Keynote speech
Prof. Andrea D'Ambrogio
Associate Professor of Computer Science
Software Engineering Laboratory
Faculty of Engineering
Dept. of Computer Science, Systems and Production
University of Roma "Tor Vergata"
link to homepage
Model-driven Approaches for the Development of Distributed Simulation Systems
Model Driven Engineering (MDE) is gaining momentum as an effective approach for dealing with the increasing complexity of software development, leading to software systems of higher quality at largely reduced time, effort and cost. The main motivation behind MDE is to transfer the focus of work from coding to modeling, by treating models, generally defined as system representations at various levels of abstraction, as the primary artifacts of development. The MDE process takes as input an abstract model of the system and then applies a series of model transformations to refine the model with platform and implementation details and finally yield the code of the system as output. The OMG's Model Driven Architecture (MDA) is the most known and supported incarnation of the MDE approach.
Notwithstanding the commonly recognized effectiveness of MDE, there is a general lack of contributions that make use of MDE techniques in the development of distributed simulation systems, although the degree of system complexity is considerably higher than that of more conventional business-oriented software systems, due to the intrinsic concurrency, the required interoperability and the intricacies of currently available simulation-oriented infrastructures (e.g., HLA - the High Level Architecture).
This talk first illustrates the opportunities and the challenges of MDE in the development of distributed simulation systems and then investigates available and prospective technical solutions that exploit MDA techniques to first give a description of the system to be simulated (i.e., the system model) and then transform this description into an high-level simulation-oriented model, which in turn is transparently translated into simulation code. The positive effect is a significant reduction of both the development effort and the level of required expertise in simulation theory and practice, especially in those contexts where acquiring such expertise might be difficult. The presented concepts are illustrated by use of an example application that deals with the development of a distributed simulation system of an heterogeneous computer network.
