DynamoDB on-demand and provisioned capacity - Serverless Applications Lens

DynamoDB on-demand and provisioned capacity

Often within your Serverless applications you will use DynamoDB, Amazon’s fully managed NoSQL database service. Amazon DynamoDB has two pricing models, pricing for on-demand capacity mode and pricing for provisioned capacity mode.

With provisioned capacity you pay for the provision of read and write capacity units for your DynamoDB tables. Whereas with DynamoDB on-demand you pay per request for the data reads and writes that your application performs on your tables.

With on-demand capacity mode, DynamoDB charges you for the data reads and writes your application performs on your tables. You do not need to specify how much read and write throughput you expect your application to perform because DynamoDB instantly accommodates your workloads as they ramp up or down.

With provisioned capacity mode, you specify the number of reads and writes per second that you expect your application to require, and you are billed based on that. Furthermore if you can forecast your capacity requirements you can also reserve a portion of DynamoDB provisioned capacity and optimize your costs even further.

With provisioned capacity you can also use auto scaling to automatically adjust your table’s capacity based on the specified utilization rate to ensure application performance, and also to potentially reduce costs. To configure auto scaling in DynamoDB, set the minimum and maximum levels of read and write capacity in addition to the target utilization percentage.

It is important to note that DynamoDB auto scaling modifies provisioned throughput settings only when the actual workload stays elevated or depressed for a sustained period of several minutes. This applies for scaling up or down the provisioned capacity of a DynamoDB table. In the case that you have an occasional usage spike auto scaling might not be able to react in time. This sometimes can be mitigated from DynamoDB burst capacity where DynamoDB reserves a portion of the unused provisioned capacity for later bursts of throughput. The burst capacity is limited though and these extra capacity units can be consumed quickly.

This means that provisioned capacity is probably best for you if you have relatively predictable application traffic, run applications whose traffic is consistent, and ramps up or down gradually.

Whereas on-demand capacity mode is probably best when you have new tables with unknown workloads, unpredictable application traffic and also if you only want to pay exactly for what you use. The on-demand pricing model is ideal for bursty, new, or unpredictable workloads whose traffic can spike in seconds or minutes, and when under-provisioned capacity would impact the user experience.

The optimal DynamoDB pricing model for your use case largely depends on the profile of your application. If the provisioned capacity model with auto scaling is the most suitable for you, we recommend to tune your target utilization for each workload to increase savings and ensure at the same time that capacity is appropriately allocated.