Agent behavior model for distributed computing management in the environment with virtualized resources

We address the management of jobs generated by scalable applications in an environment that integrates Grid system and cloud infrastructure. The complexity of the management is high due to the differences of the cloud and Grid computing models. There exist conflicts between the preferences of owners and quality criteria defined by users. Multi-agent systems achieve a significant success in solving this problem. The agents are intelligent software entities endowed with rights and responsibilities for servicing computational processes. They interact to achieve their goals. A representation of the needed knowledge and designing agent functions taking into account specifics of applications is a non-trivial problem for the known multi-agent systems used for managing distributed computing in practice. We propose a multi-agent system to manage jobs in a heterogeneous distributed computing environment that integrates Grid and cloud computing. To this end, we implement new specialized models and tools for the agent designing and operation. The model is based on the integrated use of conceptual and finite state machine programming paradigms. We demonstrate the advantages of our system on an example of the job management of a scalable application to solve optimization problems of warehouse logistics.