Real Time Distributed Simulations Using SIMSAT Linux

The need for large-scale real-time simulations is growing in demand within many diverse industries. This places ever-increasing performance requirements on the hardware platforms required to run the simulations. Possible solutions to this include the use of networked machines to distribute the simulator processing load. This paper examines some recent examples of such simulations within the European Space sector, and how SIMSAT Linux simulation infrastructure has been extended to satisfy the requirements to run distributed real-time simulations. The underlying simulation environment used by the European Space Agency (ESA) and the European Space Operations Centre (ESOC) for spacecraft simulators since the early 1990's is the SIMSAT real-time infrastructure. SIMSAT originally ran under VMS, but since then it has been ported to many different Operating Systems, i.e. Windows NT/2000/XP and then most recently SUSE Linux 9.0. It provides the real-time environment for spacecraft simulators to verify the ground segment readiness prior to launch, including the validation of the mission control system, validation of spacecraft procedures and the training of the spacecraft operators. The requirement to develop simulations for large and complex space systems within the European Space Market, for example Galileo, places heavy demands on the simulation platform. This increasing performance demand has driven the development of a distributed version of the SIMSAT infrastructure. The paper introduces SIMSAT at a conceptual level, the current SIMSAT Linux Software Architecture, and the benefits of its framework "plug-in" design which make it so versatile, flexible and extendable. This plug-in framework allows the development of new SIMSAT Linux components that plug-in to the existing SIMSAT framework, for example to distribute the running of a SIMSAT simulator across a distributed network. The system, network performance, time synchronisation, and configuration control requirements of a real-time distributed model simulation system are identified. The paper demonstrates how these constraints were solved through the software architectural design principles applied. The results show that the new Distributed SIMSAT Linux Infrastructure Software meets the system and performance requirements of running distributed models across a dedicated network. The new SIMSAT can be utilised across diverse industries to run large-scale distributed simulations. SIMSAT is an ESA product for which holds the full IPR (intellectual Property Right).