Abstract. This paper reviews recent and current efforts to implement Model-Based Systems Engineering at the Canadian Space Agency. A mission under concept development is being used both for demonstration purposes and also for real process design with SysML. Several support and bridge tools have been developed to help with key process issues. A number of technology issues have come out of our work, leading to a wish list of improvements we would like to see. Also, we have been dealing with organizational resistance that is preventing a wider acceptance of MBSE.

Model‐Based Systems Engineering (MBSE) has been developing for some time, and has recently acquired new impetus with the completion of the Systems Modeling Language (SysML). This paper envisions taking MBSE much further, to a future of highly integrated and automated design and verification coupled with advances in simulation and domain linkage to allow the synthesis of complete systems from requirements into mathematical models and then into physical realizations. This would permit the application of three of the most successful approaches from agile software development, namely rapid, iterative development of the system starting with the highest value functions, facilitating continual reassessment of the future direction, and continual regression testing to ensure that system bugs are identified and removed rapidly. We envisage the requirements and the model evolving together from proto‐requirements and proto‐model in increasing detail until the point at which the model can be realized with real hardware and software. Taking this further, the MBSE engine can perform trade‐offs and optimization on the design. Implementing this vision requires progress in a number of technologies, such as data exchange between domain tools. At this time, much engineering effort is consumed in people communicating and mediating information and translating it from one form to another (e.g. system design to mechanical design). If we can realize the vision proposed, we can remove much of the burden of information mediation and optimization, allowing engineers to focus on their expertise and larger issues. The potential savings in labour are huge.