Percona published the results from their testing of the ScaleFlux CSD 2000 Computational Storage Drive for MySQL in their blog today: How can scaleflux handle mysql workload?
Percona published the results from their testing of the ScaleFlux CSD 2000 Computational Storage Drive for MySQL in their blog today: How can scaleflux handle mysql workload?
The testing compared the CSD 2000 with a leading NVMe SSD for Queries per Second (QPS), Latency and effective storage capacity across a variety of workloads, database sizes and database configurations .
Some of the key takeaways from the testing include:
- CSD 2000’s transparent compression/decompression feature out-does InnoDB table compression, both in data compression ratio and in database performance – you can save more space on your storage while achieving 3x the performance of host compression… cost savings without compromising on performance
- Running MySQL with CSD 2000 results in up to 90% lower latency and 2x higher QPS for mixed Read/Write workloads in comparison to using an ordinary NVMe SSD.
The blog gives an overview of the testing and results. Additionally, a White Paper is available with more details on the test parameters and outcomes, with a particular focus on the larger datasets and using the CSD 2000’s transparent compression to store 2x the data of ordinary drives. This white paper is available for download in the blog and here on ScaleFlux.com
