The maintainers of the open source Valkey database this week made available an update that makes it simpler to deploy multiple instances on the same clusters as part of an effort to consume memory more efficiently.
In addition, version 9 of the key-value database adds an atomic slot migration capability that uses a snapshot-based migration process triggered by a single command to shard data across multiple databases. A child process first forks and streams data incrementally from the source to the target node, allowing both nodes to remain active throughout migration.
Once all data is migrated, Valkey performs an atomic handoff through which clients are redirected instantly to the target node without downtime or errors that previously might have been encountered using traditional slot migration techniques.
Finally, Valkey 9 also adds support for expiration dates for hash fields in addition to improving memory prefetching for pipelining commands by up to 40% higher throughput and 20% higher throughput for zero-copy responses for large requests.
Originally launched to provide an in-memory database alternative to the widely used Redis database following changes made to the way that platform is licensed. Now being advanced under the auspices of the Linux Foundation, the Valkey project now has eight maintainers and contributors from more than 150 organizations, including Aiven, Alibaba Cloud, AWS, Bytedance, Ericsson, Google Cloud, k0rdent, NetApp Instaclustr, Oracle, Percona and UpCloud.
Madelyn Olson, a Valkey project maintainer and principal engineer for in-memory databases at Amazon Web Services (AWS), said Valkey 9.0 can support more than one billion requests per second. Compared with version 8.1, Valkey 9.0 provides up to 40% more throughput to significantly improve productivity, she noted.
Longer term, maintainers of the project are also working toward providing tighter integration with large language models (LLMs) and an ability to take fuller advantage of solid state disks (SSDs) to run larger databases, said Olson.
It’s not clear how widely Valkey has been adopted as an alternative to Redis, but with more data than ever needing to be processed in real time there is little doubt that AWS, Google and Oracle are pushing organizations toward an option that costs them less to deploy. The challenge, of course, is that Redis databases are already deeply embedded within enterprise IT environments that rely on it to improve application performance by caching data in memory.
Regardless of approach, organizations of all sizes are finding the need to reduce application latency to enable near real-time application experiences, which in turn creates a need for yet another type of database for IT teams to manage alongside existing databases. The obstacle, however, is that the number of database administrators and application developers that are familiar with multiple database platforms is still fairly limited.
Hopefully, in the age of agentic artificial intelligence (AI) managing a diverse portfolio of databases will become simpler. In the meantime, however, the skills required to optimize database performance continue to be highly sought after at a time when there has never been a greater appreciation for performance.
