Requirements Traceability – Does My Data Model Matter?
Nearly all engineering organizations have one or more initiatives underway to improve their product development process. Live Traceability, Model-Based Systems Engineering (MBSE), and digital engineering are the most common areas of focus. As engineers look over the fence and make fun of marketing types for being distracted by shiny objects, marketeers look back and see a similar behavior – just with geekier objects like SysML, digital twins, and simulation. The recurring pattern we see is that at some point during the early stages of the initiative, the realization hits that the data model for requirements across teams and projects is highly inconsistent and lacks consistent relationship rules among data objects. It becomes clear at this point that further progress on the initiatives cannot be made without first fixing the inconsistent and lacking data model.
Some teams will resist an effort to establish a consistent core data model. These teams will ask to keep the flexibility to refine their own engineering data shape and that is OK. The keyword is “refine” and not “define.” Having a consistent core data model, that some teams are allowed to refine for themselves, allows for innovation around the engineering process while still enabling process-wide, integration automation, Live Traceability, model-based systems engineering (MBSE), and digital engineering.
Current Requirements Data Model
For most companies, the data model mess came into existence through a project- and document-centric mindset with legacy requirements management tools. Each project team was allowed to modify their own data structures and each set of requirements lived alone as a document in a repository. This provided project teams with flexibility but over time and over dozens, hundreds, or thousands of projects has led to a challenging situation. We often find that teams have defined the same information in numerous different ways and that even within the same teams there is significant variance across documents. In short, the best way to describe the situation is as a repository of thousands of self-contained documents and no data model exists nor even a common definition of objects upon which to achieve Live Traceability, reuse across projects, MBSE, or digital engineering.
RELATED READING: Requirements Traceability – How to Go Live
What is Necessary to Move Forward
Organizations invest in software tools but have forgotten to invest in their data. A consistent data model is the best way to maximize the benefits of software tooling, but can only be achieved by spending time on analysis.
Jama Software has developed a Data Model Diagnostic (DMD) to help tackle this challenge, taking data from your legacy tool (IBM® DOORS®), understanding its shape and size, and transitioning the data into a model-based framework (Jama Connect). The DMD automates the analysis of the existing documents to determine the most common object definitions upon which to base a consistent data model going forward. Once a data model has been determined, the next step is to implement a model-based, requirements management solution that ensures compliance is maintained. As opposed to a legacy, document-centric requirements management tool, a model-based one ensures consistent application of all objects AND defines and maintains the relationship rules among the objects. This forms a model representation of the requirements in a consistent manner across projects and is a necessary requirement for MBSE and SysML modeling.