Stop testing, start deploying your AI apps. See how with MIT Technology Review’s latest research.
Download now
share
For many years, Redis has been one of the most popular database technologies among developers.
Companies like Niantic, HackerRank, Airbnb, Hulu, and OpenAI have used Redis to improve performance and reliability during high-traffic events, build and consolidate unified data platforms, and support scalable microservices environments.
In 2024, Redis shifted its licensing plans to offer Redis Community Edition for free and offer the Redis source code through dual-source-available licensing. In reaction, some corporate contributors, such as Amazon AWS and Google Cloud, forked Redis and created Valkey, an open-source alternative to Redis.
Though there’s currently significant overlap between Redis and Valkey, the two organizations have different teams and customer bases. And though they were both once based on the same code, they’ve begun to diverge in feature sets, functionality, capabilities, contributors, and more. Over time, the divergence is only likely to increase.
Companies comparing the two need to consider each product offering as it stands today and examine the precedent and trajectory each product has established so far.
Valkey is an open-source fork of Redis version 7.2.4. Redis and Valkey are fundamentally similar: Both offer an in-memory data store and cache that developers can use to store data as key-value pairs in RAM, allowing for processing speeds that far exceed the speeds of databases that access data via a disk or solid-state drive.
In March 2024, Valkey launched with the support of The Linux Foundation, and in September 2024, Valkey announced Valkey 8.0, its biggest departure yet from the forked Redis code. As time goes on, the two codebases will only diverge further.
As Kyle Davis, a maintainer for Valkey, said in April 2024, “From this point forward, Redis and Valkey are two different pieces of software.”
Valkey, like Redis, provides an in-memory cache that provides high-performance queries with low data latency. As a result, Valkey provides some of the same features and suits some of the same use cases.
Valkey, however, lacks all the features that Redis provides—features that help developers use Redis as an operational database.
Valkey has a range of use cases and is especially useful for organizations with use cases that require low latency or real-time functionality.
Session caching: Valkey can cache frequently-accessed data client-side inside the application memory, which allows for even faster queries and reduced load on other databases. An application could use session caching to store recently viewed files, for example, allowing users to access them faster.
Message queuing: Valkey allows messages to be queued so that organizations can send different messages to different users in the correct order. If an organization was building a notification system, for example, Valkey would help it broadcast notifications to particular channels and users.
Real-time updates: Valkey, due to its high-performance, low latency design, can support systems that require real-time updates. If an ecommerce merchant needs a database to handle high volumes of traffic and transactions, Valkey can support real-time inventory updates so that customers always have an up-to-date sense of what products are available.
Like all tools, Valkey has pros and cons. Companies comparing Valkey and Redis will need to identify the pros and cons that matter most to their use cases and step back to see the big picture of both offerings to make smart decisions.
Due to the current overlap between Redis and Valkey, some of the value propositions are shared. With the release of Valkey 8.0, however, there are a few differentiating features.
For example, Valkey 8.0 enhanced I/O multithreading, which improved throughput to similar levels offered in enterprise-grade Redis. Valkey now offers improved multi-core utilization and asynchronous I/O threading, which offers a level of throughput that’s over three times greater than its previous version.
Valkey also includes experimental Remote Direct Memory Access (RDMA support), which Redis doesn’t natively support. RDMA allows computers to complete more efficient data transfers with low latency and high bandwidth by directly accessing each other’s memory.
RDMA support potentially means greater performance improvements—in the narrowly-specialized environments that benefit from RDMA—but Valkey only supports the feature at an experimental level for now.
Since Valkey is based on Redis, the primary cons organizations will want to consider when comparing the two come from the precedent Redis has set in its years of operation as compared to the relatively untested nature of Valkey.
Feature parity: Valkey is based on Redis 7.2.4, but because Valkey is a separate organization, it will not necessarily keep up with Redis updates as they come out. Already, among other features Valkey doesn’t support yet, Redis provides:
Redis 7.4, for another example, added the ability to use Redis as a session store with hash field expiration, allowing developers to unlock searchable sessions, which they can now use for multi-value indexing and to index at arbitrary depths.
Support: Valkey currently doesn’t offer operational support, a key way to get the most out of a database tool without overburdening a company’s developers.
A recent study from Percona showed that while many organizations are evaluating Valkey, 76% of them will likely need third-party enterprise support to operate it. Redis, in contrast, provides robust support plans for developers, businesses, and enterprises.
Developers who prefer reading and referencing documentation will also find much more extensive information on the Redis docs site.
Clients: Most developers use a client to interact with Redis, such as Jedis, node-redis, redis-py, NRedisStack, and Go-Redis.
These clients are designed, developed, and maintained for Redis, so compatibility with Valkey—much less full-feature support—now and in the future isn’t assured. Clients supported by Redis officially also ensure end-to-end service with trust and continuous support from a single vendor.
Established familiarity among developers: Valkey is still very new and doesn’t yet have the familiarity among developers that Redis has developed over its many years in operation.
In StackOverflow’s 2024 developer survey, for example, Redis was one of the most popular databases. 22.8% of professional developers reported completing extensive development work with Redis in the past year, 67% of developers admire Redis, and 28% of users reported wanting to use Redis in the year after this survey was taken.
Vicki Boykis, a machine learning engineer at Mozilla.ai, characterizes much of this sentiment, writing, “I love Redis with a loyalty that I reserve for close friends and family and the first true day of spring because Redis is software made for me, the developer.”
In 2024, Salvatore Sanfilippo, the original creator of Redis (better known as “antirez” in the developer community), returned as an evangelist and, in his words, a “bridge between the company and the community.”
“With the new license, the spotlight can be back on the Redis core,” he wrote, “with new, exciting features in the hands of the developers around the world.” Redis’s foundation is strong, its focus renewed, and its future full of possibility.
Redis and Valkey are based on the same code, so at first glance, the similarities between the two products will stand out. At a high level, both are similarly good fits for most companies looking for an in-memory cache.
Redis, however, offers JSON, TimeSeries, probabilistic, search and query via Redis Query Engine, vector search, RedisGears, and RedisConnect, all of which come together to help developers turn an in-memory cache into a full database tool. If a company wants to build a key-value database or store, for example, it would need this unique feature set that only Redis offers.
As Rachel Stephens, a senior analyst with RedMonk, writes “One reason for Redis’ popularity is its flexibility. While intended as a cache, developers have also used it as a key-value store, a database, and a message queue. It has grown to fill a variety of niches.”
—*
Redis remains a popular, trusted, and flexible option for development teams looking to build and deploy faster, more powerful apps. Try it for free today.
