Linux Kernel & OS

Valkey 9.2 adds Zstd compression, B+ tree sorted sets, and forkless replication

Key Takeaway

Valkey 9.2 introduces Zstd compression for persistence and replication, B+ tree sorted sets for reduced memory overhead, and forkless snapshotting as the Redis fork continues to outpace alternatives.

Summary

Valkey 9.2 has been released as the latest feature update to the prominent open-source fork of Redis, bringing substantial performance and efficiency improvements to the in-memory key-value database. The release introduces B+ tree sorted sets that significantly reduce memory overhead, Zstd compression for RDB persistence and replication that achieves better compression ratios than the existing LZ4 implementation, and forkless full synchronization to replicas that eliminates the performance penalty of fork-based operations.

The release also adds server-side hot key detection for improved operational visibility, a new DELEX command for conditional string key deletion, and dozens of bug fixes. This launch comes shortly after the competing Redict project was deprecated in favor of Valkey, consolidating the open-source Redis alternative ecosystem around Valkey as the primary community-driven fork.

Storage engine and replication changes in Valkey 9.2

  • B+ tree sorted sets: Replaces the previous sorted set implementation to deliver significant reductions in memory overhead for datasets using this data structure.
  • Zstd compression: Replaces or supplements LZ4 for RDB persistence and replication streams, providing greater compression ratios according to upstream merge request benchmarks.
  • Forkless snapshotting and replication: Eliminates the fork() system call overhead for creating snapshots and performing full synchronization to replicas, reducing latency spikes and copy-on-write memory pressure.
  • DELEX command: New conditional deletion primitive that removes a string key only when its current value matches or differs from a specified comparison value.
  • Server-side hot key detection: Adds runtime visibility into frequently accessed keys without requiring external monitoring tools or client-side instrumentation.

Why it matters

For Linux distributions and cloud operators running Valkey as a Redis replacement, the forkless replication and Zstd compression reduce operational complexity and infrastructure costs while improving replica synchronization performance. The deprecation of Redict in favor of Valkey further centralizes community engineering effort around a single open-source fork, simplifying packaging decisions for Debian, Fedora, and container maintainers who previously had to evaluate multiple Redis alternatives.

What to do

  • Update to Valkey 9.2 to get Zstd compression for RDB persistence and replication, which offers better compression ratios than LZ4.
  • Enable B+ tree sorted sets in Valkey 9.2 to reduce memory overhead for sorted set data structures.
  • Use forkless snapshotting and replication in Valkey 9.2 to eliminate fork() overhead and reduce latency spikes during full synchronizations.
  • Use the new DELEX command for conditional string key deletion when the value matches or differs from a specified value.

Original reference: www.phoronix.com →