.”

A case of joint enterprise

Zimmerman said the new Enterprise tiers had been jointly developed and engineered by Microsoft and Redis Labs and is a fully managed service operated by Microsoft. Zimmerman added the service is supported jointly, with Microsoft handling the first support call and pulling in dedicated Redis Labs support expertise as needed. She added as a native Azure service, this solution offers unique benefits like access to set-up and configuration through the Azure Portal, CLI, and Azure API. The Enterprise tiers will support integrated billing and customers can use their Azure spend commitments on the Enterprise tiers. As a result, enterprise customers who were already taking advantage of the developer experience of Azure Cache for Redis, would now be able to take full advantage of the functionality of Redis Enterprise.

In common with the existing premium tier of Azure Cache for Redis, the newly available Enterprise tiers will offer horizontal scaling, data persistence, and private link networking isolation. The Enterprise tiers are also natively integrated into popular Azure security and monitoring tools.

Both the Enterprise tiers will debut with availability of up to 99.99 per cent with zone redundant replication. Once in GA, reliability will be an even greater priority with a potential 99.999 percent SLA when active geo distribution is rolled out. Active geo distribution is based on Redis Labs’ CRDT (conflict-free replicated data types) technology, which offers unique benefits over other geo-distributed solutions to ensure the highest level of resiliency. This means customers access the database that’s closest to them, at local latency speeds.

Once in GA, users can leverage the Enterprise tiers to configure a globally available cache that can fail to another region without any data loss. This level of resiliency is particularly important to enterprise customers who can’t risk downtime and are using Redis to support business and mission critical applications. In the future, this feature will also allow customers to connect on-premises workloads with Redis Caches in Azure for a hybrid deployment.

The Enterprise tiers will also bring three of Redis Enterprise’s additional data modules which offer data structures that are not available in open source flavours of Redis. RediSearch offers secondary-indexes and full text search over Redis. RedisBloom offers the use of additional probabilistic data structures, including Bloom filters, which can be used to quickly see whether an item is present or absent from a collection with a given degree of certainty. RedisTimeSeries allows developers to use a Time Series data structure in Redis. Making the additional data modules available to developers gives them a new level of flexibility in how they could use Redis on Azure.

In addition, the new Enterprise tiers will offer substantially increased memory sizes, with 12GB to 1TB plan sizes in the Enterprise tier, and 384GB to 15TB available in the Enterprise Flash tier. The existing Azure tiers top out at 1.2TB for the clustered version of the previous tier, or 120GB non-clustered.

The flash service is based on Redis Labs technology and uses SSDs to supplement RAM. This enables the use of much bigger datasets without the cost implications of operating entirely within RAM.

Lots of 9s, and lots and lots of flash

Zimmerman added that Redis on Flash is not meant to be used for an alternative for data persistence – the same AOF and snapshot data persistence mechanisms are used with Redis on Flash.

Zimmerman said “For example, instead of keeping the entire database in DRAM, Redis on Flash keeps only the keys and the Redis dictionary in DRAM, along with the “hot values” of the dataset. The “warm values”, or the lesser used portion of the dataset, are kept on the local Flash storage, Overall cost reductions by using flash compared to an all RAM configuration can be as much as 70 per cent.

The Enterprise tiers complementary cache to Azure’s existing database services, like Azure SQL Database and Azure Cosmos DB. “Redis is very often used as a complement to back-end datastores to accelerate response times, handle the scale and reduce the load,” explained Zimmerman.

Zimmerman said among the most common use cases for Azure Cache include distributed caching, session store, message broker, as well as cloud migration and gaming leaderboards.

The private preview has sparked interest from large enterprise customers, with the increased levels of DR and availability being particular draws, Zimmerman explained. “Redis Labs is excited to already see the value the solution is bringing to our joint customers who are expanding their Redis use case support through modules and achieving a higher SLA.”

The public preview service gives developers an opportunity to preview the ways in which they can use the service to build their applications.

Private link is currently in preview, while general availability will also bring a hybrid configuration option, which Zimmerman said would be particularly useful for companies looking to migrate on-premises Redis or legacy data workloads to Azure and are looking for a fully managed native service.

She added, the service is not limited to users with existing Redis experience. Any developer looking to speed up their database operations or developing real-time applications can benefit from Redis. Plus, developers already using Redis now have the ability to do more with the introduction of Redis Modules.

“As the most beloved datastore in the world, developers love the simplicity in Redis. Redis Labs introduced modules as a way for developers to leverage a tool they already know, trust, and love, for more. With the added operational resiliency, our goal is to enable developers to build applications quickly while ensuring they operate on a global scale with the highest resiliency possible.”

Sponsored by Redis Labs


Biting the hand that feeds IT © 1998–2020