Virtual private supercomputer is an efficient way of conducting experiments on high-performance computational environment and the main role in this approach is played by virtualization and data consolidation. During experiment virtualization is used to abstract application from underlying hardware and also from operating system offering consistent API for distributed computations. In between experiments data consolidation is used to store initial data and results in a distributed storage system and offers API for distributed data processing. Combined, these APIs form solid basis of a distributed system shifting user focus from supercomputing technologies to problem being solved.