Our purpose is to identify the scope of model management and then to address the concerns of the community. To do this we may prioritize the items, and then to establish scenarios and best practices that address the concerns of the community.
The purpose of model management is to determine efficient structures to use. Structures include the repository, the models, and within the models. The considerations are many, and each one is discussed in the sections below.
Our current de facto definition of the concept of “model lifecycle management”, taken from the INCOSE IS 2014 Whitepaper, is:
Model Lifecycle Management (MLM) is a governance process synchronizing the create, read, update, and delete (CRUD) operations on heterogeneous models within the supporting modeling tools and model repositories, throughout the system development lifecycle. This is accomplished through the management of Model Configuration Items, including versions, variations, configurations and baselines of models, simulations, analysis results, and the tools that are used by multiple geographically dispersed users. In addition, MLM includes the management of all the metadata associated with the models, tools, and analysis results including who made the change, what changes were made, when and why, as well as information regarding the application of the model. A Model Lifecycle Management System (MLMS) is a set of elements that implement a model lifecycle management process, and may include people, hardware, software, data, and procedures.
This team shall be deemed successful when:
The measure of success perhaps of most interest to managers and modelers alike is: the implied internal rate of return, the rate at which the economic gains from the use of the model begins to exceed the economic costs of the development and maintenance of that model:
During the creation of a model, a certain amount of time is spent with overhead items like check-in, check-out, copying files, and deciphering complexity and relevance. The more time spent with these actions compared to the overall saved time, the less time is spent actually using a model and – most importantly – the less time spent actually achieving business goals and contributing to the business value.
Generally, when the IRR for a tactic is lower than the forecast price of money over the durations of consideration, then an enterprise may be better of not using the tactic because they'd earn a better return simply by banking the costs of development.
The Model Management Activity Team uses a private Google Group. That Group exposes the list of members to members.
Amit Fisher is the administrator of the Google Groups site. You can contact Amit at amfisher'at'us.ibm.com.
Roger Burkhart is the administrator of the OMG DokuWiki site.
Lonnie VanZandt is the chair of the MLM Activity Team.
The Model Management community authors white papers to define the scope, terminology, requirement, use cases, current practices and future direction for Model Management systems. These papers are co-edited by the Model Management community and can viewed here:
The event agenda has been annotated with notes and reference to slides presented. Please walk thus Model Management – Joe Bedocs, Brian Brady link for more details.
This file is updated from our discussion of it on the telecom. Please be sure to read the notes; not everything is on the slides. If you have comments, please send me a word document. Each comment should reference a slide number.
This is only something to throw darts at and start a more meaningful discussion. If you can send me comments by June 8th, I can get this updated for our telecom on June 18th.
During this meeting we had a discussion about sharing low level properties in a model with other models. In a specific example, Axel was interested sharing low level CAD properties like 0,0,0 reference point and axis orientation with other system models. But this need extends to all domain-specific models. We need exposure of all properties held by a model. We would like to get at the data without going through an API, without knowing the meta model, and preferably without know the repository details.
The group agreed to capture their ideas in some written form (model, drawing, sentences) and send to me by the end of the first week of March.
I will look for a place where these can be posted, and also look into setting up a model environment to capture these scenarios.
During this meeting we had a discussion about what the group wants to do. Ron spoke about metrics other than time such as accuracy, completeness, correctness, useful. That we might look at using measures of complexity, traceability and other possibilities for measures of success. The group is very interested in creating a list issues that are in the scope of model management. The group agreed to dig deeper into these areas in preparation for the IW Workshop.