SoSTrades

SoSTrades is a multi-purpose and generic simulation management and execution platform, designed to explore complex system of systems scenarios in a collaborative way. It’s maturity is demonstrated by its capability to run WITNESS integrated assessment model as shown in various communications and demonstrations.
The implementation’s main characteristics are designed with a focus on native cloud componentization and object-oriented Python. It boasts a suite of key features across various domains:SoSTrades user interface offers a web-based, multi-user environment with interactive, publication-quality graphs that cater to different user expertise levels, using Angular, Flask, NodeJS, NGINX, and Plotly technologies .In terms of collaboration, it supports multiple users working concurrently on the same study, each in their own view, ensuring traceability of input modifications and validation of both inputs and outputs. User management is robust, featuring access rights management and group categorization, alongside encrypted scenario storage. The simulation data management is comprehensive, providing each study with a unique ID and permalink. It maintains traceability across studies, processes, disciplines, and models, and includes an execution cache to support incremental computation for alternative case studies. Studies can be accessed as components via REST API and displayed on remote application dashboards through HTML widgets. Strong identification and versioning of platform, models and processes is available. Simulation definition is prescriptive, detailing models that interact with automated coupling handling and offering various numerical strategies for coupling as plug-ins, built on significant extensions of the GEMSEO library – which allow setting up a “DevOps of models” capability with auto-updating when models are added / removed / updated. Execution of simulations is cloud-based, utilizing Docker and Kubernetes to run components on a set of infrastructure resources. Lastly, the numerical aspect includes advanced optimization formulations that support multi-objective and multi-scale considerations.

Very few simulation platforms today are natively cloud or containerized. Most platforms are still heavily connected to real physical infrastructure. SoSTrades is cloud native and uses a containerized approach that allows better scalability and integration with other cloud services and applications. It for example allows dynamic resources allocation with specific characteristics along the simulation. Many existing simulation platforms today have integration mechanisms inherited from 90’s graphical models and frameworks for building visualizations, such as workflow / dataflow models popularized by systems such as AVS, IRIS Explorer, and VTK toolkit. This leads to relatively poor scalability for complex system of systems simulations as it involves a complex and interwoven structure formed by an intricate network of interconnected simulation elements that is extremely difficult to edit and maintain, even generating questions on the reliability of the simulation specification. SoSTrades relies on a complete model & parameter ontology to automatically manage interconnections between models. Namespaces can be used to confine interactions between specified models, and allow only specific parameters to interact with other namespaces. Mathematical templates are automatically used to solve loops and cross-dependencies through multidisciplinary analysis (MDA) methods. This prescriptive approach scales much better for complex simulations than pure descriptive workflow ones, and allows for easy update and incremental completion of simulation. Finally, contemporary simulation platforms often lack robust features for collaborative multi-user and multi-view functionalities, particularly in the realms of simulation and integration with SPDM systems, where traceability and validation are essential. SoSTrades was meticulously designed from the outset to facilitate collaboration. It empowers various specialists to work together seamlessly, each from their unique perspective, on a unified project. The platform ensures that all modifications to the inputs are meticulously tracked and can be reverted if necessary. It also supports a structured validation process, where experts can authenticate inputs and scrutinize results, with all actions being systematically recorded. Moreover, SoSTrades incorporates sophisticated access management capabilities, enabling users to selectively share studies with peers while maintaining the option to encrypt sensitive information. This ensures confidentiality in a competitive yet cooperative shared workspace.