Introduction

This week, my colleague, Chris Schreiber (Lockheed Martin Space), and I co-chaired an engagement session titled “Next-Generation Systems Engineering Approaches: Deep Dives” at the ASCEND 2020 virtual conference. The abstract of the session was:

This engagement session will examine two innovative approaches to systems engineering of complex aerospace systems. NASA’s Jet Propulsion Laboratory has developed an ontological modeling language and methodology, and an associated software platform called OpenCAESAR (Computer Aided Engineering for Systems Architecture), which provides an ecosystem for developing and deploying tools for model authoring, integration, analysis, and reporting, as well as a semantic data warehouse system for managing SE information. The Lockheed-Martin Space Systems Company’s approach to digital engineering leverages a robust and flexible Digital Tapestry software infrastructure, a seamless digital environment driven by federating model-based engineering data across disciplines and managing data relationships from product conceptualization to realization. After each organization’s 30-minute presentation, the presenters will lead a discussion of the similarities and differences between the two approaches, and discuss plans for future development.

I went first and presented the openCAESAR platform and how it is used at JPL to modernize the practice of systems engineering for two flight projects: Europa Clipper and Psyche. Chris went second and presented their Digital Tapestry work at Lockheed Martin Space, which heavily uses the OpenMBEE platform. In the last section of the session, we collaborated on comparing and contrasting these two approaches to modernizing systems engineering in terms of their features and architectural decisions. The following tables are the result of this comparison.

Open Access

Feature	openCAESAR	OpenMBEE
Open Access	Platform capabilities are open sourced	Platform capabilities are open sourced

Information Architecture

Feature	openCAESASR	OpenMBEE
Information Architecture	Defined as a set of interrelated DSLs represented as OWL2-DL ontologies	Defined in conjunction with CM of CM’s construct

Information Authoring

Feature	openCAESAR	OpenMBEE
Model Authoring (COTS tools)	Supports authoring using some COTS tools supported by a (CAESAR) adapter	Supports authoring using some COTS tools supported by a (Syndeia) adapter and custom development
Model Authoring (custom tools)	Supports full authoring in the DSLs directly using standard authoring viewpoints in Eclipse or VSCode (desktop or web) or custom ones implemented declaratively	Supports some limited authoring out of the box using standard viewpoints in View Editor (part of OpenMBEE) or custom ones implemented imperatively

Authoring Tool Features

Feature	openCAESASR	OpenMBEE
Correct by Construction Authoring	Using custom authoring viewpoints built with declarative UI frameworks	Using methodology-specific profiles that are checked by rules
Automated Validation	Using live rules and batch DL reasoner and a rule based audit framework	Using custom-built live and batch rules executed upon editing models in MagicDraw
Automated Refactoring	Using custom refactorings in authoring viewpoints	Using refactorings built in COTS tool
Reusable Libraries	Using reusable class and instance ontologies	Using SysML profiles and model libraries
Collaborative Authoring	Possible in gitpod.io (web-based VSCode) and in Eclipse (using CDO Server)	Possible using ViewEditor and custom plugins

Information Management

Feature	openCAESASR	OpenMBEE
Source of Authority	Information has a single authority that maintains it in a known repo, but can be imported as read-only in other repos	Information has a single-source, but can be federated and edited in multiple repos
Configuration Management	Models are configuration managed by their native authoring tools, and (in the normalized DSL format) in Git repositories with support for branches, pull requests, etc.	Models are configuration managed using their native authoring tools CM capabilities
Change Management	Models changes are (normalized and) pushed to a git branch, where they get analyzed by a CI/CD pipeline and the results are made available to peer reviewers	Models changes are made, analyzed, and reviewed in design environments, with extension for local and/or global CI/CD
Release Management	Models are published (in their ontology representation) to maven repositories with semantic versions	Models are published to the MBEE repository with semantic versions
Dependency Management	Models declare read-only dependencies on each other using maven dependencies with semantic version range (e.g., 1.2.+)	Models have version-aware (Syndeia)traceability links to each other

Information Federation

Feature	openCAESASR	OpenMBEE
Model Federation	Models are federated in repositories managed by different authoring tools	Models are federated in repositories managed by different authoring tools
Model Cross Links	Models cross links are created between the ontological representations and stored in separate ontological (integration) models	Model cross links are created using Syndeia and stored in the Syndeia link database
Model Integration	Models are normalized to common DSLs first, then their explicit or implicit cross links are used to analyze them continuously	Model have their native repositories directly linked using Syndeia links, which are used to analyze their consistency on demand.
Model Reconciliation	The ontology representations of both models are compared to calculate the deltas to apply to the target native model	The Syndeia representations of both models are compared to calculate the deltas to apply to the target native model

Information Analysis

Feature	openCAESASR	OpenMBEE
Model Analysis	Models analysis using a variety of analysis tools (including a logical reasoner) is orchestrated by a CI pipeline	Model analysis is run “locally” on various tools as part of domain process
Model Reporting	Various reporting tools are used to create reporting viewpoints (docs, diagrams, etc.)	Various reporting tools are used to create reporting viewpoints (docs, diagrams, etc.)
Report Accessibility	Historical reports are archived and organized in a (central) dashboard that allows searching and viewing all report versions by various stakeholders	Reports are generated from model on demand and archived

For more information, please contact Maged Elaasar.