June 2024 Summaries

This is an update from Redis about new features and enhancements released in June, including Redis Data Integration (RDI) now being generally available, RDI coming to Redis Insight with additional features like Redis Copilot for easy search query construction. The enhanced Redis Query Engine is also available on Redis Software, boosting query throughput by 16X. Additionally, client-side caching is now a private preview for Redis Software customers, and a new release of Terraform for Redis Cloud registry version 1.7 supports Essentials cloud subscriptions and Active-Active databases. A public preview of Redis version 7.2 and the E1 (1GB) SKU on Azure Cache for Redis Enterprise is also available, allowing users to upgrade to the latest Redis features with improved performance and availability.

Redis has introduced a new enhancement to its Redis Query Engine, which accelerates current query, search, and vector workloads. This improvement boosts the current Redis query throughput by 16X, making it faster than any other vector database tested against. The new engine is now generally available in Redis Software and will be coming to Redis Cloud this fall.

Redis Cloud Packages offer ready-to-use, pre-configured setups for deploying Redis instances fast and customized for specific use cases without manual tuning. These packages provide optimal solutions for performance, reliability, and efficiency, catering to various workload requirements and use cases such as caching, NoSQL databases, and vector databases. With flexible control and a comprehensive set of features, including automatic failover and strong security, developers can speed up their development processes, focus on innovation, and deliver great customer experiences. The packages also offer intelligent defaults, simplified management, and unmatched reliability, transforming Redis deployments and making them easier to manage and more valuable to organizations.

Redis is faster than other vector database providers in querying throughput and latency times. It outperformed Qdrant, Milvus, and Weaviate in querying throughput, achieving up to 3.4 times higher queries per second (QPS) than Qdrant, 3.3 times higher QPS than Milvus, and 1.7 times higher QPS than Weaviate for the same recall levels. On latency, Redis achieved up to 4 times less latency than Qdrant, 4.67 times than Milvus, and 1.71 times faster than Weaviate for the same recall levels. Additionally, Redis showed lower indexing time compared to other providers, with up to 2.8 times lower indexing time than Milvus and up to 3.2 times lower indexing time than Weaviate. The performance advantages of Redis were demonstrated in benchmarks against various vector database providers, including Qdrant, Milvus, Weaviate, Amazon Aurora PostgreSQL v16.1 with pgvector, MongoDB Atlas v7.0.8 with Atlas Search, and Amazon OpenSearch 2.11. Redis also outperformed other pure vector database providers in general-purpose databases with vector capabilities, such as Amazon MemoryDB and Google Cloud MemoryStore for Redis. The performance improvements were achieved through the introduction of a new enhancement to enable concurrent access to the index, allowing multiple queries to be executed concurrently on separate threads. This approach enabled efficient resource utilization across all configurations, resulting in consistent performance improvements.

Retrieval Augmented Generation (RAG) has become the standard architecture for GenAI applications requiring access to private data. The key challenge is maintaining fast application performance when incorporating AI. Paul Buchheit's "100ms Rule" suggests that every interaction should be faster than 100ms to feel instantaneous. A typical RAG-based architecture has an average end-to-end response time of 1,513ms, which is not ideal for user engagement. Redis offers three main datastore capabilities for AI: vector search, semantic caching, and LLM Memory. These features enable real-time RAG by significantly improving user experience end-to-end. By utilizing Redis' capabilities for AI, a GenAI application can achieve an average end-to-end response time of 389ms, which is around x3.2 faster than non-real-time RAG architectures and closer to the 100ms Rule.

RDI (Redis Data Integration) is a tool that synchronizes data from existing databases into Redis in nearly real-time, allowing developers to build fast apps without investing significant time and money into building their own data pipeline. RDI solves the problem of slow and expensive databases holding back app development by providing a robust and efficient method to get data into Redis and keep it sync between source databases and Redis. With RDI, companies can turn slow data into fast data, scale without limit, and stop wasting money on database costs. By integrating slow databases with Redis, RDI performs data ingestion and transformation to the Redis schema, enabling apps to access data really, really fast. RDI makes fast apps possible by keeping data in sync with configuration, not code, avoiding resource drain and teams manually building data pipelines. It also allows for seamless deployment of pipelines from within Redis Insight, performing code completion and syntax validation, validating transformation and pipeline output, and monitoring data flow and pipeline performance in an intuitive dashboard. RDI opens the door for businesses to modernize their apps without expensive database expenses, providing a cost-effective way to get fast and reliable data access.

Redis and NVIDIA NIM are being used to accelerate GenAI app development by providing a fast data platform for real-time data and AI applications. This combination enables companies to build and deploy GenAI apps faster, while staying ahead of the challenges of software development in this field. By using Redis as their vector database and NVIDIA NIM for model deployment, developers can skip the setup and maintenance of full-stack infrastructure, streamlining AI model deployment with pre-built cloud-native microservices that deliver optimized inference on NVIDIA accelerated infrastructure. With Redis and NVIDIA NIM, companies can reduce costs and speed up responses to provide real-time experiences for their users. The demo showcases how easy it is to get started with this technology, allowing developers to build a simple chatbot that uses RAG with a Redis vector database and NIM for model and inferencing, generating accurate responses in seconds.