ScyllaDB Benchmarks

Selecting the right database for your application is an important task. At ScyllaDB we constantly test Scylla against Cassandra for throughput and latency. Database latency determines your application speed, while database throughput defines how many users your application can service. ScyllaDB uses various testing tools such as industry YCSB benchmarks and dedicated Cassandra Stress tests.

4-node Scylla 2.2 (i3.metal) vs 40-node Cassandra 3.11 (i3.4xlarge)

  • 2.5X reduction in AWS EC2 costs
  • Dramatic improvement in availability through 10X reduction in cluster size
  • Up to 11X improvement in 99th percentile latency
  • Amazon EC2 Bare Metal Instances are a great platform for Scylla

Throughput Results of Scylla 1.6 vs Cassandra 3.0 Using Amazon Web Server i2.8xlarge

  • Amazon Web Servers are a great platform for Scylla

  • Scylla uses your resources on AWS in the most efficient way

  • Scylla outperforms Cassandra on small or large instances, using SSD or EBS based volumes

  • Applications using Scylla can perform 10x better and are more reliable

Throughput Results of Scylla 1.6 vs Cassandra 3.0 Using Amazon Web Server c3.2xlarge

  • Scylla’s compelling advantage can be seen even in EBS-backed small instances. 21% – 145% gain on top of Cassandra

  • Worry-free compaction time

  • Up to 10 times lower 99.9% latency

Throughput of ScyllaDB vs Cassandra Using 2TB Data for Different YCSB Workloads

  • Scylla 1.4 vs. Cassandra 3.9

  • Scylla outperforms Cassandra by a factor of x37

  • Samsung NVMe SSDs drive Scylla’s latency lower than and in-memory workload on Cassandra

  • Scylla handles a data set ten-times bigger than Cassandra with comparable or better performance

Take it for a Spin. See the performance you can expect from Scylla. Our Scylla Test Drive demo runs a live stress test against the requirements you set.


Let’s do this

Getting started takes only a few minutes. Scylla has an installer for every major platform. If you get stuck, we’re here to help.