Performance Benchmarking Report

Performance Benchmarking

Performance Benchmarking Database Platforms and Cloud Native Storage Solutions for Kubernetes

Kubernetes is rapidly becoming the de-facto standard for efficient and resilient application deployment. In common with all infrastructure frameworks, Kubernetes needs persistent storage and in particular storage that is managed from the cluster and applications directly.

Container-native storage has emerged as a new category of software-defined storage where the traditional features of data storage (resiliency, data protection, scalability) are built into containers running within a Kubernetes cluster itself.

Performance and the efficient use of resources are critical in achieving scalability with container-native platforms. The inefficient use of storage hardware translates directly into additional costs that compound as systems scale. Therefore container-native storage needs to deliver high performance with low latency and minimal overhead – while still providing data integrity.

In this report done by Architecting IT, they conducted a series of tests to compare a range of commercial and open-source container-native storage solutions, using three common database platforms - PostgreSQL, MongoDB, and Redis against three metrics:

  • Latency
  • Bandwidth
  • Throughput

Download Report

Ondat Outperforms All Container-Native Storage Solutions

Of the five platforms tested in this paper, Ondat achieves the best results across all the tests. The platforms tested were:

  • Ondat v2.9.0 – commercial software from Ondat (formerly StorageOS).
  • Portworx Enterprise v2.11 – commercial software from Pure Storage. 
  • OpenEBS/cStor v3.3.0 – an open-source initiative initially developed and supported by MayaData. 
  • Rook/Ceph v1.9 – an open-source solution that uses Ceph as the storage platform and Rook as a management and presentation interface. 
  • Longhorn v 1.3.2 – an open-source storage solution developed by Rancher Inc, now part of SUSE.

“Our research concludes that Ondat delivers the most efficient performance for throughput and latency, compared to all the competitor solutions in this test. Ondat delivered the best overall performance but crucially also had the most consistent response times when looking at the outlier data.” said Chris Evans, Founder and Principal Analyst of Architecting IT.

PostgreSQL 

In the Postgres testing, replication at the volume level is used to protect data, performing tests with and without replication. There are two graphs for Postgres, showing latency in milliseconds (smaller numbers are better) and throughput in transactions per second (larger numbers are better).

1.Chart-1

1.Chart-1