DynamoDB Dynomite Essentials

Important: These docs are for the outdated Jets 5 versions and below. For the latest Jets docs: docs.rubyonjets.com

Here are some DynamoDB essentials to help you learn how to use Dynomite better.

What is DynamoDB?

DynamoDB is a NoSQL database.

  • It does not rely on a traditional relational data model
  • It stores data in a key-value format
  • It’s schemaless

Thoughts

DynamoDB works differently from a traditional SQL database server. Schemas do not have to be defined ahead of time. A lot of people emphasize that this is a benefit. In some ways, I agree, but I also think there are some cons.

Pros:

  • You can “add” columns on-the-fly by simply inserting data with the new columns. Since there is no schema, you don’t have to worry about migrations to add a column to a table.
  • Adding columns on the fly is a huge benefit for large databases. However, most web apps are not large enough to worry about it.
  • Adding indexes is an asynchronous operation that does not lock other database operations. However, adding indexes can take some time. See Wait Times.

Cons:

  • Adding a column is not a big deal for most small and medium size database sizes. Relational databases perform just fine, are more common, and are well understood.
  • Schemaless can be too loose. You can accidentally insert data with a “new column,” and it’s available. This can create messy data. The pro can also be a con. Dynomite helps prevent this with different undeclared fields behaviors.
  • To use the database in a performant way means you must use the query api call instead of the scan call. You must be explicit about it. Dynomite helps with this by being smart about querying by automatically discovering indexes and uses the query when possible.
  • Adding indexes can take a while, IE: 8m+ for small tables, and must be added sequentially.
  • Cost: AWS DynamoDB specifically can cost a lot of money. GSI indexes are essentially copies of the table. It can double the storage costs for that table. Also, provisioned throughput costs money. For small tables, using on-demand throughput can save money.

Even though I’ve listed more cons than pros here, DynamoDB can work quite well for your use case, and Dynomite helps improve the experience quite a bit.

In this section: