E.V. Bolgova, A.V. Bogacheva, M.A. Balakhontseva, K.V Knyazkov, A.V. Dukhanov, S.E. Khoruzhnikov

This paper describes the approach to organization of virtual laboratory practices (VLP) using the cloud computing technologies within the model of AaaS (Application as a Service). This model chosen within the development of inter-disciplinary areas and relevant educational programs, including area of Computational Science. It provides the development and use of composite applications (CA) - a set of oriented to decision of whole problem interacting cloud services. The most common approach to the representation of the CA is the formalism of the workflow (WF), which allows you to describe the relations between the particular operations in a distributed environment as a directed graph. A composite application is created by linking existing software packages into a complex application that runs in a "cloud" of distributed resources. This paper proposes the set of learning models using cloud-based VLP on the base of promising model AaaS. They were created in the class of directed acyclic graph, whose nodes are data-processing units with the input and output parameters, and arcs are the relations between the nodes. For example, unit may be launch of the package, procedure, unit to vary parameters, conditional branching, cycle. There are proposed learning models with using hidden WF and open WF without editing ability, as well as with the ability to changing open WF and debug it. Also, the model for verification of learning is presented. These models are analyzed in the context of evaluate of descriptors of knowledge level, skills and personal qualities. The end of the paper presents the software tool to automate the creation and configuration VLP (with use the proposed models), based on cloud computing platform CLAVIRE.