Redis wants to offer real-time data with AI

Redis wants to offer real-time data with AI

Redis Labs has announced its long-term vision for the company. The company wants to create a real-time data platform that can serve as the foundation for all kinds of modern applications.

At the RedisConf 2021, Redis talks about a vision where integrated databases are more consistent, offer latencies of less than a millisecond and are augmented with artificial intelligence. The company wants its users to focus more on creating features for their apps that provide value to customers, while giving them a better understanding of the transactions and operations that run within their organisations. On that basis, customers should be able to make better decisions.

High-performance database management

Redis is the developer of the open source database management system of the same name. The system runs in memory and is known for its high performance. It can serve as a database, but also as a cache or a message broker. Further features include built-in replication, a lightweight scripting language, the ability to remove the least recently used cache and customisable levels of on-disk persistence.


Higher consistency has already been achieved to a large extent with the release of RedisRaft. This deployment option will be part of the upcoming 7.0 version of Redis. Users will not have to make a compromise between strong consistency or eventual consistency. It should also offer greater scalability and performance.

Yiftach Shoolman, the founder and CEO of Redis, tells SiliconAngle that with RedisRaft, Redis has managed to pass the Jepsen Test. This is an industry standard that measures the consistency of distributed databases. No problems were observed, while high performance was also guaranteed. According to Shoolman, no cloud database provider has yet passed this test.


This, in combination with other improvements to the RediSearch search engine, ensures that RediSearch works much faster than databases such as MongoDB and Elastic, by a factor of 3 to as much as 37 percent. These data models can now also be deployed locally so that they can be physically closer to the end-user. This should reduce latency even further.


Finally, Redis introduced RedisAI. This is a feature store that should make it possible to serve AI models closer to the features. This should simplify the architecture and improve the performance of AI applications by a factor of 10 to 100.


RedisAI is available for on-premises deployments as of today. In the second half of the year it will also be released for Redis Enterprise Cloud. RediSearch is in private preview and will become generally available in the second half of the year.

Tip: Redis is now the most popular database within AWS