Table of Contents
- From Open Source to Serverless
- Cassandra TCO Test Setup
- Test Results
- About DataStax
- About William McKnight
- About Jake Dolezal
This study examines the full cost and true value of Cassandra self-managed on Google Cloud (GCP) and the cost of a fully managed serverless Cassandra service. We included dedicated compute hardware (for self-managed Cassandra), cost per read and write operation (on serverless Cassandra), storage growth (each write operation adds new data) and people cost in our three-year total cost of ownership calculations. People costs take into account that certain capabilities in serverless Cassandra needed for the workload were not available in self-managed Cassandra, requiring workarounds. We used market rates and typical splits of full-time equivalent (FTE) and consulting to determine our people costs. The production staffing FTE for serverless Cassandra was .2, for example.
A realistic performance test based on a use pattern relatable to a modern enterprise was established using NoSQLBench, a performance-testing tool for the NoSQL ecosystem. This established the basis for our serverless Cassandra pricing and helped determine the configuration of the self-managed platform.
We selected a 15-node self-managed Cassandra cluster of n2-highmem-16 virtual machines using the Ubuntu 18.04 operating system and installed the latest version of OSS Apache Cassandra for our test. Since serverless Cassandra is fully managed, there was no configuration to select.
The use case was built on a consumer application, which typically experiences a peak and a trough of activity each day. In our workload, on average the peak is around midday and the trough around midnight. For simplicity, and given the ongoing nature of online commerce and business activity today, we also assumed this usage pattern continued 7 days a week, 365 days a year.
For serverless Cassandra, we used the best pricing available during the initial public release of the DataStax Astra serverless offering. For self-managed Cassandra on GCP, we considered the cost of an enterprise-grade deployment, including two 15-node production clusters (for disaster recovery) and one 15-node development/test cluster. We also provisioned twice the amount of projected storage needed.
Our final three-year total cost of ownership figures for the study showed $762,430 for serverless Cassandra and $2,988,654 for self-managed Cassandra—making serverless Cassandra 75% less expensive than self-managed Cassandra in our test. We found serverless Cassandra to have 95% fewer production costs, 3 times less migration complexity, and 22% lower infrastructure costs than Cassandra.
This test shows the immense value of choosing a serverless Cassandra service to deploy Cassandra for an enterprise project.