Search

Distributed hybrid simulation is an approach to large-scale testing in which the system under test is split into several sub-structures which are tested or simulated in different locations. Data are passed between the sub-structures at each timestep so as to ensure that the distributed experiment realistically simulates the full system under test. This approach optimises the use of resources at different locations to achieve a more representative experiment. While different software to conduct distributed simulations exists, there are no standards and specifications to organise and plan the experiments, and as a result the different systems lack inter-operability. To address these issues, we have developed a high-level specification called Celestina, which provides a framework for conducting a distributed experiment. Celestina specifies the services to be implemented, under three main headings of networking, definition and execution, and supports the data exchange during a simulation. It does not force any particular implementation or method of data exchange. This paper summarises the Celestina specification and describes one implementation. Lastly, a validation experiment is presented, involving distributed numerical simulations of an earlier local hybrid experiment, in which Celestina controls the experiment planning and data exchange effectively and with minimal computational overhead. [ABSTRACT FROM AUTHOR]