Scalable Model-Centric Engineering Analysis Workflows Using Distributed Task Queues with the Tradespace Analysis Tool for Constellations

Earth-observing space systems increasingly consider combined operation of multiple systems in missions that require computationally intense simulation analysis and exhibit large architectural trade spaces. A long-term objective seeks to develop transparent, automatic, and fast mission performance evaluation techniques for preconceptual design studies in a digital engineering ecosystem. This paper develops and evaluates a scalable distributed task queuing system using the Tradespace Analysis Tool for Constellations (TAT-C) to evaluate constellation revisit and data latency performance. Implemented with the Celery Python library, the task queue manages and distributes analysis tasks across multiple workers. Specialized analysis workflows use chain and group task primitives to decompose large tasks into subtasks for distributed processing. A computational experiment evaluates execution time for a global coverage analysis task with 6–48 satellites, comparing 1–4 workers to a baseline with a single execution thread. Regression analysis characterizes the effects of task size and number of workers on execution time, finding effective scaling performance with an increasing number of workers and up to a sevenfold speedup compared to the baseline. Discussion reflects distributed computation challenges such as developing efficient workflows and balancing processing and input/output demands. Results demonstrate practical feasibility of scaling up computationally intensive analysis beyond the resources of a single host.