As VMware customers have continued their digital transformation and modernization journeys over the years, we’ve seen directly how relational databases have proved to be reliable and stable resources for data management, while traditional disk-based databases can struggle to meet increasing demands. I, personally, have witnessed many cases where slow disk data access introduced system bottlenecks. In particular, slow disk access latency tends to increase as the number of users with complex and or repetitive queries grows. This results in poor user experiences and unsatisfied customers.
High-volume transactions and low-latency data requirements may not be suitable for traditional disk-based relational databases. Real-time transactional data transformation activities (such as moving from monolithic architectures to microservices) increase the number of applications. In these cases, it is often difficult to scale traditional relational databases to satisfy the increasing number of applications.
Fast, modern caching at scale
Real-time transactional, analytical, and AI applications need fast data access in order to improve the user experience. This is where caching solutions like Tanzu for Valkey are important. Valkey is an open source, high-performance key/value and streaming database. VMware Tanzu for Valkey is an enterprise-ready version of Valkey that’s integrated into the Tanzu ecosystem. Its goal is to simplify provisioning, operation, scaling, and maintenance of Valkey instances.
Tanzu for Valkey supports low-latency in-memory caching and messaging. In-memory caching helps reduce read latency and offload repeated queries from disk-based databases. It can also act as a primary database, running standalone or in a cluster with replication and high availability. Valkey instances can be attached to modern applications for scalability and low-latency data access.
A competitive alternative to Redis
Redis changed its open source licensing model beginning with version 7.4, instead adopting a restricted license. As a result, the open source community created a fork of Redis named Valkey. Since then, numerous performance improvements and added features have made Valkey an exciting choice as an open source alternative to Redis.
Valkey’s use of the Redis’s RESP protocol makes it compatible with Redis applications for drop-in replacement. If an application uses a Redis client library, it can likely connect to Valkey without any code changes.
Improvements since Redis OSS (7.x)
Valkey adds multi-threaded scalability and significant performance improvements. For example, benchmarks highlight 200% performance improvement and approximately 20% memory reduction compared to Redis OSS 7.4.
Valkey added native JSON support. It adds vector database capabilities for AI/ML. Valkey LDAP integrates with existing identity management infrastructure. It supports query data access using JSON Path semantics.
Valkey contains availability-zone deployment awareness for client data access, along with enhanced observability metrics. The open source community has contributed significant improvements.
Tanzu for Valkey use cases
Tanzu for Valkey provides fast access (low latency, high throughput) and in-memory caching of all sorts of data types, such as strings, numbers, hashes, sets, and JSON structures. Valkey supports searching for data with single-digit millisecond latency and high-throughput queries per second (QPS) for hash or JSON data types with complex filters (AND OR logic) using JSON PATH. It also supports GEO spatial searches to find records based on Global Positioning System (GPS) coordinates.
Applications can implement a look-aside caching pattern to improve database read access. It is a good choice for web session state caching and API rate limiting when combined with Spring projects for Java applications.
Easy event streaming and event logs
Tanzu for Valkey supports event streaming use cases, similarly to Kafka and RabbitMQ Stream. Apps can listen for events or replay events as needed. Valkey Streams (similar to Redis Streams) is ideal for event-driven architectures for use cases such as ordered delivery of messages. For example, Valkey’s event streaming can instantly ingest ordered products in real time from a Producer App. Since Valkey operates in-memory, it delivers submillisecond latency to a Consumer App to process the orders.
Valkey Streams allows applications to react instantly to events like orders, product updates, and other areas of interest. It maintains a durable and ordered event log. This enables replaying data for auditing, recovery, or analytics purposes. Multiple Consumer Apps can independently read the same stream from a Producer App. Valkey Streams unlocks scalable low-latency, in-memory, real-time processing for event-driven architectures.
Tanzu for Valkey as a vector database
Valkey Search offers a high-performance vector similarity search engine. Imagine your company’s knowledge is scattered across many documents. Using Valkey as a vector database can help turn your business domain documentation into relevant insights for your AI/ML use cases powered by in-memory speed.
Valkey’s vector database implementation is optimized for AI-driven workloads and can support handling billions of vectors. It implements exact search matching using K-Nearest Neighbors (KNN) or Approximate Nearest Neighbor (ANN) search. Applications can combine semantic searching with metadata pre- or in-line filterings.
Combining Tanzu for Valkey with Spring AI can help provide large language models (LLM) business domain content from Word documents, PowerPoint documents, web pages, Retrieval Augmented Generation (RAG), or even Cache Augmented Generation (CAG) use cases.
Included within Tanzu Data Intelligence
Tanzu Data Intelligence contains a portfolio of data products for both transactional and analytical data use cases. It supports data lakehouse architectures for data science and AI teams. It supports real-time data, federated query services, data warehousing, data lakehouse, and distributed data compute services.
Tanzu Data Intelligence contains many popular components, such as VMware Tanzu RabbitMQ, VMware Tanzu Greenplum, VMware Tanzu Data Lake, VMware Tanzu GemFire, VMware Tanzu for Postgres, and VMware Tanzu for Valkey, with support for MadLib, PostgresML, Apache, Iceberg, Apache Spark, and more. There’s also drag-and-drop event streaming, like Tanzu Data Flow in the VMware Tanzu portfolio, to keep Valkey up to date with data in external sources.
Customers can leverage Tanzu for Valkey’s speed, versatility, and scalability for caching and real-time access.
The VMware Tanzu team’s expertise
VMware has a long history developing the popular Spring Data Redis library.
Spring projects such as Spring Cache, Spring Cloud Gateway, Spring Integration, and Spring Cloud Data Flow (now Tanzu Data Flow) use caching technologies like Tanzu for Valkey. Valkey integrates easily with Spring Boot and Spring Data.
Caching-dependent applications have simplified server deployments with the Tanzu Platform Tile since approximately 2017. And the VMware Tanzu team has been an early adopter of Valkey for containerization, first with Bitnami and now Tanzu Application Catalog.
Tanzu for Valkey simplifies Day 1 and Day 2 operations on Kubernetes and Cloud Foundry with award-winning 24×7 enterprise support. Many companies also turn to VMware Tanzu for security invulnerability remediation. Valkey on the Tanzu Platform has the potential to significantly increase the return on investment (ROI) with provisioning and management automation.
In summary, Tanzu for Valkey is a robust, open source, and cost-effective alternative to Redis, backed by the VMware Tanzu team’s expertise. Tanzu for Valkey offers a comprehensive data services offering, especially for applications requiring fast, scalable, and highly-available data access. To learn more about Tanzu for Valkey, reach out to us here.
