The talk is based on my original article that I wrote back in 2011 and followed up with friend feed case study. We will do a walk through history of HSTORE, and how we can now use JSONB support in PostgreSQL, discuss what makes it enticing, and comparable to NoSQL stores like MongoDB or CouchDB.

We will look at how a FriendFeed clone could have been designed today using basic document store (JSON documents), build indexes on these deep nested fields for our various queries, and use partitioning features to distribute the data. The best part about all this is you get all of that schema free storage/lookup benefits, with solid ACID guarantees that other NoSQL stores can't match.