S3 is not a database, it is only a storage system. It lacks a database engine and many storage features, but it can be pretty useful for saving media files such as photos, videos, and music.
This approach is already very popular. For example, if you develop an application that uses a MongoDB database, you could use MongoDB GridFS to store large binary data. However, the most efficient solution is to offload this kind of data to cloud services because the machines responsible for your database are usually the most expensive ones. It means that the cost per gigabyte in a database is usually higher than a cloud storage service, such as S3.
In our serverless store, we are storing the product images in SimpleDB/DynamoDB as string fields. Instead of saving the full binary data, we save just the URL of the image file. Example:
https://s3.amazonaws.com/serverless-store-media/product-images/lonely-bird.jpg
When we receive this information in the frontend, the <img> element...