New ScyllaDB Release Achieves 50% Higher Throughput, 33% Lower Latency

SUNNYVALE, Calif, February 20, 2024 — ScyllaDB, the database for predictable performance at scale, today announced its latest release, ScyllaDB Enterprise 2024.1.0.  The new release introduces significant performance improvements: up to 50% higher throughput, 35% greater efficiency, and 33% lower latency than the previous enterprise release.

In benchmarks, ScyllaDB 2024.1 achieved more than 1.5X higher throughput compared to 2023.1. In some cases, this can translate to a 35% reduction in the number of vCPUs required to support a similar load. This enables a similar reduction in vCPU cost. Additionally, tests achieved 33% lower latency (both mean and P99), even with the higher throughput. Setup details and results are available in this blog.

ScyllaDB Co-founder and CEO Dor Laor introduced this performance boost at the recent ScyllaDB Summit, which hosted a community of over 7,500 developers, engineers, and database enthusiasts. Laor explained, “We achieved this performance boost through profile-guided optimization. The new performance level keeps ScyllaDB at the forefront of fast and efficient databases; it’s ideal for business-critical workloads that require ultra-low latency, even beyond 1M ops/sec.” Laor’s keynote, which also introduced ScyllaDB’s revolutionary “tablet” architecture, is now available on-demand alongside 30+ tech talks by ScyllaDB users such as Discord, Expedia, Paramount, and more.

Also in this new release:

• Repair-Based Node Operations provides a more robust, reliable, and safer data streaming for node operations like node-replace and node-add/remove.
• Extended Encryption at Rest allows use of Amazon Key Management Service (KMS) to store and manage keys and setting default EaR parameters for all cluster tables.
• Extended Guardrails framework protects ScyllaDB users and admins from common mistakes and pitfalls related to replication.