1-Would it be more efficient to store objects that contain its relations if you only do (simple) read operations? (e.g. JSON database)

2-Instead, do graph DB engines try to break through bottlenecks for big data and analytics scenarios?