database design – Why big data analytics over peta-byte datasets leads to NoSQL not RDBMS

Computational and storage requirements of applications such as for big data analytics. business intelligence and social networking over peta-byte datasets have pushed SQL-like centralized datasets to their limits. This led to the development of horizontally scalable, distributed non-relational data store, called NoSQL databases.

According to the literature, it is because that the current relational data model has deficts and problems in the modeling of data and constraints of horizontal scalability over several servers and big amounts of data.

The only reason I can figure out is that NoSQL supports unstructured data. What about other reasons? When it comes to large scale distributed databases, why the choice of NoSQL is almost always prefered over RDBMS?