The main problems with existing RDMS are-

• Have fix schema : RDBMSs are more suitable for structured data, such as weblog, sensor and financial data.
• You loose lots of important info which doesn’t fits in schema : Schema is fixed, we have a policy of “Define Schema first, then Data”
• Expansion of schema is not easy
• Growing amount of unstructured data which doesn’t fit there
• Let’s rid away from fix schema – Flexible storage format – Raw form

Most current generation RDBMSs Online Transaction Processing (OLTP) were originally written for
shared memory multi-processors in the 1970,
those include a suite of features —

• disk-resident B-trees and heap files,
• locking-based concurrency control : Suitable where we transactions are going on
• support for multi-threading