A Composable Framework for Multiple Resolution, Cross-Layer Modeling and Simulation

We propose an agent-based framework and simulation infrastructure for the integrated modeling and simulation of multiple layers that is applicable towards the simulation of large, complex networked system of systems. This framework enables a process and methodology of iterative modeling and refinement and progressive drill down into additional detail during various parts of the system lifecycle. The key innovation of this approach is using agent-based, dynamically composable simulation model elements that can be assembled within and across layers to create multiresolution, cross-layer simulations. In the Phase I effort, we will identify key application layer and network layer abstractions that will enable the creation of reusable simulation elements, executable models, and components. The core computational model of the simulation approach is based on a proven agent-based simulation framework that leverages a distributed, parallel discrete event simulation system that has been successfully applied to large-scale cross-layer models. Additionally, we will devise a general metrics collection framework that will allow users to flexibly collect and utilize results. The approach will be demonstrated in the context of a use case consisting of a distributed system developed using the service oriented architecture paradigm.