T.N. Churov - Junior Research Scientist, National Research University of Information Technologies, Mechanics and Optics
D.A. Nasonov - Junior Research Scientist, National Research University of Information Technologies, Mechanics and Optics
E.V. Bolgova - Junior Research Scientist, National Research University of Information Technologies, Mechanics and Optics

The article offers a new approach to organize the system of cloud computing of the second generation - a personal cloud model. Not like other cloud computing technologies, while performing the tasks of the user this model dynamically generated a single cloud space, which includes the services of his or her local computer and remote cloud resources of other providers. This approach allows us to run partial-private distributed applications and maintain all the standard features of control systems with complex scientific applications. In addition to the description of the general concept of this technology, the article presents the results of its implementation and application in the cloud platform CLAVIRE. The personal cloud technology is based on the generation of simple and convenient mechanisms of integration of computing resources owned by the user into a common pool of clouds - resources. At the same time it supports the ability to embed a local personal computers as well as special high-performance systems. It describes the features of software packages and prescribes their launch modes when embedding a resource, in addition to standard data access. Package descriptions are also grouped in a special register and presented to the user for editing. Apart from that, the system supports the functions of access control, provenance of the progress of composite applications, and pricing of the custom resources - use. To comply with this concept in the platform CLAVIRE there were developed new components, such as resource base, resource account component, and provenance component. To test the personal cloud model technology, a high-performance software complex to simulate the conformation-dependent properties of proteins for rational design of medications has been implemented on its basis. On the basis of this complex the experimental studies of the effectiveness of the personal cloud model were implemented. These experiments revealed the time costs, but they are still comparable with the overall scale of variability while working in the cloud.