MBSE Wiki

Download

Figure 1 shows the steps in the conversion process from a SysML v1 model to a SysML v2 model which includes: (1) pre-process the SysML v1 model to prepare it for the transformation, (2) transform the SysML v1 model to a SysML v2 model, (3) post-process the SysML v2 model to better leverage the SysML v2 capabilities, and (4) validate that the SysML v2 model accurately reflects the intent of the SysML v1 model.

Figure 1 Model Conversion Process

In addition, further steps may be required to assess the impact of the SysML v2 model on existing artifacts that were derived from the SysML v1 model. The derived artifacts may need to be updated for the SysML v2 effort, but this is considered outside of the scope of the SysML v1 to SysML v2 model conversion. Each of these steps is summarized below.

This step involves pre-processing the SysML v1 model to prepare the model for transformation. The required pre-processing will depend on the transformation capability that the modeling tool provides so it is important to understand the tool capability and limitations. Performing the standard SysML v1 to SysML v2 model transformation requires that the SysML v1 model con-form to the SysML v1 specification, so the pre-processing of the model should ensure that SysML v1 model conforms to the SysML v1 specification. Any tool specific extensions along with other tool customizations to the model may need to be removed. However, the use of stereotypes and profiles should be supported by the transformation.

There are certain features of SysML v1 such as adjunct properties that have no explicit mapping to SysML v2. Part of the pre-processing could be to remove these elements or assess the impact of the transformation on these features and note that they may need to be addressed in the post-processing.

Circular dependencies should be identified to determine if and how they may impact the transformation and addressed accordingly. The SysML v1 model may also need to be reorganized to enable an incremental conversion process.

Creating a well-formed SysML v1 model that conforms to good practice will facilitate the conversion process. Model validation errors should be resolved to ensure the model is well-formed. Standard modeling conventions should be applied such as consistent naming conventions and minimizing ambiguities and redundancies. The amount of effort to pre-process can be compared to the additional effort to transform and post-process. The value of converting the SysML v1 model should be weighed against the level of effort to create the SysML v2 model from scratch.

This step involves transforming the SysML v1 model to a SysML v2 model. A SysML v1 model can be transformed to a SysML v2 model using a tool that can execute the standard SysML v1 to SysML v2 transformation specification. The standard transformation requires that the SysML v1 model be conformant to the SysML v1 specification in order to be transformed to a conformant SysML v2 model. It is important to understand the tool capabilities and limitations relative to the standard SysML v1 to v2 transformation specification.

The SysML v1 to v2 transformation specification defines the rules for transforming each kind of element in SysML v1 to a corresponding element in SysML v2. The transformation also includes rules for cases where there is no corresponding SysML v2 element. For example, a block in SysML v1 includes a meta property called 'isEncapsulated'. There is no equivalent concept in SysML v2 since the SysML v2 language designers did not see a need for this. However, there is a rule for how to address this in the transformation.

The tool should generate validation errors and warnings to indicate what aspects of the transformation were not successful. In addition, a manual inspection should be performed to compare the SysML v2 model with the SysML v1 model.

This step involves post-processing the SysML v2 model to leverage the SysML v2 capabilities. The transformed SysML v2 model may need to be reorganized and refactored to fully leverage the SysML v2 capabilities. The reorganizing and refactoring should apply the usage-focused modeling paradigm to more fully leverage the SysML v2 capabilities. This paradigm is briefly dis-cussed in the section entitled “Post-process the SysML v2 Skyzer model.”

It is imperative to validate that the SysML v2 model accurately reflects the intent of the SysML v1 model. This can be done by comparing the two models. This may include reproducing selected views of the SysML v2 model such as a system hierarchy and carefully comparing it with the system hierarchy in the SysML v1 model. It is anticipated that tool vendors may be able to generate automated comparison reports to assist in the inspection. Comparing execution and analysis results from the SysML v2 model with the corresponding execution and analysis results of the SysML v1 model may also assist in the validation.

A tool is expected to support the SysML v2 standard views which can render similar information that is contained in the nine standard SysML v1 diagrams. However, the layout information is not pre-served and would need to be adjusted manually to align with the original SysML v1 diagram.

The following are additional steps outside of the scope of the SysML v1 to SysML v2 model con-version, but may be critical to the program’s success.

This step involves identifying artifacts that were generated or derived from the SysML v1 model and assessing the potential changes to these artifacts that result from the SysML v2 model. The artifacts can include specifications, architecture description documentation, requirements traceability reports, related analysis, test plans, and others.

This step involves updating the derived artifacts as needed. This step may be dependent on integrations between the SysML v2 modeling tool and other modeling tools and repositories.

The conversion process should be performed incrementally rather than converting the entire model in a one-time batch process. As part of the pre-processing, the SysML v1 model can be partitioned to minimize coupling and enable incremental transformation. For example, the model can be partitioned into packages that contain the structure, behavior, and re-quirements and further partitioned into mission, system, and subsystem levels. The incremental conversion process can first transform the structure, then transform the behavior, and then transform the requirements.

It should be noted that the transformation is currently one-way from SysML v1 to SysML v2, and that there is no standard transformation specification to transform a SysML v2 model to a SysML v1 model. This is because many of the capabilities in SysML v2 are not supported in SysML v1. For example, SysML v1 supports a block decomposition but does not support a SysML v2 part decomposition.

The transformation of a classified SysML v1 model should preserve all classification markings in the SysML v2 model. A standard security extension should be applied that leverages the metadata capability in SysML v2. A well-defined process should be es-tablished to ensure all markings are properly applied. This should include manual inspection of the model. The same classification procedures that apply to the SysML v1 model should apply to the SysML v2 model and must be adhered to.

The SysML v2 API configuration management services can be leveraged to manage the configuration of the SysML v2 models beginning with the initial transformation. Typical branch and merge concepts can be used to manage updates to the model. The configuration management of the SysML v2 model should be managed in the broader context of the overall engineering ecosystem that typically involves a product lifecycle management environment and workflow/issue management applications such as Jira.