You forgot ... Most likely, it will be much faster (but it will depend on your model. If you have a simple model, I recommend just stick to RDB).
If you have many-to-any, recursive relationships, nested subgraphs, etc ... then if it will be much faster.
Of course, it will also work on much less hardware. Research at IBM shows that you will save about 50% of the CPU at your mid-tier (for example, you are web servers where all type translations occur for RDB).
In addition, you do not need a database expert. If you are an OO guy, you already have a set of skills ... especially if you know Hibernate / nHibernate, because the lifecycle concepts of an object from Hibernate and ORM space were mainly captured from the OODB world and transferred to the RDB camp.
In addition, you can iterate much faster in development cycles without having to constantly contact the database administrator and request schema updates.
.... RDB will eventually switch to mainframe. Everything around, but not used for new projects. These Cobalt programmers didn't believe at this time ... so don't worry with all the speakers.
source share