API
< Back to all clusters
[TECHNOLOGY] · 2 sources

started · updated

Valkey releases version 9.2 with AI inference and memory optimizations

Valkey, the Linux Foundation-backed open-source datastore, has released version 9.2, introducing several architectural improvements aimed at reducing operational costs and memory usage. A major feature is the introduction of forkless snapshots, which utilize a background thread instead of a separate process. This change prevents the temporary memory spikes and doubling of RAM usage typically associated with traditional copy-on-write snapshots.

The update also optimizes large sorted sets using a B+ tree-based architecture, resulting in up to 105% faster updates and 36% to 43% better memory efficiency for large datasets. For replication, new continuous stream compression can reduce cross-region traffic by 50% to 65%, while LZ4 compression can shrink backup files by up to 80%.

Additionally, the release includes new capabilities for AI inference, such as distributed key-value caching designed to reduce GPU time and costs. The update also introduces Path Hash, a native data type for efficient hierarchical lookups, which can assist AI systems in reusing previously computed context.

Entities

AWS · Alibaba Cloud · Linux Foundation · Madelyn Olson · Valkey