Questions about efficiency, costs and usage of Spice vs. Direct Query in QuickSight

Hello everyone!

I’m researching query methods in Amazon QuickSight and would like your help. I have a few specific questions:

  1. Which is more efficient in terms of performance and cost between Spice and Direct Query? In which scenarios should each one be used?

  2. What are the types of charges associated with Direct Query? In addition to the cost per query, are there additional charges such as Data Transfer?

  3. Regarding incremental updates in Spice, how does it handle operations such as DELETE? Is there a mechanism similar to a crawler that maps the data schema?

  4. If I need to update a table frequently as a full update, is it still worth using Spice?

I appreciate any contribution that can clarify these questions!

Hi @ArthurMendes

SPICE (Super-fast, Parallel, In-memory Calculation Engine) is Amazon QuickSight’s in-memory engine designed for faster performance by preloading data.

SPICE is more efficient for performance because it stores data in memory, resulting in low-latency queries and better scalability for large datasets. It’s cost-effective if your queries are run frequently and you want to avoid paying for every query separately.

Direct Query connects to your data source in real-time, allowing up-to-date access without loading data into SPICE. However, this method can be slower, especially with complex datasets or if the source is under heavy load. Costs can accumulate due to both query execution and potential data transfer charges.

SPICE supports incremental updates, which allow you to append or refresh parts of your dataset without reloading the entire dataset.

  • If you need to remove data or handle deletes, you may have to re-import the dataset (full refresh) or apply logic in the dataset transformation stage to exclude unwanted records.
  • SPICE does not have an automatic mechanism like a crawler to map schema changes. You need to manually update the schema in QuickSight dataset.

SPICE is suitable when you have large datasets that are frequently accessed in analysis and dashboards, benefiting from pre-aggregation and faster reporting.

Direct Query is suitable when you have smaller datasets and require real-time reporting where data freshness is critical.

Please refer the below documentation this might be helpful for you.

QuickSight pricing - Business Intelligence Service – Amazon QuickSight Pricing – AWS

2 Likes

Hello, @Xclipse !

I want to thank you so much for your contribution, it was amazing! I really appreciate you sharing your knowledge.

By the way, do you know where I can find information about the amounts charged from Direct Query?

Thank you very much!

Hi @ArthurMendes, thank you. As for pricing, it depends on the underlying services such as Athena or other intermediaries. I would recommend the following:

For Athena:

Or check out the general pricing calculator:

Also, please reach out to your account manager/billing, they can help you navigate the pricing structure.

1 Like