It’s important to understand the capacity and pricing for your ScyllaDB Cloud use case. You’ll find that ScyllaDB generally delivers better price-performance than other databases due to its close-to-the-metal, shard-per-core, shared-nothing design, which fully utilizes the hardware it runs on.
Stand Alone Pricing
You can use the ScyllaDB Cloud Calculator to obtain an estimation of pricing and performance for ScyllaDB Cloud. This data is reflective of basic assumptions of performance and is not to be used in place of rigorous capacity planning and proof-of-concept (POC) testing. (See “Known Limitations” below.)
Comparative/Competitive Pricing
You can also use the ScyllaDB Cloud Calculator to compare ScyllaDB Cloud to the following NoSQL databases:
- Amazon DynamoDB
- DataStax Astra
- Amazon Keyspaces
Note that while many features across these databases may be the same or similar across the different offerings, there will be instances where specific features are not available, or there may be similar features that vary widely in implementation. You are encouraged to contact us to find out more about your specific needs.
ScyllaDB strives to maintain current pricing on these competitors in our calculator. Yet if you find any discrepancy between our calculator and what you find on a competitor’s website, please let us know.
Reserved, Provisioned, or On-Demand Pricing
Pricing for cloud services varies greatly dependent on how you wish to consume them over time. So let’s define a few terms and describe the differences.
For ScyllaDB Cloud
- Reserved — this is when compute and storage assets — servers or nodes — are set aside in ScyllaDB Cloud exclusively for your use for 1 year.
- On-Demand — this is when compute and storage assets — servers or nodes — are set aside in ScyllaDB Cloud exclusively for your use for on an hourly basis.
For other Cloud Vendors
- Provisioned — Cloud vendors do not set aside specific assets, but contract for a certain performance capacity measured usually as certain throughput levels of reads or writes.
- Reserved — Cloud vendors may also provide reserved capacity, set-aside compute resources for larger workloads.
- On-Demand — Cloud vendors can also provide throughput as-needed. You pay for what you use.
Reserved or provisioned pricing is most appropriate to traffic that has regular, steady and predictive workload patterns. For example, if you have the same number of IIoT sensors reporting in at regular time intervals each day with the same amount of data each time. Even as you increase the number of IIoT devices, your traffic pattern grows predictively.
However, sometimes your traffic varies radically based upon stochastic events, or varies widely based on the time of the month or year. Perhaps this is a new workload, where you cannot predict the load it might see based upon lack of historical demand data. While you can project capacity, estimations might vary widely from actual requirements. For those sorts of workloads, on-demand pricing is more appropriate.
Known Limitations
The pricing we provide is based upon a set of assumptions for the calculator and cannot account for every possible feature or use case. Some examples of the features our ScyllaDB Cloud Calculator does not account for currently include:
- Lightweight Transactions (LWT) or other form of consistency/transaction mode
- Materialized Views and Secondary Indexes
- Change Data Capture (CDC) or streaming data features
- Full table or range scan performance
- Highly variable payload sizes
- Different classes of operations created with Workload Prioritization
Example Use Cases
ScyllaDB outperforms its competition while costing less in most but not all use cases. Let’s take a look at how it compares using a few specific examples. (Prices used in these examples are as of March 2021.)
Write-Heavy
ScyllaDB Cloud has a fast write path, making it perfect for write-heavy workloads. Write-heavy workloads can include time series data, or ingestion from many data points, such as for the Internet of Things (IoT). In this example we’ll have a 10% read / 90% write workload, with 4 terabytes of data stored and a typical 1kb payload.D
