Need data model advice for an AWS app

zumpoof used Ask the Experts™
I'm building a news application. Each article can be tagged with words like "Politics", "Food", "San Jose, CA". I'd like users to be able to select a tag, and see all related articles. What would be the recommended way of storing this data?

My instinct is to use S3 for each news article and tag, and then to use DynamoDB to store the relationship of tag to article (edges). So, when a user wants to view articles for a tag, I'd retrieve the S3 object for that tag, do a DynamoDB query for the ID's of articles associated with that tag, and then retrieve the S3 objects for each article.

Does this sound like a good plan, or is there a better way?
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
IT Consultant
i think it depends on how S3 tags are used to classify an article and its content.

say if search criteria are limited to article catigories only, it will be rarely having more than 100 categories, commonly about 20. in this case, a tags-only solution might be simple yet working well.

say if search critrria are for randomly defined keywords or attributes, especially one attribute may have multiple related attributes depending on the content of an article, then the tagging mechanism and query APIs of S3 can't handle that in an efficient way. in this case, DynamoDB is a good choice to manage the attributes, keywords, relationships and their index.

actually S3 plus DynamoDB is one of the best practices in managing metadata for S3 based BOLB data. below AWS blog is a very good reference and just for your scenario. please review.

hope it helps,

Do more with

Expert Office
Submit tech questions to Ask the Experts™ at any time to receive solutions, advice, and new ideas from leading industry professionals.

Start 7-Day Free Trial