Redis Labs Establishes New Standard for Instant Multi-Model Databases

Redis Labs, the home of Redis and provider of Redis Enterprise, broke new ground in the database industry today, at RedisConf19, with the introduction of two new data models and a novel data programmability paradigm for multi-model operation.

RedisTimeSeries: Process high-volume data
RedisTimeSeries is designed to collect and store high volume and velocity data, to the scale of billions of data points, and organize it by time intervals. RedisTimeSeries enables organizations to easily distill useful data points with built-in capabilities for downsampling, aggregation, and compression. This provides organizations the ability to query and extract data in real-time for rapid analytics.

RedisAI: Run AI data models within Redis
RedisAI eliminates the need to migrate data to/from different environments and allows developers to apply state-of-the-art AI models to where the data lives, in Redis, dramatically improving the speed with which analytics can be conducted and actions taken. By integrating with common deep learning frameworks including TensorFlow, PyTorch, and TorchScript, and by utilizing Redis Cluster capabilities over GPU-based servers, RedisAI reduces processing overhead and dramatically accelerates the time to insights.

RedisGears: Operate multiple models simultaneously
RedisGears, an in-database serverless engine, is based on the efficient Redis Cluster distributed architecture to enable infinite programmability options supporting event-driven (asynchronously) or transaction-based (synchronously) operations.

With these innovations, Redis can now manage multiple data models driven by a single trigger or application’s request, including native data structures, search, graph, streams, AI, time series, document, and probabilistic data structures. Redis’ true multi-model database architecture is supported by:

• Direct Inter-Model Communications enabling models to consume other models to enrich their functionality. For instance, graph traversal in RedisGraph can start with a full-text search provided by RediSearch to efficiently reach the relevant nodes in the graph.
• Single Dataset Copy paradigm powering efficient inter-model communication with minimal processing and data-copies overhead.
• Full in-database programmability that can automatically process, transform, and query structurally different types of data.

Redis can now run virtually any data model and execute millions of operations per second while maintaining sub-millisecond latency. Redis Enterprise enhances this further by adding greater reliability, linear scalability, deployment flexibility of operating across multiple clouds and in hybrid environments with active-active CRDT-based technology. Redis Enterprise supports globally distributed architectures to power modern applications incorporating high-speed transactions, recommendation engines, data ingest, session management, real-time analytics, caching, and numerous other instant experience use cases. Redis Enterprise can run entirely in DRAM or be tiered intelligently across RAM, in cost-effective NVMe based SSDs or in persistent memory.

“Redis is the first, instant multi-model database with infinite in-database programmability options,” said Yiftach Shoolman, co-founder and CTO at Redis Labs. “With these new advancements, Redis users can process requests across multiple data models asynchronously or synchronously, with sub-millisecond latency at any scale.”

“The number of environments which an organization might want to combine different types of data is increasing and expanding. Because of this, it is our belief that multi- and hybrid-cloud multi-model databases represent the future of modern data-centric applications,” said Philip Howard, research director at Bloor Research. “Redis Enterprise supports multiple logical views into its database, now to include time series data and deep learning models, and has taken this one step further to deliver event-driven data transformations from one model to another, in real-time, in memory. Redis Enterprise will eliminate the need for organizations to deal with siloed data environments and multiple APIs.”